CULTIVATING WISDOM THROUGH A SERVICE-LEARNING EXPERIENCE 
 
 
 
 
A DISSERTATION 
SUBMITTED TO THE FACULTY OF THE GRADUATE SCHOOL 
AT THE UNIVERSITY OF MINNESOTA, TWIN CITIES 
BY 
 
 
 
ANDREW BAILEY 
 
 
  
 
 
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS  
FOR THE DEGREE 
DOCTOR OF PHILOSOPHY 
 
 
 
 
Dr. Keith C. Russell, Adviser 
 
 
April 2009 
 
  
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
© Andrew William Bailey 2009 
 
i 
 
Acknowledgements 
 
 There are many to whom I owe much gratitude for their contribution to this 
project.  First, I would like to thank the circle of professionals who guided me through 
this mind-bending process.  Thanks go to Dr. Corliss Outley, who somehow saw enough 
potential in me to bring me onto the program.  My deep gratitude goes to Dr. Keith C. 
Russell, for adopting me and providing just enough challenge and support to keep me on 
this side of the edge.  He listened graciously, though skeptically, to my proposal to 
research Wisdom and helped me keep my feet on the ground, even if my head was in the 
clouds.  Thanks also to Dr. Leo McAvoy, for his genuine concern and thoughtful 
guidance.  His balance of theoretical competence and practical know-how are an example 
to aspire to; a wisdom demonstrated by a life well-lived.  Thanks to Dr. Jerome Stein for 
making me explore ideas I never would have encountered on my own, and for helping to 
smooth over some of my rough edges.  His wit, humor, and authenticity managed to 
sneak the learning past my guard, despite my best efforts.  Finally, thanks to Dr. Stephen 
Ross, who adopted me for the second time and saw me through the important final steps.  
His feedback was invaluable in the final writing stages and in navigating all the red tape. 
 
  I am very grateful to my cohort; Tami, T, Mike, Gen, H.K., Sean, & JoePa, for 
help in wrestling with difficult topics and for the laughs that kept me sane through it all.  
Thanks also go to David and Trisha Fenrick, for their support and friendship.  Amy and 
Ed, I can’t thank you enough for the good times and good conversations.  Life would be 
much diminished without such a good group of friends. 
 
 Thanks to the STLF, to Irene, Greg, and Brian.  You guys are about as much 
inspiration as I can handle.  I’m honored to work with such an aspiring group of leaders 
and hope to live up to your standards someday.  I look forward to seeing what you will 
accomplish in the future.  I expect great things, and know you expect the same from me. 
 
 Finally, I am completely in debt to my wife, and my family back home in VA.  
Big accomplishments are never the work of one man, but a result of the love, sweat, and 
hope garnished on him by others.  Mom and Dad always supported my adventures, even 
when they didn’t seem to add up.  They provided everything I need to flourish, including 
the freedom to grow.  Thanks to Gam, for being the anchor of our family and for her 
unending love and interest in her grandchildren.  Thanks to my brothers, Greg and Erick, 
for cultivating my tenacity from an early age, and for keeping me in my place.  And 
finally, to Stephanie, my wife, who literally made this whole endeavor a possibility.  Her 
moral, mental, emotional, and spiritual support nourish me every day. 
 
ii 
 
Dedication 
 
 
 
 
 
 
 
 
 
To Stephanie, my best friend, supporter and the delight of my heart.   
Thank you for joining me on this adventure, and for the color you add to my life. 
iii 
 
Abstract 
 
Recent research has indicated that the acquisition of knowledge alone is not 
sufficient to ensure the common good.  As such, this study measured the effectiveness of 
a service-learning experience in cultivating various personal and interpersonal assets 
necessary for optimal human development; collectively defined as Wisdom.  Wisdom 
was measured as a latent variable that is evidenced by cognitive, affective and reflective 
effect indicators (Ardelt, 2003).  A quasi-experimental design was used to measure the 
difference in wisdom attributes for those involved in a service-learning program (n= 288) 
as compared to a similar control group (n= 321).  Wisdom was measured before, after, 
and one-month following an intense service-learning tour.  Structural Equation Modeling 
(SEM) was used to determine the comparative influence of Social Environment, Social 
Relationships, Openness to Experience, and Civic Attitude on Wisdom domains.  Latent 
Growth Modeling (LGM) was used to determine the growth trajectories of wisdom over 
the course of five weeks, and to identify variables that influenced those trajectories.  
Finally, trip components were ranked by mean scores as to their perceived importance in 
facilitating growth in the three Wisdom domains. 
 Results indicate that service-learning participants reported significant gains in 
Wisdom domains as a result of the trip, with overall Wisdom scores remaining significant 
one month after the trip.  Control group participants reported declines in all measures, 
with the exception of Civic Attitude.  In addition, the proposed SEM fit the data well, 
providing deeper insight into predictors of the development of Wisdom in early 
adulthood.  Implications for the fields of education and recreation are discussed within 
the context of this study and previous relevant research. 
iv 
 
TABLE OF CONTENTS 
CHAPTER I ........................................................................................................................ 1 
INTRODUCTION ........................................................................................................... 1 
Statement of the Problem ............................................................................................ 4 
Purpose of the Study ................................................................................................... 6 
Theoretical Models and Hypotheses ........................................................................... 7 
Definition of Terms ................................................................................................... 16 
Limitations of the Study ............................................................................................ 16 
Outline for Dissertation Proposal .............................................................................. 18 
CHAPTER II ..................................................................................................................... 19 
REVIEW OF THE LITERATURE ............................................................................... 19 
Experiential Learning ................................................................................................ 19 
Service-Learning ....................................................................................................... 21 
Wisdom ..................................................................................................................... 34 
Wisdom and Experiential Learning........................................................................... 42 
CHAPTER III .................................................................................................................... 45 
METHOD ...................................................................................................................... 45 
Research Design ........................................................................................................ 45 
Participants ................................................................................................................ 49 
Instruments ................................................................................................................ 52 
Data Analysis ............................................................................................................ 55 
CHAPTER IV ................................................................................................................... 65 
RESULTS...................................................................................................................... 65 
Summary of the Instrument ....................................................................................... 65 
Response Bias ........................................................................................................... 68 
Selection Bias ............................................................................................................ 68 
Construct Validity and Reliability............................................................................. 69 
MANOVA and t-Test Analyses ................................................................................ 72 
Service-Learning Component Ranking ..................................................................... 75 
Structural Equation Model ........................................................................................ 78 
Latent Growth Analysis ............................................................................................ 85 
CHAPTER V ..................................................................................................................... 93 
CONCLUSIONS ........................................................................................................... 93 
Cross-Sectional Conclusions ..................................................................................... 93 
Longitudinal Conclusions ....................................................................................... 104 
Latent Growth Analysis (LGA)............................................................................... 112 
Conclusions and Implications ................................................................................. 115 
APPENDIX A. PARTICIPANT CONSENT FORM ..................................................... 120 
APPENDIX B.  PRE-TRIP QUESTIONNAIRE ............................................................ 121 
APPENDIX C.  POST-TRIP SURVEY SUPPLEMENT ............................................... 126 
BIBLIOGRAPHY ........................................................................................................... 128 
 
v 
 
LIST OF TABLES 
 
Table 3.1. Time-Table for Data Collection  ...................................................................... 48 
Table 4.1. Items and Factors Included on the Supplemental Post-Test Questionnaire  ... 67 
Table 4.2.  Independent t-test on Initial Status of Treatment and Control Groups  ......... 69 
Table 4.3.  F-Statistic Values for Outcome Variables Using Various Sampling Methods 69 
Table 4.4.  Cronbach’s Alpha, Skewness & Kurtosis Levels for Dependent Variables  ... 70 
Table 4.5.  Fit Indices for the CFA of the Wisdom Construct  .......................................... 71 
Table 4.6.  Inter-domain Covariance Within the Wisdom Construct  ............................... 71 
Table 4.7. MA/OVA for Difference Scores of Wisdom Sub-Domains.............................. 73 
Table 4.8.  MA/OVA for Difference Scores of Civic Attitude and Openness .................. 74 
Table 4.9.  Paired t-Test for Pre and Post-Test Civic Attitude and Openness Scores  ..... 75 
Table 4.10.  Reliability Measures for Trip Component Factors  ...................................... 76 
Table 4.11.  Component Ranking for the Reflective Domain ............................................ 76 
Table 4.12.  Component Ranking for the Affective Domain ............................................. 77 
Table 4.13.  Component Ranking for the Cognitive Domain  ........................................... 77 
Table 4.14.  Ranking of Total Scores for Trip Components  ............................................. 78 
Table 4.15.  Standardized Direct Effects of Predictors on Wisdom Domains .................. 79 
Table 4.16.  Fit Indices for the Multi-group Confirmatory Factor Analysis  ................... 81 
Table 4.17.  Comparison of Fit Indices for the Proposed Model and Alternative      
Models ............................................................................................................ 82 
Table 4.18.  Percent of Variance Explained in Each Item ................................................ 84 
Table 4.19.  Correlations of SEM Variables ..................................................................... 85 
Table 4.20.  Chi-Square test for the Difference in Pre and Post and Follow-up      
Wisdom Scores ............................................................................................... 87 
Table 4.21.  Predictors of Latent Intercepts and Growth Trajectories of Wisdom  
Domains  ........................................................................................................ 89 
Table 4.22.  Time Invariant Predictors of LGA for Wisdom, Civic Attitude, and   
Openness   ...................................................................................................... 90 
Table 4.23.  Covariance of Time-Variant Predictors and the Wisdom Construct  ........... 92 
vi 
 
LIST OF FIGURES 
 
Figure 1.1.  Kolb’s Experiential Learning Model with Learning Styles ............................ 8 
Figure 1.2.  The Virtues in Kolb’s Learning Model ........................................................... 9 
Figure 1.3.  SEM Path of Wisdom Development through Service-Learning ................... 14 
Figure 2.1.  Service-Learning Model with Kolb’s Experiential Learning Model ............ 23 
Figure 2.2.  Model of the Outward Bound Process .......................................................... 32 
Figure 2.3.  Baltes Theoretical Framework of Wisdom ................................................... 37 
Figure 3.1.  LGA for Wisdom with 3 Dimensions ........................................................... 58 
Figure 3.2.  LGA for Wisdom with 3 Predictors .............................................................. 69 
Figure 3.3.  Three Dimensional Wisdom Construct  ........................................................ 61 
Figure 4.1   Mean Difference scores for Control and Treatment Groups on Wisdom Sub-
domains .......................................................................................................... 73 
Figure 4.2.  Mean Difference Scores for Control and Treatment Groups on Civic 
Attidude and Openness .................................................................................. 74 
Figure 4.3.  Standardized Regression Weights for Predictors on Wisdom Domains ........ 80 
Figure 4.4.  Full Final Model ............................................................................................ 83 
Figure 4.5.  Mean Structure Analysis for Pre, Post, and Follow up Wisdom Scores ........ 86 
Figure 4.6.  Comparison of Group Mean Trajectory with Individual Trajectories .......... 88 
 
 
 
1 
 
CHAPTER I 
 
INTRODUCTION 
 
“If we are willing to concede education as the process of forming fundamental 
dispositions, intellectual and emotional, toward nature and fellow men, philosophy may 
even be defined as the general theory of education (Dewey, 1916, p. 331; emphasis in 
original).” 
 
The scope of educational ideals has been narrowed significantly in the modern 
era.  Since Plato’s portrayal of the good and just philosopher, or “lover of wisdom” (Lee, 
1987), the aims of education have waned to a point where many institutions today are 
concerned only with cognition.  The disregard of other important developmental assets 
has caused alarm amongst many top educators.  Sternberg (2003) states that, “In U.S. 
society, cognitive skills have become practically equated with intellectual skills- the 
mental bases of intelligence.  This equation is a mistake (p. 147).”  Other researchers 
contend that an increase in global interdependence and the concurrent increase in 
egocentrism is a volatile combination that must be countered by placing a stronger 
emphasis on civic responsibility in education (Delve, Mintz, & Stewart, 1990; Putnam, 
2003).  If the future is plagued with conflict, the instability is not something that exists 
“out there somewhere,” but a current that flows from inside individuals.  To ensure both 
individual and social well-being, it may be necessary to not only teach students to recall 
facts and to think critically, but to think wisely as well (Sternberg, 2003). 
Wisdom is becoming an increasingly popular concept in various fields of study 
(Assmann, 1994; Baltes, Gluck, & Kunzmann, 2005).  Though most disciplines have 
avoided proposing a model of a virtuous life (Sternberg, 1998), the recent call for a 
positive psychology is leading the way to a conceptualization of “optimal human 
2 
 
performance” (Baltes, Gluck, & Kunzmann, 2005).  This search for the pinnacle of 
human development has brought about the re-discovery of the ancient concept of 
wisdom.  The leading wisdom researchers come from the field of gerontology, perhaps 
intrigued by Erikson’s (1982) conceptualization of wisdom as the eighth and final stage 
of human development. 
Wise persons have been found to posses rich knowledge and experience in 
matters of the human condition, self-knowledge, openness for new experiences, the 
ability to learn from mistakes, and good intentions in action (Baltes, Gluck, & 
Kunzmann, 2005).  Wisdom is thought to arise from intense learning and practice within 
a social-cultural environment which encourages its pursuit.  It emerges as a result of a 
variety of experiences that interact and collaborate.  Wisdom involves the orchestration 
of many factors (cognitive, personal, social, and spiritual) and is developed through “the 
guidance of mentors and the mastery of critical life experiences (Baltes, Gluck, & 
Kunzmann, 2005, p. 332).” 
As a multi-dimensional construct, wisdom represents a measure of human 
development based on the balance of several intrapersonal subsystems (Sternberg, 1998).  
Though it has been traditionally seen as an exclusive virtue of the elderly, recent research 
has challenged these assumptions.  The “seeds” of wisdom are evident even in early 
adolescence and “wisdom performance” has been shown to increase sharply between the 
ages of 15 and 25 (Pasupathi, Staudinger & Baltes, 2001; Richardson & Pasupathi, 2005).  
Few studies have been done to determine the impact of “interventions” in facilitating 
wisdom performance.  However, research suggests that an appropriate intervention would 
include experience and opportunities for reflection.  Specifically, these experiences 
3 
 
would occur within a variety of positive and supportive social contexts and would include 
the opportunity for group collaboration (Staundinger & Baltes, 1996), as well as moral 
challenges that incite some degree of profundity (Webster, 2003).  In addition, there must 
be ample opportunity for personal and group reflection, which transforms knowledge into 
wisdom and transforms the individual in the process (Ardelt, 2003).   
 Concrete experience, collaboration, challenges, and opportunities for reflection 
are essential elements in experiential education, having been espoused for decades.  The 
philosophy of John Dewey is central to the theory of experiential education.  Dewey 
(1998) emphasized that education is not just a preparation for future vocation, but rather a 
continual process by which one learns from life experiences.  This type of education 
develops each person to their fullest capacity so s/he can evaluate new situations and 
respond accordingly.  Instead of “filling” each student with accumulated facts, Dewey 
envisioned a method of education that developed a person’s ability to perceive, to make 
connections, and to successfully adapt when faced with new challenges. 
The key to this kind of development is reflection upon experience.  By reflecting 
on life experiences, one can make connections to other information and experiences, 
which then leads to the perception of meaning (Dewey, 1934).  In order for experience to 
lead to meaning, it must occur within a specific context (setting or milieu) and it must 
generate tension through challenges (physical, mental, emotional).  This tension 
motivates the participant to adapt to the situation, which generally results in growth 
through a transformation of knowledge, beliefs or values.  This growth then becomes a 
continuous process as new knowledge is applied to new contexts, leading to increased 
meaning and fulfillment.  Wisdom and experiential learning share several core elements.  
4 
 
This study will provide insight into the potential for facilitating the development of 
wisdom through an experiential learning program for young adults. 
Statement of the Problem 
“[Wisdom’s] survival implication is as an antidote to knowledge that pursues selfish, 
short-term, and limited goals that often turn out to have disastrous consequences for the 
very persons they were intended to benefit- not to mention other people or the nonhuman 
environment (Csikszentmihalyi & Rathunde, 1990, p. 48).”   
 
Psychologists suggest that educators must no longer avoid making value 
judgments about the means and ends of a “good life” (Schwartz, 2000; Seligman & 
Csikzentmilhayi, 2000).  By focusing on the techniques of progress, our society has 
neglected to reflect on the goals of that progress (Baltes, Gluck, & Kunzmann, 2005; 
Marcel, 1954; Pacione, 1999; Schumacher, 1977).  This neglect is being realized in 
various ways, including a loss of a sense of direction (Dyke, 2006; Schwartz, 2000), an 
increase in violent crimes, family breakdown, psychosomatic complaints 
(Csikzentmilhayi, 1999), as well as the modern environmental crisis (Gottlieb, 2001; 
Mayer & Frantz, 2004; Schultz et al., 2004).  That these deficits have occurred 
simultaneously with vast increases in IQ scores (Flynn, 1998), testifies to the impotence 
of knowledge alone in dealing with modern dilemmas.   
Much of the lack of empirical attention paid to a cultivation of the “good life” can 
be attributed to a lack of efficient psychological measurements for intangible variables.  
Early successes in behaviorist and positivist research at the beginning of the 20
th
 century 
may have led to a devaluation of more “vague” and “fuzzy” psychological concepts 
(Holliday & Chandler, 1986).  Wisdom may have been further devalued because of its 
association with old age; a less productive time of life that is not highly esteemed in 
5 
 
modern societies (Holliday & Chandler, 1986; Marcel, 1954).  Furthermore, wisdom’s 
multi-dimensional nature has rendered it difficult to capture empirically.  Modern 
psychometrics, coupled with new analytic techniques such as latent variable modeling, 
are proving valuable for the investigation of this complex construct.  In this way, 
researchers have been able to reach beyond the scope of “Scientism” (Habermas, 1970) to 
explore the interactive and dynamic process of latent constructs such as wisdom (Ardelt, 
2004, Baltes, 2004).   
While previous research has contributed much to our understanding of certain 
phenomenon, many psychologists and philosophers are calling for a longer-term, more 
complex measurement of human development (Baltes, Gluck, & Kunzmann, 2005).  
Knowledge, it seems, is not enough to ensure the “common good.”  With the 
development of instruments and methods for the measurement of wisdom comes the 
opportunity to begin research into the facilitation of growth in wisdom.  A few efforts 
have been made to improve wisdom performance.  Baltes and Smith (1996) were 
successful at increasing scores for wisdom-related knowledge using imagery and social 
collaboration.  This study, however, included only adults and there was no longitudinal 
analysis to determine if the results had any long-term impact.  Additionally, questions 
were raised as to the validity of their method of measurement as a true indicator of 
wisdom.  Further research needs to be done to determine if wisdom can be facilitated in 
other populations using a variety of measures with robust psychometric properties. 
Previous research has reported that wisdom increases most sharply between the 
ages of 15 and 25 (Baltes, Gluck, & Kunzmann, 2005).  This identifies the period of 
adolescence as a key stage for facilitating the growth of wisdom.  Richardson and 
6 
 
Pasupathi (2005) have identified possible antecedents to wisdom that are evident in 
adolescents.  So far, however, there have been no efforts to facilitate the growth of these 
antecedents through interventions or intentional programming.  Sternberg (2003) has 
implemented a curriculum designed to improve wisdom performance in adolescents, but 
the results of this study have not yet been determined.  Further research into the 
facilitation of adolescent wisdom should be the focus of future research. 
Pilot studies conducted for this research have revealed promising findings.  
Ardelt’s (2003) Three-Dimensional Wisdom Scale (3D-WS) was used to measure gains 
in affection, reflection, and cognition of students involved in a week-long service-
learning tour.   These students (ages 14-19) showed significant gains in affection and 
cognition after the service-learning experience.  This was the first time the 3D-WS had 
been used with a population of this age, and further research must be done to establish the 
reliability of the instrument with adolescents.  Furthermore, there was no follow-up 
measurement, no control group, and no efforts were made to determine which aspects of 
the service-learning process contribute to the growth in wisdom domains.  This study 
addresses many of the weaknesses of the pilot study and seeks to identify the key 
components of the trip and determine how they contribute to each wisdom domain.   
Purpose of the Study 
 The purpose of this study is to determine the impact of an experiential education 
program that incorporates service-learning on scores in wisdom domains.  This study 
measured the impact of a nine-day service-learning trip on wisdom in early adulthood, 
and determined which aspects of the trip influenced growth in each wisdom domain.  
Trip participants were given Ardelt’s (2003) 3D-WS before and after the experience, then 
7 
 
again one month after completion of the program to determine the stability of outcomes.   
Gains in wisdom domains were assessed in terms of mean scores.  Post-trip 
questionnaires included a list of trip elements, which the participants rated in terms of 
importance for their learning in various domains. 
Theoretical Models and Hypotheses 
Experiential Learning  
Many models of experiential learning have been proposed, the best-known being 
that promoted by Kolb (1984).  Kolb’s experiential learning model (Figure 1.1) is based 
on a four-stage theory developed by Lewin and Dewey (Dyke, 2006).  Kolb’s model is 
often criticized as being too simplistic, but this simplicity has enabled it to serve as the 
basis for many other learning models.  Experiential learning theory (ELT) defines 
learning as “the process whereby knowledge is created through the transformation of 
experience.  Knowledge results from the combination of grasping and transforming 
experience (Kolb, 1984, p. 41).”  Service-learning is a type of experiential learning 
through which participants address human and community needs together, with 
structured opportunities for reflection and discussion.  The experiences are intentionally 
designed to create opportunities for community service and engagement. 
Kolb (1984) viewed concrete experience as the first step in the learning process.  
Concrete experience is defined very loosely and can include any variety of activities or 
situations that a person encounters in life.  In the second stage of Kolb’s model the 
individual reflects on their experience, comparing it with previous knowledge and beliefs.  
Having reflected on the experience, the individual then “abstractly” applies his/her new 
insights to other aspects of his/her life.  This abstraction is then tested by applying the 
8 
 
new insights to future experiences (experimentation).  This model was chosen because it 
has been established through a vast body of research (Kolb, Boyatzis, & Mainemelis, 
2000) and because it is sufficiently parsimonious to be applied to the program involved in 
this study. 
Further development of Kolb’s model included the addition of learning styles to 
each quadrant.  Those in the Divergent quadrant have dominant learning abilities 
associated with Concrete Experience and Reflective Observation.  These people tend to 
be imaginative and emotional, and display excellent abilities in brainstorming and 
viewing situations from various points of view.  Those in the Assimilating quadrant enjoy 
abstract ideas and are often more concerned about the logical soundness of a theory than 
they are about its practical value.  Those who are of the Converging learning style enjoy 
finding practical uses for ideas.  As such, they tend to enjoy technical tasks with solvable 
problems, rather than “thorny” social issues.  Finally, Accommodating learners prefer to 
Figure 1.1.  Kolb’s Experiential Learning Model with Learning Styles. 
9 
 
learn from hands-on experience, to explore new challenges, and to test out different 
approaches in the field (Kolb, 2000). 
Due to the influence of radical educators, such as Paulo Freire, Kolb added  
another layer to his learning model to emphasize that ELT promotes the development of a 
culture as well as that of individuals.  As illustrated in Figure 1.2, Kolb distinguished 
values from facts, and relevance from meaning, placing them in the location where he 
supposed each concept played the most important role.   
 
 
Values, for instance, predetermine the concrete experiences in which one engages, 
while facts help one to abstract and apply the ideas we gain from those experiences.  
Furthermore, in the relevance stage one organizes the observations of their experience 
and then attempts to apply that meaning purposefully through experimentation.  Finally, 
Kolb adds a virtue to be associated with each quadrant of the model.  Love enables the 
INTEGRITY 
Feeling 
 
 
 
 
 
 
Courage 
Wisdom 
Thinking 
Love 
Acting Observing Meaning Relevance 
Justice 
V
a
l
u
e 
F
a
c
t 
Figure 1.2. The Virtues in Kolb’s Learning Model (Stewart, 1990) 
10 
 
learner to go beyond their self-centered observations and empathize with others.  Justice 
helps one to negotiate a set of accumulated facts in order to make fair and inclusive 
judgments.  Wisdom goes beyond the accumulation of facts to assist in making 
meaningful decisions about one’s behavior.  Lastly, courage provides the motivation to 
find meaning in one’s actions and to persist in the application of acquired knowledge 
(Stewart, 1990). 
The Experiential Learning Model represents a mode of learning that is available 
to everyone, regardless of age and access to educational resources.  As illustrated above, 
it is an effective model for all types of learners and supposedly improves the experience 
of both the learner and the culture within which the learner is engaged.  Though ELT is 
not teacher-dependent, it assumes that there are intentional opportunities for concrete 
experiences and reflection (Kolb, 1984).  Wisdom also develops as a result of experience 
and reflection.  Based on the Experiential Learning Model, the following null-hypothesis 
was tested: 
1a.  Those participating in a service-learning trip will demonstrate no higher gains 
in wisdom domains than the control group. 
In addition, to testing the above hypothesis, this study sought to explore the 
relationship of important elements in experiential learning.  As such, the following 
research question was addressed: 
1b. Which components of the service-learning experience are perceived as most 
valuable for growth in each wisdom domain? 
 
 
11 
 
Three Dimensional Model of Wisdom   
Ardelt’s (2003) conceptualization of wisdom is based on the pioneering implicit 
study performed by Clayton and Birren (1980).  That study produced a list of 
characteristics associated with wise people, and categorized these traits into three 
categories: Cognition, Affection, and Reflection.  Items measuring the cognitive 
dimension assess one’s willingness and ability to understand phenomena on a deep level, 
knowledge of the ambiguity of human nature, knowledge about the limits of knowledge, 
and an awareness of life’s uncertainty.  While this dimension measures one’s beliefs 
about the world, which may be affected by perspective-taking, this dimension does not 
measure perspective-taking directly.  The affective dimension measures one’s demeanor 
and sympathetic compassion towards others.  This domain is “other-oriented” and wise 
ratings should demonstrate a lack of self-centeredness.  The final dimension, reflection, 
measures the degree to which one tries to overcome their subjectivity, their willingness to 
look at situations from various points of view, and their acceptance of responsibility for 
their life circumstances (Ardelt, 2003). 
 All three domains are considered to be necessary but not sufficient indicators of 
wisdom.  Measured in this way, wisdom is viewed as a personality characteristic, and not 
as a type of performance that would vary from one context to another (Ardelt, 2003).  An 
effective integration of the three dimensions would indicate the presence of wisdom 
within the individual.  For example, a person who scored high in the cognitive dimension 
could be intelligent, but they would not be wise if they lacked compassion.  Conversely, a 
compassionate person may be very sympathetic toward another person, but may not be 
very effective at interacting with them or helping them without background knowledge 
12 
 
and perspective.  According to Ardelt (2003), the reflective domain is the “hub” of 
wisdom, as it simultaneously increases one’s willingness to understand life on a deep 
level and their compassion for others.  However, reflection is supposedly associated with 
age (Assmann, 1994) and a pilot test supported this assumption (Bailey & Russell, 2008).  
Therefore, it may be unlikely that scores would increase in the reflective domain over 
such a short period of time.   
Based on the above model the following null-hypotheses were formed: 
2a.  There will be no difference in pre- and post-trip wisdom domain scores for 
service-learning participants. 
2b.  There will be no difference in pre- and follow-up wisdom domain scores for 
service-learning participants. 
Proposed Structural Equation Model  
 Previous studies have identified several important antecedents to wisdom.  
Staudinger & Pasupathi (2003) found that intelligence (particularly crystallized 
intelligence; i.e. verbal and math performance) and openness to experience were the best 
predictors of wisdom-related performance in adolescents.  In addition, Ardelt’s (2000) 
longitudinal study of elderly adults identified one’s social environment in early adulthood 
as the only significant predictor of wisdom as measured 40 years later.  Previous research 
has suggested that experience in a variety of social contexts is conducive to wisdom 
(Baltes, Gluck, & Kunzmann, 2005).  Accordingly, a preliminary research study (Bailey 
& Russell, 2008) revealed a significant positive correlation between the amount of social 
engagement and wisdom scores.  The authors suggested that this may be related to the 
concept of Social Capital as conceptualized by Coleman (1990), and that a high level of 
13 
 
engagement with a variety of social groups could broaden one’s perspective and increase 
one’s openness to experience.  Finally, as wisdom is a virtue for the “common good” 
(Baltes, Gluck, & Kunzmann, 2005) and service learning increases awareness and 
concern for the “social good” (Rhoads, 1998), it is conceivable that one’s civic attitude 
could play a key role in wisdom development.  Thus, the proposed model included 
measures of civic-efficacy, civic responsibility, and democratic attitude.  
 This study tested the hypothesized and theoretically-grounded structural equation 
model as illustrated in Figure 1.3.  Social environment, and social relationships are 
exogenous variables that are hypothesized to influence one’s openness to experience and 
their civic attitudes as well as directly affecting overall wisdom.  Openness to experience 
and civic attitudes affect all three domains of wisdom, being stimulated by the treatment 
of the service learning experience.  Since the service-learning trip is based on Kolb’s 
(1984) Experiential Learning Model, the treatment can be viewed as a series of cycles 
that involve experience and reflection.  These components have a direct effect on wisdom 
scores as well as an indirect effect which is associated with openness and civic attitudes.   
14 
 
 
Figure 1.3.  SEM Path of Wisdom Development through Service-Learning 
Based on the above model, the following hypotheses were developed: 
 3a.  Social environment and social relationships will have a significant direct 
effect on wisdom scores. 
 3b.  Openness to experience and civic attitude will have a stronger direct effect on 
wisdom scores than will social environment and social relationships. 
 In addition, the following null hypotheses will be tested: 
3c.  Service-learning participants will demonstrate no higher gains in civic 
attitude and openness than the control group. 
3d.  There will be no difference in pre and post-trip scores for civic attitudes and 
openness-to-experience for service-learning participants. 
Significance of the Study 
 This study contributes to the existing literature that emphasizes the need for a 
more comprehensive measure of human development (Baltes, Gluck, & Kunzman, 2005; 
Social  
Environment 
Social  
Relationships 
Cognition 
Affection 
Reflection 
openness 
Civic 
attitude 
W 
I 
S 
D 
O 
M 
Treatment 
Demographics 
15 
 
Sternberg, 1990; Sternberg 2005).  Current measures of progress are too narrow to ensure 
a future directed toward the “common good” (Sternberg, 2003).  As such, it is vital that 
future research contributes to more complex conceptualizations of human development.  
In addition, it helps to verify the measurement of wisdom through a self-report measure 
that can be easily administered in other settings. 
This study represents the first known attempt to facilitate the growth of wisdom 
domains through experiential learning.  Previous studies on service-learning have 
reported gains in conceptually-related outcomes.  Much research has been done to 
determine the nature of wise persons (Clayton & Birren, 1980; Holliday & Chandler, 
1986), and to develop measurements that accurately assess wisdom (Ardelt, 2003; Baltes 
& Smith, 1990; Sternberg, 1998; Webster, 2003).  Additionally, many researchers have 
identified experience and reflection as two central vehicles for the cultivation of wisdom 
(Ardelt, 2004; Assmann, 1994; Webster, 2003).  To date, there are no published studies 
that have tested the methods of experiential learning as a facilitator of growth in wisdom. 
The proposed research also has practical implications for the field of experiential 
education.  Theorists have contended that the methods of experiential education are 
useful for developing more than encyclopedic knowledge (Dewey, 1916; Freire, 1993; 
Kolb, 1984).  Unfortunately the process of experiential learning has proved difficult to 
verify.  This study will provide information that connects the components in a service-
learning environment to their associated outcomes.  In this way, future programming can 
be assessed and modified to provide experiences that are more effective for long-term 
growth.   
 
16 
 
Definition of Terms 
1. Affection: The presence of positive feelings (i.e. sympathy and compassion) and 
the absence of negative emotions and behavior towards others (Ardelt, 2003). 
2. Cognition:  A person’s ability and desire to understand life, its significance and 
deeper meaning, particularly with regard to intra- and interpersonal matters 
(Ardelt, 2003). 
3. Reflection:  The consideration of life events from a variety of perspectives with 
the intent of gaining insight into the true nature of things.  Reflection develops 
self-awareness and simultaneously increases a person’s cognition and affection 
(Ardelt, 2003). 
4. Wisdom:  A combination of personal characteristics and dispositions that is 
considered optimal for individual and social well-being.  These are developed 
through a combination of experience and reflection, and are evidenced by the 
presence of affective, cognitive, and reflective indicators (Ardelt, 2003). 
5. Service-Learning:  A form of experiential education in which students engage in 
activities that address human and community needs together with structured 
opportunities intentionally designed to promote student learning and development 
(Jacoby, 1996, p. 5). 
Limitations of the Study 
 There are several limitations that should be addressed at the onset of this study.  
First, the program participants were a self-selected group that cannot be considered as 
representative of the entire student population.  Thus, the results of this study cannot 
be indiscriminately generalized beyond the scope of late adolescents/ young adults 
17 
 
attending a Midwestern college or university.  The sample was relatively homogenous 
in regards to race and socio-economic status.  In addition, these students could be 
considered highly motivated and pro-socially oriented, since they were paying to 
participate in a service-learning trip over Spring Break. 
 While this study included a follow-up measurement that provides some 
longitudinal support, a 1-month follow-up assessment may not be sufficiently long-
term to accurately assess one’s growth toward wisdom.  As a complex measure of 
human development, it may be more appropriate to measure these participants over 
the course of several years.  This is especially true given that reflection is considered 
the “hub” of wisdom for this research.  A follow-up scheduled several months or 
years after the experience may provide a better assessment of the true impact of the 
experience, as the participants would have more time to process the trip.  
Unfortunately, such methods are beyond the scope of this study. 
 While this study did include a control group, the participants were not randomized 
and the control group was non-equivalent.  Indeed, the control group differed from 
the treatment group on most measures of interest before the treatment was in effect.  
A variety of analyses were conducted using a variety of sampling methods, showing 
little variation in outcomes.  In addition, the SEM structure was consistent for both 
the treatment and control groups, indicating that, while the mean scores for outcome 
variables differed, the structure of relationships between the variables in the study 
was consistent across groups.  Finally, the effect sizes for the treatment group remain 
significant even when accounting for the disparity in pre-trip measures.  Future 
researchers may prefer to use a purposive sample that equates more closely with the 
18 
 
service-oriented disposition of service-learning participants (i.e. social work 
students). 
Finally, wisdom is still relatively young as a psychological construct.  The 
construct is still being defined, and any claim regarding the increase of wisdom will 
necessarily depend upon the developments of future research.  In addition, the use of 
self-report measures for wisdom is not universally accepted (Baltes, 2004).  As with 
any self-report measure, one must assume a certain level of self-awareness and 
honesty in responses. 
Outline for Dissertation Proposal 
 This proposal presents relevant information regarding the cultivation of wisdom 
through experiential learning.  Chapter I provides a brief background of the variables, 
as well as the rationale for and significance of this study.  Chapter II provides a 
review of associated literature that pertains to experiential learning and wisdom.  
Chapter III outlines the proposed research design and methodology.  The results of 
quantitative statistical analyses are presented in Chapter IV and a discussion of these 
results is presented in Chapter V. 
 
 
19 
 
CHAPTER II 
REVIEW OF THE LITERATURE 
This chapter provides a summary of the literature that is relevant to the 
independent and dependant variables of this study.  The first section details the literature 
that applies to the independent variable, experiential learning.  After a general discussion 
of experiential learning, the literature that pertains specifically to service-learning is 
reported.  The next section consists of the literature relevant to the dependant variable, 
wisdom.  This section presents the main approaches used by wisdom researchers and 
reports the findings that support the rationale for this study.  The final section makes the 
connection between experiential learning and wisdom, as well as providing a summary of 
the issues most pertinent to this study. 
Experiential Learning 
 Experiential educators consider life (or life experiences) to be the basis for any 
knowledge (Jarvis, Halford & Griffin, 1998).  While this statement is not revolutionary, 
its implications for learning certainly can be.  Educational theorists such as John Dewey 
and Paulo Freire emphasized that experience-based education is vital to inspire critical 
thinking, to maintain individual freedom, and to insure that learning becomes a life-long 
process (Dyke, 2006).  Experiential educators utilize concrete experience and reflection 
to help an individual transform his/her knowledge system, and to indirectly transform 
society.  After a brief discussion of the difference between passive and experiential forms 
of education, this section will detail previous research regarding service-learning and 
adventure trips to demonstrate how wisdom can be best cultivated through experiential 
education. 
20 
 
 Traditional styles of passive education have been equated with ineffective and 
authoritarian styles of teaching, also referred to as “banking knowledge” and “rote 
learning” (Dyke, 2006; Moacir, 1994).  This method of education views the learner as a 
passive, empty vessel that is filled by a knowledge expert.  The learner is not encouraged 
to think critically, but is encouraged to memorize given information.  Students are not 
actively involved in the learning process, and instead of acquiring the skills that will 
enable them to continue learning throughout their lifetime, they are taught to be 
dependant on the knowledge of authorities (Jarvis, Halford & Griffin, 1998).  These 
students often leave with an inflexible set of rules and “descriptive knowledge” (Kekes, 
1995) that may help them to get a job, but won’t always help them to negotiate difficult 
life problems that require a wise interpretation of that knowledge.  Thus, passive 
education can improve knowledge (cognition) of a particular subject matter, but it may 
not cultivate multi-dimensional, wise individuals. 
 Conversely, experiential learning places the emphasis on the student, 
downplaying the role of the authority.  Students are encouraged to take an active role in 
their own learning, thereby increasing their motivation to learn (Jarvis, Halford & Griffin, 
1998).  The teacher is viewed as a facilitator who guides the student through the process 
of learning.  Thus, the emphasis is placed on the process of learning as well as the 
content.  This student can then continue the process of learning throughout his/her 
lifetime by applying this process to new experiences.  In addition to relaying specific 
knowledge (cognition), experiential education also teaches the practice of reflection 
(Dyke, 2006), increases affection through personal contact (Rhoads, 1998), and provides 
students with the skills necessary to re-discover ancient concepts and socialize them to 
21 
 
their cultural context (Moacir, 1994).  This re-discovery and modern application of 
ancient truths is a central theme in the concept of wisdom (Ardelt, 2004). 
Since reflection plays such an important role in experiential learning, a clearer 
definition seems appropriate.  The term “reflection” is often used very loosely in a variety 
of contexts.  From a strictly scientific perspective, reflection may be as simple as 
recording observed data and using that data to hypothesize about a future experiment.  
Takahashi (2002) called this reflection on sensory material “reflective thinking” in order 
to distinguish it from the deeper type of reflection which leads to self-awareness and 
wisdom.  While “reflective thinking” promotes knowledge of how empirical objects 
interact, “reflective understanding” promotes personal meaning through dwelling on the 
cultural, historical, and personal significance of experience.  Other theorists attempted to 
define this deep reflection with terms such as “emancipatory learning” (Mezirow, 1990), 
“secondary reflection” (Marcel, 1954), and “second-loop learning” (Jarvis, Halford & 
Griffin, 1998).  Deep reflection requires one to stand apart from their ego (Marcel, 1954) 
in order to view things from a different perspective.  This separation from the ego then 
reveals fallacies in one’s belief system and allow them to return to the concrete situation 
with a renewed understanding.  Wisdom develops as a result of this continual effort of 
detached concern (Csikszentmihalyi & Rathunde , 1990)) and of establishing connections 
and making meanings where none are readily apparent (Helson & Srivastava, 2002).   
Service-Learning 
Service Learning has been defined as “a form of experiential education in which 
students engage in activities that address human and community needs together with 
structured opportunities intentionally designed to promote student learning and 
22 
 
development (Jacoby, 1996, p. 5).”  It has also been referred to as “citizenship 
education,” a term that emphasizes a concern for the social good (Delve, Mintz, & 
Stewart, 1990).  As such, it is concerned with the personal development of the individual 
participants as well as the communities in which they serve.  The inclusion of service-
learning in schools has increased drastically in the past two decades.  A 1999 survey 
found that 83 percent of high schools reported opportunities for community service 
activities, as compared to a mere 27 percent in 1984 (Spring, Dietz, & Grimm, 2006).  
More than one-third (36%) of youth polled had participated in community service as part 
of a school activity or requirement.  Reported outcomes for service-learning participants 
include increased efficacy and civic engagement (Spring, Dietz, & Grimm, 2006), 
improved academic performance and behavior (Frederickson, 2000; Lundy, 2007; Scales, 
et. al, 2006) and increases in empathy, cognition, self-concept, and social development 
(Berger & Milem, 2002; Lundy, 2007; Waldstein & Reiher, 2001).  This section will 
detail underlying theories of service-learning and present previous research relevant to 
the current study. 
The Service-Learning Model   
As a form of experiential learning, service-learning models fit well into the scope 
of Experiential Learning Theory (ELT).  Figure 2.1 illustrates one attempt to converge 
the elements of service-learning with Kolb’s (1984) ELT model.  The service-learning 
model consists of five phases that participants proceed through due to involvement in 
community service.  These phases include: exploration, clarification, realization, 
activation, and internalization.  The learner’s progress through these stages is mediated 
by their type of involvement (i.e. direct or indirect), their length of time involved, the 
23 
 
needs and outcomes of the participants, and the balance of challenge and support offered 
by the experience. 
 
 
Figure 2.1. Service-Learning Model with Kolb’s Experiential Learning Model 
(Delve, Mintz, & Stewart, 1990, p.37). 
One factor that service-learning contributes to the model is the emphasis of 
values.  In fact Delve, Mintz, and Stewart (1990) assert that “…involvement in 
community service is symbiotic with values development (p. 8).”  Their theory is based 
24 
 
on a combination of Perry’s (1970) Cognitive Development Model, Kohlberg’s (1975) 
Moral Development Model, and Gilligan’s (1982) Model of the Development of 
Women’s Moral Judgment.  As the learner progresses through the five phases of the 
service-learning model, they also progress through the stages of development as 
expressed by Perry, Kohlberg, and Gilligan.  Thus, according to Perry’s model, the 
participant would begin with a simple dualistic mindset that views things as either “black 
or white.”  As they progress through the service-learning model, the participants begin to 
see “multiplicity” in issues, understanding that there may be more than one right answer.  
Finally, the learner grows into the understanding that knowledge must be viewed within 
its own context, a mature stage referred to as “relativism.”  Similarly, the learner 
proceeds through Kohlberg’s stages of moral development beginning with a moral stance 
of avoiding punishment and ending with a respect for the dignity of all individuals.  
Gilligan’s model emphasizes the difference in moral development between genders, and 
the need to consider different intervention methods for each.  In addition, Gilligan 
legitimizes the concept of caring, a concept that is central to service and to wisdom.   
 The Service-Learning Model illustrates how one’s thought processes can grow 
from simple, concrete functioning toward a more complex and integrated style through 
the methods of service-learning.  Though the principles in this model support the 
framework for this study, Kolb’s (1984) model was deemed more appropriate for this 
research.  The service-learning model represents a longitudinal measure of development 
that goes beyond the scope of the current study.  As such, it is perhaps best applied to 
programs that are offered through schools and engage the learner at small increments 
over the course of a semester.  The program utilized in this study incorporates elements 
25 
 
of the service-learning model along with elements in adventure or expeditionary learning 
models to provide an intense, but short-term experience.   However, an understanding of 
the role of service-learning in values development is imperative when measuring a value-
laden concept such as wisdom. 
Service-Learning Literature   
Research on service-learning is vast and plentiful.  Earlier studies were plagued 
with inadequate methodology, most of which reported outcomes anecdotally.  A recent 
thrust of literature has focused on legitimizing the field using established measurements 
as well as experimental and quasi-experimental designs.  The following studies provide a 
sample of the recent quantitative, longitudinal, and qualitative studies and their reported 
participant outcomes. 
 Schmidt, Shumow, and Kackar (2007) analyzed data collected from the National 
Household Education Survey to determine the effects of service-learning on academic 
adjustment, behavior problems, civic efficacy, and civic knowledge.  This study was 
designed to confirm the findings of previous research with the use of a large, nationally 
representative data set.  The sample consisted of 4,306 high school respondents who 
attended public and private schools.  They assigned the respondents to groups according 
to the number of hours spent in service activities, whether their service was voluntary or 
required, and the type of service they were involved in (i.e. direct contact, indirect, and 
service to environment or animals).   
 Findings suggest that participation in any type of service was associated with 
significant increases in grades and civic knowledge, and a significant decrease in 
behavior problems.  There was no difference in outcomes according to hours of service.  
26 
 
Additionally, the outcomes were consistent for those who volunteered to participate and 
those who engaged in mandatory service.  Those who had direct contact with the 
individuals being served tended to have higher grades, while those involved in indirect 
service to organizations reported more civic efficacy and civic knowledge (Schmidt, 
Shumow, & Kackar, 2007).  These findings provide strong support for the personal and 
civic outcomes attributed to service-learning.  The results also imply that different 
activities may lead to different outcomes.  While this study provides a good general idea 
of the outcomes associated with service-learning, its scope is too broad to understand the 
true nature of a service-learning experience.  In addition, this type of survey research 
cannot establish causality or duration of outcomes due to the lack of a control group and 
longitudinal measures. 
 Astin, Sax, and Avalos (1999) set out to further legitimize the outcomes of 
service-learning experiences with the use of longitudinal, multi-institutional data.  The 
researchers utilized information from surveys completed in 1985 by 279, 985 college 
freshmen at 546 colleges and universities.  The Student Information Form (SIF) survey 
included info on personal and demographic characteristics, high school experiences, 
expectations about college, values, goals, self-concepts, and career aspirations.  A follow-
up survey was completed by the same students in 1989, to measure their perceptions of 
the college experience and to serve as a post-test for the information gathered on the 
original survey.  Finally, a nine-year follow-up survey was completed by a sample of 
12,376 students who had completed the previous two surveys.   
 Significant behavioral outcomes associated with service included: attending 
graduate school, earning higher degrees, donating money to one’s alma mater, socializing 
27 
 
with others of different race, and participating in community work after college.  In the 
realm of values, undergraduate service is associated with helping others in difficulty, 
participating in community action programs, participating in environmental cleanup 
programs, promoting racial understanding,  and developing a meaningful philosophy of 
life (Astin, Sax, & Avalos, 1999).  These findings validate the long-term impact of 
service-learning in both the cognitive and affective domains.  Furthermore, the effects 
remained significant even when controlling for service after college, indicating that 
service in the late adolescent/ early adult years can be a powerful source of long-term 
development.   
 While the two previous studies provide evidence of a relationship between service 
and positive development, the lack of a control group presents problems for identifying 
causality.  Batchelder and Root (1994) conducted a quasi-experimental study of 96 
students enrolled in small, mid-western liberal arts colleges.  Forty-eight of these students 
were enrolled in courses that incorporated service-learning into the curriculum.  The 
other 48 students were enrolled in courses with similar content that did not include 
service-learning activities.  Measurements included pre and post responses to social 
situations, journal analysis, and an evaluation of the service-learning experience.   
 Those students involved in service activities demonstrated a greater resolve to act 
in the face of uncertainty and a greater awareness of the multiple dimensions and 
variability involved in dealing with social problems, than their counterparts in the control 
group (Batchelder & Root, 1994).  These findings are extremely significant due to their 
relevance to the topic wisdom.  A tolerance for ambiguity and an awareness of the 
“thorniness” of life situations are important characteristics of people nominated as wise 
28 
 
(Batles, Gluck, & Kunzmann, 2005).  The fact that those who participated in service-
learning grew in this respect supports the idea that wisdom is gained from experience 
with real-life situations, and not from knowledge acquisition alone.  In addition, those 
involved in the service-learning component demonstrated significant increases in 
Prosocial Decision-Making, Prosocial Reasoning, Self-Reflective Empathetic Reasoning, 
and Internalized Reasoning.  Empathy and prosocial values have been associated with 
wisdom in previous research, and an increase in these measures could be associated with 
a growth in wisdom domains (Batles, Gluck, & Kunzmann, 2005; Bailey & Russell, 
2008). 
 A similar study was conducted by Lundy (2007), who assessed the exam scores 
and empathy ratings of 192 college students enrolled in a developmental psychology 
course.  Students either completed a service-learning project, an interview project, or a 
research paper as their major class assignment.  Analysis revealed that, while there were 
no significant differences for the first exam scores, those involved in the service-learning 
project scored significantly higher than those in the other two groups on exams two, three 
and four.  In addition, those students in the service-learning group showed a significantly 
greater change in empathy scores than students in the other two groups.  In fact, empathy 
scores for the interview and research paper groups decreased over the course of the 
semester.  Once again, evidence suggests that service-learning makes a significant 
contribution to cognitive and affective development, two core domains of wisdom.   
The Learning Process   
While all of the above studies provide support for the positive impacts of service-
learning, they contribute little to the understanding of the learning process.  In order to 
29 
 
replicate the positive results reported by these studies, one must identify the key elements 
involved in the process and determine how they interact.  Since service-learning can be 
performed in a variety of ways, it may be necessary to evaluate these elements within the 
unique context of each program.  A few notable studies have attempted to unravel the 
service-learning process, and this section will present some of the key findings from that 
research. 
 Rhoads (1998) conducted a phenomenological study in an attempt to dig deeper 
into the participants’ experiences during service-learning activities.  His contention was 
that “fostering a deep commitment to caring is the postmodern developmental dilemma 
all of education faces, including higher education.  If we are to promote democratic 
citizenship in these challenging times, then we must foster in our citizens a commitment 
to caring (Rhoads, 1998, p. 6).”  Over the course of six years, Rhoads utilized naturalistic 
inquiry to collect data in the form of surveys, interviews, observations and journal 
analysis.  Data were collected from participants in a variety of programs, including one-
week intensive trips and long-term activities that were woven into school curriculum.   
The findings were categorized into three interactive themes: Self- Exploration, 
Understanding Others, and The Social Good (Rhoads, 1998).  Self-Exploration included 
elements of identity clarification, going “outside of oneself” to connect with others, and 
self-interrogation.  Service-learning, then, can help students to critically examine 
themselves, and to understand their place in the scheme of things.  The second theme, 
Understanding Others, included topics such as: placing faces with the statistics (about 
homeless people, etc.), overcoming stereotypes and generalizations, and being sensitized 
to cultural diversity.  The final theme, Social Good, consists of increased awareness of 
30 
 
social responsibilities, of the complexity of social problems (i.e. inequality), and of the 
inefficiency of piecemeal service programs.  These themes arose from being personally 
involved in service to others and from the opportunity to reflect on the service through 
journaling and discussions.  Self-awareness, compassion and a concern for the common 
good are all characteristics associated with wise persons (Baltes, Gluck, & Kunzmann, 
2005).  Thus, the service-learning process may be instrumental in the development of 
wisdom.  
Another study conducted by Lakin and Mahoney (2006) included 60 urban sixth-
grade students who participated in a mandatory service-learning project.  Forty of the 
students (two classes) served as a treatment group, while another 20 students (one class) 
served as a control group.  This study attempted to measure the outcomes associated with 
the service-learning experience as well as assessing two factors that were assumed to 
influence the process.  Outcome measures included prosocial attitudes and self-efficacy.  
It was hypothesized that these outcomes would be mediated by empowerment and a sense 
of community in the learning process.  Empowerment was defined as “gaining mastery 
over one’s life” and a sense of community included “a sense of belonging and influence 
in a group.”   Empowerment was fostered by allowing youth to be the primary decision-
makers, with adults serving as facilitators who asked guiding questions and provided 
assistance when needed.  Democratic decision-making processes, leadership discussions, 
and student-led subcommittees were used to promote a sense of community. 
Results indicated that the program was successful in providing an empowering 
environment that fostered a sense of community (Lakin & Mahoney, 2006).  Qualitative 
data included comments regarding the increase in students’ beliefs that they can make a 
31 
 
difference, and support for the idea that group decision making increases self-efficacy 
and connectedness to one another.  In addition, the process experience explained 33% of 
the variance in the students’ intent to be involved in future community action, and 29% of 
the variance in post-test scores on self-efficacy.  This study provides a first-look at the 
influential processes involved in community service.  Programs that involve opportunities 
for youth leadership and decision-making may be more effective at promoting self-
efficacy and group cohesiveness.  Additional research needs to be done to further explain 
the processes involved in service-learning activities.  Also, since the program in Lakin 
and Mahoney’s (2006) study was a mandatory, semester long project, further research 
involving various program models and participants would help validate these findings. 
Adventure Education   
The variety of experiential program models is somewhat problematic when trying 
to understand the underlying process.  The program being researched in the current study 
includes many of the service-learning elements described in the above studies (Lakin & 
Mahoney, 2006; Rhoads, 1998), but the program design is more akin to that of adventure 
education models, such as Outward Bound.  The Walsh and Golins (1976) model (Figure 
2.2) illustrates some of these similarities.  In this model the learner is placed into a unique 
physical and social environment and given a set of problem-solving tasks to overcome.  
These tasks place the learner in a state of dissonance that must be overcome with some 
measure of adaptation.  This idea of “adaptive dissonance” comes from earlier 
educational theorists such as Dewey, Lewin, and Piaget, and also provides the foundation 
for Kolb’s (1984) Experiential Learning Theory.  With the successful negotiation of this 
dissonance, the learner acquires a sense of mastery that influences their future 
32 
 
experiences through increases in self-esteem, self-awareness, and self-efficacy, among 
other things.   
  
Figure 2.2. Model of the Outward Bound Process (McKenzie, 2003, p.9). 
 
Certain elements of the Walsh and Golins (1976) model are clearly comparable to 
elements described in the service-learning literature.  The problem-solving tasks, for 
example, resemble the challenges as described by Delve, Stewart and Mintz (1990).  
Mastery was one of the key process factors identified by Lakin and Mahoney (2006), 
though they referred to it as “empowerment.”  Also, the social environment includes 
elements such as group autonomy (i.e. ownership and decision-making), conflict 
resolution, and reciprocity, all of which have been identified as important factors in 
service-learning environments (Lakin & Mahoney, 2006, Rhoads, 1998).  One factor that 
is not present in the service-learning literature is the novel physical environment.  While 
33 
 
this may not play a central role in school-based service programs, it could certainly 
influence the process of an intense service-learning trip.  Long hours on a bus and nights 
spent on the floors of school gymnasiums, for example, could certainly influence the 
group dynamics and overall outcomes of a service-learning experience. 
 Adventure programs suffer from the same lack of process-knowledge as do 
service-learning programs (Hattie, et. al, 1996).  The Walsh & Golins model served as the 
foundation for Outward Bound trips for almost 30 years before it was empirically tested 
(Sibthorp, 2003).  Recent research has focused on defining this model.  A study 
conducted by McKenzie (2003) is especially relevant to the current proposal. 
 McKenzie (2003) collected data from 92 students who were involved in an 
Outward Bound experience.  This study was designed to test the well-known Walsh & 
Golins Model (1976) and to identify key elements in the learning process.  Through a 
series of questionnaires, interviews and observations, McKenzie (2003) found 29 course 
components that influenced program outcomes.  These components were then ranked as 
to their importance regarding the major categorical outcomes associated with Outward 
Bound experiences.  The elements considered most important for the outcome of Self-
concept included: “achieving individual success,” “being physically challenged,” “taking 
responsibility for yourself.”  The components that most influenced Motivation included: 
“achieving individual success,” “learning new skills,” “instructor’s personalities,” 
“instructors as role models,” and “being challenged physically.”  Finally, the components 
that most affected Interpersonal skills were: “problem solving as a group,” “achieving 
group success,” “having leadership responsibilities,” “relationships with other group 
members,” and “instructor’s personalities.” 
34 
 
 These findings indicate that different components of the trip have a different 
influence on each program outcome.  This not only has implications for the modification 
of the Outward Bound model, but it helps make a connection between each of the 
program components and the reported outcomes.  Much more research is needed to 
further decipher the process of similar program models.  While the service-learning 
process is not identical to the Outward Bound process, they do share many program 
components, and it is conceivable that they could overlap considerably.   
Wisdom 
 Much research has been done to demystify the concept of wisdom.  Early research 
focused on implicit theories of wisdom in an effort to understand its common meaning 
(Clayton & Birren, 1980; Holliday & Chandler, 1986).   These exploratory studies 
provided the basis for an increasingly large body of literature, including a collaborative 
volume that consisted of 12 chapters in which wisdom was defined 12 different ways 
(Sternberg, 1990).  Having a solid body of foundational research, explicit theorists then 
began to explore the potential of measuring wisdom and wisdom-related performance.  
Baltes and Smith (1990) developed a complex qualitative method of measuring “wisdom-
related knowledge,” while others constructed self-report wisdom scales, designed to 
measure personal characteristics that represent the wisdom as a latent variable (Ardelt, 
2003; Brown, 2004; Webster, 2003).  Since there is no universally accepted definition of 
wisdom, there is also no universal measurement.  Research design is highly debated 
among wisdom researchers (Ardelt, 2004; Baltes, 2004).  This section will describe the 
three main categories of wisdom research, reporting the major foundational studies for 
each. 
35 
 
Implicit Research   
Implicit research has helped to define the concept of wisdom through analysis of 
historical texts and through qualitative methods that establish commonly held views 
about the characteristics which wise people have (Assmann, 1994; Clayton, 1980; 
Sternberg, 1985; Holliday & Chandler, 1990).  This foundational knowledge is 
imperative for any understanding of the concept.  These studies have determined that the 
conceptualization of wisdom is affected by social, environmental and political context, 
and by a person’s values and beliefs (Baltes, Gluck, & Kunzmann, 2005).   
 The first implicit research study on wisdom was conducted by Clayton and Birren 
(1980).  The participants in the study included 83 adults who were divided into three 
categories: 31 were considered “young,” 23 were “middle-aged,” and 29 were “old.”  
Utilizing a list of 12 wisdom descriptors that were generated from earlier research, 
Clayton and Birren had these participants rate the similarity of all possible word pairs.  
The descriptors included “experienced,” “intuitive,” “introspective,” “pragmatic,” 
“understanding,” “gentle,” “empathetic,” “intelligent,” “peaceful,” “knowledgeable,” 
“sense of humor,” and “observant.”  Multidimensional scaling analysis was conducted to 
determine the similarities between the pairs.  The analysis yielded 3 distinct dimensions, 
which were labeled “cognitive” (i.e. knowledgeable, experienced, intelligent), “affective” 
(i.e. understanding, empathetic, peaceful), and “reflective” (i.e. introspective and 
intuitive). 
 A similar study was conducted by Holliday and Chandler (1986) in 2 parts.  In the 
first study 150 adults from three different age categories (50 young, 50 middle-aged, 50 
old) were asked to describe six different types of people: wise, shrewd, perceptive, 
36 
 
intelligent, spiritual, and foolish.  Qualitative analysis revealed 79 distinct descriptors 
which were referred to as “wisdom attributes.”  A list of these 79 attributes was then 
given to another 150 adults in the same age categories who were asked to rate each of the 
attributes using a Likert scale from 1 (almost never true of wise people) to 7 (almost 
always true of wise people).  Factor analysis of the descriptors identified 5 underlying 
dimensions of wisdom.  These wisdom dimensions included: “exceptional 
understanding,” “judgment and communication skills,” “general competencies,” 
“interpersonal skills,” and “social unobtrusiveness.”  Ardelt (2003) contends that closer 
examination reveals the presence of cognitive, affective, and reflective elements within 
the five factors proposed by Holliday and Chandler (1986).  Thus, there does seem to be 
relative agreement on the elements of wisdom, though those elements are categorized 
differently by researchers. 
Wisdom-Related Performance (WRP)  
 Researchers from the Max Planck Institute in Berlin developed their own unique 
approach to measuring wisdom.  Rather than relying on folk-concepts of wisdom, Smith 
and Baltes (1990) developed their theory based on their own expertise in human 
development.  They define wisdom as “expert knowledge involving good judgment and 
advice about important but uncertain matters of life (p.95).”  Thus, wisdom is evidenced 
by the existence of five essential criteria: rich factual knowledge, rich procedural 
knowledge, life span contextualism, relativism, and uncertainty.  To assess wisdom-
related performance, they present scenarios to participants that represent morally difficult 
situations or life planning problems.  For instance, subjects may be presented with the 
story of Joyce, a 60-year old widow who is debating whether or not to give up her 
37 
 
business and move closer to her son, in order to help care for his two small children.  
Participants are then asked to think out loud as they contemplate the dilemma.  Two 
trained judges rate the responses based on the five established wisdom criteria.  Those 
responses that include more of the criteria are considered wiser.  Figure 2.3 illustrates the 
framework of wisdom as developed by the researchers at the Max Planck institute. 
 
Figure 2.3. Baltes Theoretical Framework of Wisdom (Baltes, 2003) 
 
 In one study utilizing the above methods, Staudinger and Baltes (1998) measured 
the wisdom-performance of 36 psychologists, with 54 other highly-educated 
professionals serving as a control group.  The premise if this study was to determine if 
certain types of life experience may facilitate the development of wisdom-performance.  
The participants, ranging in age from 25 to 82 years, responded to two morally 
challenging scenarios and completed a battery of personality and intelligence measures.   
38 
 
 Analysis indicated that practice as a clinical psychologist accounted for the largest 
amount of unique variance (26%) in wisdom scores.  Personality measures were stronger 
predictors of wisdom performance than were intelligence variables.  Professional 
specialization and personality scores shared 9% of the variance.  The strongest 
personality predictors were Openness to Experience and a middle range location on the 
Introversion-Extroversion dimension.  These findings highlight professional 
specialization and personality measures as important variables in the cultivation of 
wisdom.  Furthermore, the authors contend that the reason psychologists scored higher on 
wisdom performance was because they were routinely exposed to difficult and “thorny” 
life problems.  Others (Ardelt,2004; Webster, 2003) speculate that clinical psychologists 
merely score higher on measurements developed by other psychologists.   
 In another study conducted by Staudinger and Baltes (1996), 244 participants 
completed wisdom tasks to investigate the impact of collaboration on wisdom-
performance.  The participants consisted of 122 middle-aged participants (61 women and 
61 men) together with a person of their choice with whom they felt comfortable 
discussing life problems.  The first interview session was conducted using the same 
methods as previous research (described above).  During the second session, the dyads 
were assigned to one of five experimental conditions.  Three of these conditions involved 
only one participant in each dyad while the other two conditions allowed for 
collaboration.  
The “external dialogue plus individual appraisal,” and the “internal dialogue” 
conditions resulted in significantly higher performance levels than the other three 
conditions.  The first condition included dialogue along with additional time for personal 
39 
 
reflection before responding.  The latter condition involved one participant in an 
“imaginary” dialogue with a person they considered to be wise.  These results suggest 
that real or “virtual” collaboration can improve wisdom-related performance.  Group 
collaboration is a central component in programs that utilize experiential learning, and 
reflective imagery, a form of “virtual” collaboration, is often encouraged through 
metaphors and guided reflection. 
 Another study which utilized the methods proposed by the Max Planck Institute 
was conducted to observe the differences in wisdom performance as evidenced in 
different age groups (Staudinger & Pasupathi, 2001).  For this study, 14 heterogeneous 
adolescents (ages 14 to 20) were compared to a sample of 58 young adults (ages 21 to 
37).  Again, participants responded out loud to difficult, ill-defined life dilemmas and 
were rated by two independent judges along the five wisdom criteria.  The results 
indicate that age is positively related to wisdom-performance in adolescents.   This 
relationship changes, however, beginning at age 23.  For adults 23 years and over, there 
is no significant relationship between age and wisdom-performance.  Further analyses 
indicated that all respondents scored higher on factual knowledge than procedural 
knowledge, with this being especially true of adolescents (Staudinger & Pasupathi, 2001).   
Finally, adolescents scored higher on tolerance of uncertainty and lower on value 
relativism, with adults showing the opposite trend.   
These results indicate that, though age may be a strong predictor of wisdom in 
adolescents, the effects of the age variable lose their potency in the mid-twenties.  In 
addition, wisdom may be constructed differently according to age.  For example, while 
the adults showed consistent scores across wisdom tasks, the adolescent scores were 
40 
 
higher when the tasks were more relevant to their age context.  These findings reveal 
another level of complexity within the wisdom-construct that will require further research 
to unravel.   
 Though much research has been conducted by the researchers associated with the 
Max Planck Institute, there are questions as to the methods used and whether they are an 
appropriate assessment of wisdom.  First, their definition of wisdom places it very close 
to the construct of Crystallized Intelligence (Catell, 1977).  This may indicate that these 
researchers are measuring cognition alone to the neglect of other important wisdom 
domains (i.e. reflection and affection).  In addition, Ardelt (2004) points to the findings of 
the Staudinger and Baltes (1998) study reported above to suggest that personality plays a 
more significant role than intelligence in the construct of wisdom.  In fact, it is likely that 
one’s personality would serve as a confound for professional specialization, since a 
general disposition toward service to others may lead one into the profession of clinical 
psychology.  Furthermore, since the scenarios are generic, and there is no associated 
affection for the respondent, the responses given cannot be considered a measure of how 
the individual would truly respond or act if confronted with the scenario in real life 
(Ardelt, 2004; Webster, 2003).  Finally, the procedures used by the Max Planck 
researchers are inefficient when trying to measure larger sample sizes.  For these reasons, 
a number of researchers have begun the development of self-report wisdom 
questionnaires. 
Wisdom Instruments  
In order to measure wisdom with the use of questionnaires, the wisdom construct 
is considered to be a latent variable that it evidenced by the effects of sub-domains.  
41 
 
Ardelt (2000) conducted the first longitudinal study of the wisdom construct, utilizing 
latent factor techniques to determine wisdom in old age.  The original data were collected 
40 years earlier by trained coders in extensive interviews.  Of the 248 women who 
participated in the original 1968 study, 82 were able and willing to participate in Ardelt’s 
follow-up assessment.  Wisdom in old age was operationalized as a latent variable with 
cognitive, affective and reflective effect indicators, based on Clayton’s (1986) original 
implicit study.     
Though internal reliability levels were relatively low (some even below .5), the 
study reported some illustrative findings.  First, the respondent’s wisdom in old age was 
positively affected by their social environment in early adulthood, while all other 
variables were unrelated to wisdom in later years.  This highlights the age of early 
adulthood as a vital time for the cultivation of wisdom, regardless of objective 
circumstances or personality characteristics during childhood.  Only wisdom and 
financial situation had a positive significant effect on subjective well-being in old age 
when all other variables were held constant.  In addition, wisdom had a positive influence 
on physical health and on the quality of women’s family relationships.  Thus, the 
cultivation of wisdom, beginning in early adulthood, can lead to existential well-being in 
old age: a picture akin to the eighth stage of human development portrayed by Erikson 
(1982). 
Finally, Webster (2003) developed a completely independent Self-Assessed 
Wisdom Scale (SAWS), based on a five-factor conceptualization of wisdom that is linked 
to Erikson’s theory of generativity and the concept of ego integrity.  The SAWS consists 
of 30 items on a 6-point Likert scale that measures experience, emotions, reminiscence, 
42 
 
openness, and humor.  Webster conducted 3 initial studies to establish the validity and 
reliability of the SAWS.  Total scale reliability was assessed with a sample of 179 
Canadian men and women.  The instrument was found to be reliable (alpha= .78) but the 
domains of openness and humor overlapped, resulting in low factor loadings for those 
domains.   
In the final study 39 men and 46 women completed three instruments to measure 
wisdom (SAWS), generativity (Loyola Generativity Scale), and ego- integrity (Taft and 
Nehrke, 1990).  Generativity and ego-integrity are two constructs that were previously 
associated with wisdom as conceived by Erikson (1982).  The SAWS was found to be 
positively correlated to both generativity and ego-integrity, supporting its value as a 
measurement of Erikson’s eighth stage of development.  Concerns about Webster’s scale 
include a lack of expert analysis in development of the scale items, a lack of support for 
the inclusion of humor as an aspect of wisdom, and Webster’s heavy reliance on the 
relatively weak correlations of the SAWS to generativity and ego-integrity as evidence of 
construct validity. 
Wisdom and Experiential Learning 
The construct of wisdom shares many elements that are fostered through 
experiential education.  Both emphasize the need for first-hand experience.  While a list 
of facts can be memorized using a variety of methods, the development of wisdom 
requires a broad range of life experience.  This necessity of experience is the main reason 
that wisdom is typically associated with the elderly (Assmann, 1994).  Furthermore, some 
authors contend that wisdom comes, not from a vast array of general experience, but 
specifically from those difficult life experiences that allow for some degree of moral 
43 
 
profundity (Webster, 2003).  These difficult experiences require the individual to 
confront the complexities of life and wrestle with the underlying issues, thereby 
developing mastery by overcoming this uncomfortable dissonance (Ardelt, 2003; Kolb, 
1984; Walsh and Golins, 1976).  Service-learning programs are particularly effective at 
generating this dissonance, as they encourage participants to “go beyond themselves,” to 
view life from different perspectives, and to consider how their actions affect the “social 
good” (Delve, Mintz & Stewart, 1990; Lakin & Mahoney, 2006; Rhoads, 1998),   
Experience is a necessary, but not sufficient element in the development of 
wisdom.  Another vital component of wisdom development, and of experiential learning, 
is reflection.  “Reflective thinking simultaneously fosters a deeper understanding of life 
and human nature and the development of sympathy and compassion for others (Ardelt, 
2003, p. 300).”  In other words, an individual’s will to act (volition) is directed by his/her 
knowledge (cognition) and “feelings” about others (affection) through the mediation of 
reflection.  A lack of reflection not only inhibits one from understanding life on a deeper 
level, but it can reduce one’s ability to control urges (James, 1896; Pascual-Leone, 1990) 
that lead to impulsive behavior.  Furthermore, the practice of reflective methods can 
generate habits that may enable one to make wiser choices in future situations (Baltes & 
Staudinger, 2000). 
The key to any model of experiential learning is reflection.  Reflecting on the 
experience is what enables an individual to truly learn, to transform himself and to 
transform society (Dewey, 1916; Dyke, 2006; Moacir, 1994; Jarvis, Halford & Griffin, 
1998).    Reflection allows the learner to interpret their new knowledge based upon their 
own beliefs and life context.  They must then decide if their current system of knowledge 
44 
 
and beliefs can accommodate this new knowledge.  If it cannot, and the knowledge is 
compelling enough, the individual may have to “transform” their system of beliefs to 
include this knowledge (Mezirow, 1990;  Piaget, 1997).   
While the research does not contend that there is a magical formula that can foster 
wisdom in any situation, it is conceivable that the elements of experiential learning theory 
(Kolb, 1984) could contribute to wisdom performance.  The service-learning experience 
being utilized in this study, the Pay It Forward Tour (PIFT), includes elements associated 
with each stage of the Experiential Learning Model.  The stage of “concrete experience” 
will be evidenced by the activities (games, service, cultural tours…) and social and 
environmental encounters involved in trip.  “Reflective observation” is performed via 
debriefing after specific activities, through group discussions, journaling and periods of 
solitude.  “Abstract conceptualization” is commonly facilitated through intentional small 
and large group discussions that to help the participants connect their reflections to 
experiences in their daily lives.  For example, a participant who just participated in a day 
program at a battered women’s shelter would have time to reflect on that experience 
alone and in small groups.  Then the trip directors would lead a discussion designed to 
intentionally highlight the underlying social issues (i.e. gender, power, inequality, etc.).  
By identifying the general issues, the participants can then make connections to other life 
experiences that are affected by the same issues.  The final stage of “active 
experimentation” occurs on the trip as the participants apply their new-found self-
knowledge to future concrete experiences, and later as the participants apply this 
knowledge to their daily lives upon returning home. 
 
45 
 
CHAPTER III 
METHOD 
 This chapter outlines the proposed steps involved in measuring the impact of a 
service-learning experience on wisdom domains.  The first section details the research 
design followed by a discussion of the participants involved in the study, including a 
description of the sample and methods of selection.  The next section includes a 
discussion of the survey instruments, including pre-established instruments and one 
developed with the aid of the programmers.  The chapter concludes with a discussion of 
the data analysis methods that were used to answer the hypotheses and research questions 
as detailed in Chapter I. 
Research Design 
 This study was based on a quasi-experimental, repeated measures design that 
included two measurement points for the control group, and three points for the 
experimental group.  Both groups were given a pre-test survey and a follow-up survey 
one month after the conclusion of the service-learning trip.  The experimental group was 
given an additional survey immediately after the trip experience so that an immediate 
assessment of program impact could be discerned. 
 The experimental group consisted of the entire population of college students 
participating in a nine-day Spring Break service tour.  These students completed a survey 
before, immediately after and one month after the service-learning experience.  The 
control group consisted of a similar group of college students who were not participating 
in the service tour.  These students were comprised of a convenience sample of students 
enrolled in classes at the University of Minnesota.  Pre-test measurements ensured the 
46 
 
integrity of the study by accounting for any differences in the experimental and control 
groups before the experience.  The control group provides an important measure of 
comparison in this study that helps to determine causality in wisdom development.  The 
lack of a control group has been a consistent weakness in previous service-learning 
research (Lakin & Mahoney, 2006).  Previous studies have used control groups to 
measure service-learning outcomes (Batchelder & Root, 1994; Lakin & Mahoney, 2006; 
Lundy, 2007) and wisdom-related performance (Baltes et. al, 1998), but none have 
measured the influence of service-learning on wisdom domains. 
 The quantitative data utilized in this study were collected using paper and web-
based surveys.  Table 3.1 illustrates the data collection time-table.  The control group was 
given paper-based surveys in their respective classrooms during the week prior to the 
service-learning trip.  They were informed of the voluntary nature of the study and signed 
consent forms to indicate their willingness to participate (Appendix A).  The consent 
forms served as a method of identification for each survey, and were removed from the 
survey packet after providing each survey with a code.  Coding the surveys provided for 
participant confidentiality while permitting the researcher to pair the pre and post survey 
instruments.  The follow-up surveys were given to the control group five weeks after the 
initial instrument was completed.  Since these surveys were filled out during normal class 
periods, time-management was be greatly simplified.  Each survey included the date 
completed to ensure consistency in data collection.  All data were entered manually into 
Microsoft Excel for easy transferability into statistical programs (i.e. SPSS, Amos). 
 The experimental group was given paper-based surveys during their pre-trip 
meetings, one week prior to the experience.  The participants included students from 
47 
 
local colleges and universities in the Twin Cities region.  The STLF leaders held pre-trip 
meetings at each college and university one week prior to the tour.  The participants were 
informed of the voluntary nature of the study and signed consent forms indicating there 
agreement to participate (Appendix A).  The surveys were included in the list of 
paperwork that each participant must complete before stepping on the bus.  While the 
surveys were voluntary, their inclusion on the list of paperwork provided a way to 
systematically collect surveys from each participant.  This also provided a way to account 
for those few students who could not attend the mandatory pre-trip meeting, and to 
follow up with them before the experience.  Using the same methods as practiced for the 
control group, the consent forms were removed from the instrument packet once the 
surveys had been coded for pair-wise comparisons.  
 The post-experience surveys were completed on the bus as the participants 
returned from their final destination.  Given that the participants were returning from 
Washington, D.C., they had ample downtime to complete the forms at their leisure.  The 
instrument packets were distributed and collected by the STLF leaders for each bus and 
sealed in an envelope for confidentiality.  Pre-trip training for the STLF leaders ensured 
that they were informed about the importance of participant confidentiality, and agreed to 
handle the data using the specified procedures.  These surveys also included a cover page 
with personal information that was removed once the instruments had been coded for 
identification.   
The one-month follow-up survey for the experimental group was internet-based.  
This was reasoned as the most efficient way to collect the final instruments since there 
was no formal post-trip meeting.  Previous evaluations performed with the Pay It 
48 
 
Forward program achieved a very high response rate (95%) for college-age participants 
using web-based surveys.  Since all of the participants are college students, it was safe to 
assume that they have access to the internet and that they are competent with computers.  
The participants received an email three weeks after the trip to remind them of the study 
and to request their participation in the online survey.  Another email was sent to the 
participants one week later with a link to the web-based survey instrument and a unique 
password which ensured that all participants responded only once.  This password also 
aided in tracking the responses, so that follow-up emails could be directed only to those 
participants who had not yet responded.  Tracking and emailing began the first week the 
survey was available and continued until two weeks after the initial survey posting.  The 
online surveys were tracked by date, so that any discrepancy of measurement could be 
accounted for. 
  
Table 3.1.  
Time-Table for Data Collection 
 January 
2008 
February 
2008 
March 2008 April 2008 May 2008 
Tx Group 
& STLF 
Staff 
Begin 
training 
STLF staff 
Co-Design SL 
component 
Questionnaire 
5
th
-12
th
 Pre-
trip meetings: 
1
st
 Instrument 
collected 
12
th
 Pre-
survey 
emails 
Enter and 
analyze data 
Tx Group   13
th
-22
nd
 Pay 
It Forward 
Tour: 2
nd
 
instrument 
collected 
17
th
-30
th
 
Collect 
online 
follow-up 
surveys 
 
Control 
Group 
  5
th
- 12
th
 
Baseline 
Instrument 
17
th
 – 24
th
 
Collect 
follow-up 
surveys 
 
 
49 
 
Participants 
 The participants in this study included 615 college-aged students, who were 
currently attending a college or university in the Midwestern United States.  Of these 615 
students, 288 received a service-learning “treatment” while the other 327 served as a 
control group.  The experimental group consisted of all the participants involved in the 
Pay-It-Forward service learning tour in March, 2008.  The control group consisted of a 
similar group of students currently enrolled in classes at the University of Minnesota, 
Twin Cities.   
 The STLF does not collect demographic information for its programs, but the 
following additional descriptors may provide some insight into the nature of the PIFT 
clientele.  Nearly all of these participants (99.6%) have had formal classes in leadership 
at the college level.  Over half (53%) have been involved in a “significant” service-
learning experience in the past, either in school or on their own.  Seventy percent report 
having traveled outside of the United States at least once, while 22% have been out of 
North America four or more times.  Almost all (99%) of these students have a leadership 
role in at least one club or extra-curricular activity.  Finally, 87% of the PIFT participants 
are considering a career in the non-profit field. 
Organization of Focus  
The Pay It Forward Tour (PIFT) is the signature program of a non-profit 
organization called Students Today Leaders Forever (STLF).  The company was founded 
by four college freshmen in 2003 and based on the principles set forth in the novel Pay It 
Forward written by Catherine Ryan Hyde.  Although STLF is associated with the 
University of Minnesota, the company receives no University funding and they currently 
50 
 
don’t offer any college credits to their participants.  STLF offers five day experiences for 
high school students and nine day trips for college students during which students 
proceed from a Minnesota locale to another location (i.e. Chicago, Washington D.C.) 
while engaging in different service-learning and cultural immersion experiences each 
day.  These trips are funded primarily by participant fees.   
 The participants are not involved in the planning or trip logistics.  At the 
beginning of a program, the participants will meet at a local school or church gymnasium 
for a brief series of introductions and games.  After these activities, the participants board 
a coach bus (many tours include several buses) and proceed to their first destination.  The 
bus is a central part of the program, and it is possible that the long hours on the bus, 
which provide a wealth of opportunity for intense interaction, could contribute to the 
program outcomes.  When they reach their first destination, the participants engage in a 
service or cultural activity and then proceed to their temporary home.  Sleeping quarters 
may include the floors of school gyms, church auditoriums, or the basements of local 
community members’ houses.  Service activities could place students in Goodwill store 
warehouses or out repairing trails at a state park.   
While the activities vary from year to year and from bus to bus, the structure of 
the experience is largely the same.  Certain environmental (i.e. indoor camping, long bus 
rides) and programming components (i.e. group development, debriefing) are relatively 
consistent.  The STLF has developed a program manual which includes leadership 
expectations and programmatic standards.  The leaders are selected in October and are 
assigned to a core group (4 members) who then work together to design a trip for a group 
of 40 peers.  This bus core is also responsible for recruiting enough college students to 
51 
 
fill their bus.  The leaders participate in several retreats and leadership meetings designed 
to prepare them to facilitate a successful PIFT experience.  Their leader’s manual 
includes a list of standard games and reflective activities that are facilitated on all PIFT 
events.  These include ice-breakers, bus activities, accountability partners, structured 
reflective activities, and a final ceremony to be conducted at their “celebration city” (i.e. 
the final destination city).   
At the beginning of the trip, the buses all leave from one central location and 
travel on their own route to the celebration city.  Each bus participates in unique service 
activities in different cities.  At the end of the trip, all the buses reconvene in the 
destination city (usually Washington, D.C.) to share stories, view pictures, and reflect on 
the impact they have made during their tour of service.  While the service activities are 
unique, the games and reflective activities are consistent across buses.  One example of a 
popular reflective activity is called the “yarn activity”.  During this activity, the 
participants sit in a circle and pass a ball of yarn around randomly with each participant 
holding onto a portion of the string.  Before they can pass the ball of yarn forward, the 
participant must first share a personal struggle with the group.  The struggles shared 
range from surface issues (i.e. no sleep) to very vulnerable issues (i.e. childhood abuse), 
the choice of which is completely up to the participant.  At the completion of the activity, 
each participant is holding a piece of the yarn which has been woven from person to 
person.  The activity emphasizes the fact that, despite obvious differences, they are all 
connected by common life struggles.  Similar activities are conducted throughout the 
week, encouraging the building of relationships and reflection on the daily experiences. 
  
52 
 
Instruments 
 In this study, Ardelt’s (2003) 3-Dimensional Wisdom Scale (3DWS) was 
used to measure cognition, affection, and reflection.  The 3DWS is a 39-item 
questionnaire utilizing a Likert scale format (1= Strongly Agree, 5=Strongly Disagree).  
The cognitive dimension consists of 14 items (i.e. ‘People are either good or bad’; ‘I am 
hesitant about making important decisions after thinking about them’), the reflective 
dimension consists of 12 items (i.e. ‘When I’m upset at someone, I usually try to put 
myself in their shoes for a while’; ‘I always try to look at all sides of a problem’), and the 
affective dimension consists of 13 items (i.e. ‘I can be comfortable with all kinds of 
people’; ‘If I see people in need, I try to help them one way or another’).  Wisdom is 
considered to be a latent variable that is evidenced by the effect indicators of the 3 sub-
domains.  The three subscales are dispersed throughout the questionnaire and several 
items are reverse-scored.  Cronbach’s alpha values for the cognitive, reflective and 
affective domains of the 3DWS are .78, .75, and .74 respectively and they have an overall 
range of .71 to .85 (Ardelt, 2003). 
 To assess theorized relationships, other personal and demographic information 
was collected on the pre-trip questionnaire.  Age and gender were accounted for, as they 
have been shown to affect wisdom scores in adolescents (Baltes, Gluck, & Kunzmann, 
2005; Bailey & Russell, 2008).  School enrollment was assessed to account for any 
variation in scores across schools.  Social environment was measured with four items 
which describe the education of parents, intellectual interests of parents, and educational 
aspirations of the participant.  Social relationships were measured using six items 
53 
 
designed to indicate the number of formal and informal relationships in which the 
participant is involved.   
 In addition to scores on the 3D-WS (Ardelt, 2004), analyses were conducted to 
measure change in civic attitudes and openness to experience.  Civic attitudes were 
measured using the 5-item Civic Attitude Scale (Mabry, 1998).   Reliability for this scale 
was reported at .80.  A subscale of the NEO-PI (Costa & McCrea, 1985) consisting of 12 
items was used to measure the participant’s openness to experience.  The openness to 
experience subscale, one of the well-known “big five” personality traits, has 
demonstrated an internal consistency ranging from .86 to .95, with a stability ranging 
from .51 to .83 over six years of longitudinal measurements.  A copy of the pre and post-
trip questionnaire packets can be referenced in Appendix B. 
Service-Learning Components   
To assess the importance of main elements in the service learning process, a 
questionnaire was developed with the help of the trip leaders.  Using focus group 
interview methods (Gray, 2004), a contingent of 18 trip leaders were asked to generate a 
list of trip elements that they deem important for participant growth.  The insight of the 
trip leaders proved invaluable to the survey development, as they had all been 
participants on previous PIF tours and gave feedback from multiple perspectives.  Having 
participated in the PIFT the previous year, these students were training to lead their first 
tour over the coming spring break.  While focus groups can be difficult to coordinate 
(Gray, 2004), these trip leaders were heavily invested in the program design and were 
highly motivated to evaluate trip outcomes.  The focus group sessions occurred during 
the regularly scheduled pre-trip meetings for PIF leaders.  Though focus groups are 
54 
 
typically associated with qualitative research, the members of the focus group in this 
study were not being measured, but were developing an instrument to measure the 
program impact on participants.  For this reason, these methods are more akin to action 
research than mixed-methods (Creswell, 2003). 
 The first half of the session involved brainstorming and component identification.  
Trip leaders were given time to free-write and discuss trip components in small groups of 
four or five people.  The group then came back together to identify and list as many 
important components of the program that they could recall.  The second stage of the 
session consisted of converging the similar concepts and reducing the list to a 
manageable size.  Since the post-trip instrument packet already contained close to 60 
items, the goal was to reduce the component list to 15 items or less.  Initial convergence 
was done with entire group of trip leaders.  The leaders were then given 10 “votes” (i.e. 
stickey dots) to place next to any of the items they choose.  If they thought one item was 
the only important one, they could place all ten votes on that item.  Those items with the 
most votes were kept and added to the post-trip survey. 
 The final instrument design resembled the questionnaire developed by McKenzie 
(2003) to test the Outward Bound model.  The elements were rated by the trip 
participants (1-5 Likert scale) as to their importance in the development of each wisdom 
domain.  A complete list of the items was laid out in a matrix-type format on the survey.  
For the “affection” dimension, participants were asked to rate each item as to its 
importance for generating their compassion and concern for others.  Reflective and 
cognitive domains were rated in the same way, with the definitions of each being 
consistent with those assessed by the 3D-WS instrument (Ardelt, 2003).   Items were then 
55 
 
ranked by mean scores for each domain.  A sample of this questionnaire can be seen in 
Appendix C. 
Data Analysis 
Validity and Reliability   
The 3D-WS has been established as a reliable and valid test instrument (Ardelt, 
2003).  However, the instrument developer has asked to be informed of the factor 
loadings in order to further confirm the reliability with a young-adult population.  Thus, a 
confirmatory factor analysis (CFA) was conducted using AMOS 16.  In accordance with 
Kline’s (2005) suggestion, multiple fit indices were examined to test the overall fit of the 
three-factor wisdom model as well as the fit of the hypothesized structural equation 
model (SEM).  Specifically, this study examined the Chi-Square statistic (X
2
), root mean 
square of approximation (RMSEA), the comparative fit index (CFI), the Tucker-Lewis 
Index (TLI), and the Expected Cross-Validation Index (ECVI).  Further description of 
CFA methods are detailed in the SEM analysis. 
Outcome Measurements   
To assess the difference in outcome scores between the control and experimental 
groups, a repeated-measures Multiple Analysis of Variance (MANOVA) was conducted 
on each dimension of the 3D-WS and on the Civic Attitudes and Openness to Experience 
scales.  A repeated measures MANOVA with treatment as the between subjects effect, 
was conducted to determine if there is a difference in baseline scores between the 
experimental and control groups and to compare the group differences at 1 month from 
the conclusion of the experience.    This analysis provided the necessary information to 
determine group differences before the treatment, and differences in group trajectories 
56 
 
due to the treatment.  Given an effect size of .25, an α <.05, and sample size of 600, the 
power for this analysis exceeded .90 (Faul, Erdfelder, Lang, & Buchner, 2007). These 
results address hypotheses 1a, 3c, and 3d. 
Service-Learning Components 
This analysis identified those components of the service-learning experience 
which were most salient for increases in each wisdom domain.  This was determined by 
ranking the components under each wisdom dimension (cognitive, affective, reflective) 
according to mean ratings.  Those components with higher mean ratings were identified 
as significant vehicles for the attainment of wisdom via each indicator domain.  This 
information addresses research question 1b.  In addition, it provides the basis for a 
program model, as well as creating a framework for future service-learning research. 
Latent Growth Analysis   
LGA, also referred to as latent growth modeling (LGM) and latent curve analysis 
(LCA), is related to other multi-dimensional analyses, such as hierarchical linear 
modeling (HLM).  In essence, it is a method of SEM that estimates the trajectory of 
growth in a multi-dimensional construct over time.  LGA is unique in its ability to 
describe individual behavior in terms of initial status and developmental trajectories 
(Hancock, Kuo & Lawrence, 2001).  By using a combination of correlations over time, 
changes in variance, and shifts in mean values, LGA can determine the variability across 
individuals, as well as a providing the means to test the contribution of other variables 
and constructs in the regards to initial status and trajectories. 
Figure 3.1 illustrates the latent growth model for the second-order Wisdom 
construct.  Latent factors are represented as ovals, errors and residuals as circles, and 
57 
 
observed variables, as squares.  Because growth coefficients are unknown qualities that 
vary across individuals, they are viewed as latent variables (Muthen & Curran, 1997).  
Conceptually, the growth curve is similar to a regression equation.  The first 
measurement point becomes the reference (intercept) for the second and third 
measurements.  Accordingly, the parameters for the “Intercept” factor are constrained, 
meaning they are not free to be estimated.  This is important as a reference point, but it is 
also essential in order for the model to be identified.  The constraints applied to the slope 
parameters are designed to show the growth trajectory based on the intercept scores.  
These constraints begin with zero, and increase incrementally as more time points are 
added.  Since the proposed study contains only 3 time points, the parameters were labeled 
0,1, and 4.  This labeling implies that a linear, positive change is expected. 
The next important feature in the model is that the intercept and slope factors are 
specified to covary (through their residuals).  This implies that the individual’s initial 
status on wisdom will have an effect on the rate of growth in wisdom.  A positive 
covariance indicates that higher initial wisdom scores are associated with higher 
increases in wisdom over time.  A negative covariance indicates that higher initial scores 
are associated with lower increases over time (Kline, 2005).  Thus, LGA allows one to 
account for individual trajectories, including potential interaction effects due to initial 
status.  Though MANCOVA can produce similar information if the baseline score is 
included as a covariate, the initial status factor provides a more relevant covariate.  
Muthen and Curran (1997) assert that, “The baseline variable may in fact be seen as a 
fallible indicator of this factor (p. 380).”  Previous studies on wisdom have used the 
MANCOVA method described above (Staudinger & Pasupathi, 2003). 
58 
 
 
Figure 3.1.  LGA for Wisdom with 3 Dimensions 
 
The mean associated with the intercept factor represents the average initial 
wisdom score for the entire group (Kline, 2005).  The residual term associated with the 
intercept factor represents the variation of individual scores around the mean initial 
status.  Likewise, the mean for the slope factor indicates the average rate of increase for 
the entire group, while the slope’s error term indicates individual variation around the 
mean rate of growth.  Finally, the error terms for the wisdom scores are allowed to covary 
across time.  This is an important feature that differentiates LGA from other forms of 
multivariate analysis.  It is likely that error variances for individual scores will be related.  
ANOVA assumes that these errors are equal across time and completely unrelated.  These 
assumptions are often violated for repeated measures data (Kline, 2005).  A test of the 
multi-dimensional model was conducted first to identify any areas of misfit before adding 
predictors.   
Intercept 
Int.Status 
Slope 
Growth 
1 
1 
1 
1 
0 4 
 
Wisdom1 Wisdom 2 Wisdom 3 
Ref Aff Cog Ref Aff Cog Ref Aff Cog 
 
59 
 
Based on the results of the SEM analysis, appropriate predictors were added to the 
LGA in order to determine their influence on wisdom trajectories.  Figure 3.2 illustrates 
the LGA for the second-order wisdom construct with two time-invariant predictors and 
one time-variant predictor.  For clarity, errors and covariances are not shown.  The time-
invariant predictors are the exogenous variables in the SEM model, which are measured 
only once and not expected to change as a result of the treatment.  Since it is likely that 
openness will change over the course of the treatment, it will be included as a time-
variant predictor, which has a direct influence on wisdom scores over time.  This gives 
deeper insight into the development of wisdom as influenced by each predictor.  This 
analysis will provide the necessary information to address hypotheses 2a and 2b. 
 
 Figure 3.2.  LGA for Wisdom with 3 Predictors 
 
 
 
Intercept Slope 
 
 
 
Wisdom1 Wisdom 2 Wisdom 3 
Ref Aff Cog Ref Aff Cog Ref Aff Cog 
Social 
Envir 
Social 
Relatio 
Open 1 Open 2 Open 3 
60 
 
Structural Equation Model 
The proposed SEM model was tested with the use of AMOS 16.  In essence, the 
SEM (Figure 1.3) was an exploratory analysis in which a regression was conducted for 
all predictors on all mediators and dependent variables.  The proposed model is a mixed-
method SEM that includes a measurement of the wisdom construct as well as a theorized 
path to wisdom.  This type of assessment is also referred to as a Structural Regression 
(SR) model (Kline, 2005).  Models that include measurement and path components are 
susceptible to interpretational confounding if both components are measured 
simultaneously (Anderson & Gerbing, 1988).  As such, the analysis involves a two-step 
process:  1) Assessment of the wisdom construct through Confirmatory Factor Analysis 
(CFA) and 2) Assessment of the latent path model.  The following sections will address 
the processes and considerations necessary in addressing both types of SEM analysis.  
Construct Validity (CFA)  
 Before attempting to assess the overall fit of the path model, it was first necessary 
to establish the validity of the proposed construct (Figure 3.3).  While the wisdom 
construct measured by the 3D-WS (Ardelt, 2003) has been reported as valid in other 
studies, the instrument may require certain adaptations to prove its applicability to a 
variety of populations.  The appeal of this conceptualization of wisdom is in large part 
due to its parsimony as compared to other complex models (Ardelt, 2003).  While 
parsimony can aid in the application of a construct to various contexts (Kline, 2005), it is 
likely that the relationships between the observed variables and the factors could vary 
considerably with different populations.  This is especially true when dealing with latent 
factors, which require a certain level of intercorrelation among variables without 
61 
 
complete multicollinearity (Kerlinger & Lee, 2000).  The process of assessing validity 
through CFA involves: model specification, model identification, assessment of fit, 
model modification, and model selection. 
 
Figure 3.3. Three Dimensional Wisdom Construct (Ardelt, 2003) 
 
Assessment of fit is the crux of SEM analysis and is often the most ambiguous 
step.  This is due to inconsistencies in practice and reporting, as well as inconsistencies in 
the basic language used by SEM analysts (McDonald & Ho, 2002).  Given findings of an 
inadequate model fit, which is likely in social science research, the analyst should 
investigate the adequacy of the parameter estimates and of the model as a whole (Byrne, 
2001).  While global measures of fit are important in understanding overall model fit, 
they tell the analyst little about the individual parameters.  Parameters can be assessed 
using the factor loadings (coefficients) associated with each variable.  These loadings 
should make conceptual sense, in regards to their size and sign (i.e. positive or negative).  
Inappropriate estimates indicate an insufficient model.  Inappropriate and/or non-
significant parameters can be deleted from the model for the sake of parsimony, assuming 
that their removal makes conceptual sense (Byrne, 2001).   
Affection 
Reflection 
Cognition 
Wisdom 
13 observed  
indicators 
14 observed 
indicators 12 observed 
indicators 
62 
 
When assessing the goodness of fit for the model as a whole, the researcher may 
be faced with a seemingly contradictory set of fitness indices.  In CFA, model testing is 
seen as a hypothesis test.  The null hypothesis is that the specified model is the 
appropriate model for the population.  Unlike standard hypothesis tests, the analyst does 
not want to reject the null hypothesis.  Even if a model is found to be statistically 
significant, however, that does not mean that it is of practical import.  The debate over 
the practice of significance testing in psychology has led to the development of various 
indices to help the researcher determine the real value of the proposed model (Byrne, 
2001).  Since each of the fit indices has its own strengths and weaknesses, the model’s 
fitness was assessed using multiple indices (Hu & Bentler, 1999).   
The X
2
, RMSEA, CFI, TLI, and ECVI statistics were used to determine the 
significance of the model (Kline, 2005).  The X
2 
represents a test of the null hypothesis, 
or the extent to which all the residuals in the model are zero (Byrne, 2001).  The number 
reported as the X
2
 statistic (shown as CMIN in Amos) is the amount of discrepancy in the 
entire model.  For this reason, a lower CMIN value is better.  Unlike a typical 
significance test, the researcher does not want to reject the null hypothesis, so a high p-
value (>.05) for the X
2
 statistic is desirable.  The CFI measures the proposed model 
against a hypothetical “independence” model, in which all of the variables are considered 
independent.  Values for this index range from zero to 1.00, with a value of >.95 
indicating a good fit (Byrne, 2001; Hu & Bentler, 1999).   
The RMSEA measures the proposed model against a contrived population 
covariance matrix.  This index also ranges from zero to 1.00, but a lower value (<.08) is 
desirable (Byrne, 2001).  Finally, the ECVI tests the likelihood that the model would 
63 
 
cross-validate with similar sized samples from the same population.  While there are no 
universal cutoff values for this index, it is recommended that the ECVI for the proposed 
model be lower than comparative models (i.e. independence & saturated models).  By 
addressing all of the above indices, the model will be assessed with a level of rigor that 
will minimize the likelihood of Type I or Type II error (Hu & Bentler, 1999). 
To identify specific areas of misfit in the data, the analyst should consult the 
standardized residuals and the modification indices (Byrne, 2001).  If the researcher 
decides to respecify the model, they will then be operating in a post hoc, exploratory 
mode (Byrne, 2001).  Though this is not uncommon in SEM applications (Anderson & 
Gerbing, 1988), the researcher should be wary of arbitrary modifications that may have 
serious effects on the construct.  The most important criteria for determining whether or 
not to respecify the model is whether or not it makes conceptual sense to free up the 
parameter.  Even if the change could be backed up theoretically, the analyst should strive 
to adhere to the “parsimony principle” (Kline, 2005).  This principle states that the 
simplest model for describing the data is preferred over complex models that may have a 
marginally better fit. 
Path Analysis   
Once an appropriate measurement model was identified, the researchers 
proceeded with the path analysis.  If CFA and path analysis were performed in a single 
step, it would be extremely difficult to identify the source of model misfit.  Once the 
construct was established as valid, any misfit in the full model could be attributed solely 
to the regression paths between the latent variables.  The path analysis was largely 
exploratory, and every variable was allowed to associate with all other variables and 
64 
 
factors.  While there was substantial theoretical backing to support the inclusion of all of 
the variables in the model, there was little to no basis for specifying the relationships 
between these variables a priori. 
The process of model fit assessment and respecification was similar to that 
described in CFA.    First, the individual parameters were analyzed for significance.  Any 
non-significant parameters were removed sequentially, beginning with the least 
significant relationships.  The model was reassessed after each deletion to take account of 
how the modification had affected the entire model (Kline, 2005).  In this manner, the 
exploratory model was reduced to a level of parsimony that may be appropriate for cross-
validation in future research.  All modifications were carefully considered in the effort to 
maintain conceptual meaningfulness.   
The initial analysis was conducted using only the pre-trip scores from the 
experimental group.  All non-significant parameters were deleted from the model and a 
series of modification indices were assessed to determine which structure has the best fit 
(Byrne, 2001; Kline, 2005).  Multi-group CFA was then conducted with treatment as the 
grouping variable to be sure that the proposed model was appropriate for both the 
treatment and control groups.  Given that the model showed a strong fit for both groups, 
the path can be considered valid.  This analysis provides the necessary information to test 
the proposed SEM model across both groups.  These results provide information 
regarding the structure of wisdom development and address hypotheses 3a and 3b. 
65 
 
CHAPTER IV 
RESULTS 
 This section reports the results of quantitative analyses designed to examine the 
effects of a service-learning program on the Wisdom of trip participants.  After a brief 
description of the questionnaire and sampling method, the response rate and reliability 
measures are presented.  This is followed by a presentation of the analytical results for 
each stated hypothesis.  This study included cross-sectional and longitudinal analyses to 
determine the relationships of predictive and mediating variables, their comparative 
influence on the Wisdom construct, and the effects of a service-learning trip on the 
mediating variables and overall Wisdom of trip participants.  Wisdom was the dependent 
variable, indicated by its three sub-domains (Affection, Cognition & Reflection).  
Openness to Experience and Civic Attitude served as mediators, which were expected to 
vary over time and to influence Wisdom scores.  Social Environment and Social 
Relationships were the independent predictor variables.  Other variables which were 
accounted for include: age, gender, grade-point average, and average amount of time 
spent in personal reflection. 
Summary of the Instrument 
 The pre-test questionnaires (see Appendix B) were distributed to the 
treatment group one week before the trip and collected at the time of trip departure.  A 
total of 288 students participated in the Pay It Forward Tour.  From this sampling frame, 
a total of 235 completed the pre-trip instrument completely, resulting in a response rate of 
82%.  The post-test measurement was given to the treatment group immediately after the 
trip experience.  These questionnaires included the additional section (Appendix C) 
66 
 
designed to measure the impact each trip component had on each Wisdom domain.  
Participants were given the questionnaires during the long bus-ride home after the final 
ceremony in Washington D.C.  Of the 288 participants, 227 (79%) questionnaires could 
be matched for comparison of pre and post-trip measurements.    
The follow-up measure for the treatment group was administered during the fifth 
week after the conclusion of the trip using an online survey tool.  The participants 
received an email one week prior to the release of the survey tool to convey the 
importance of their participation.  A link was then emailed to the participants six times 
over a two-week period to encourage them to complete the questionnaire.  A total of 191 
participants responded to the online survey, resulting in a 66% response rate.  
Unfortunately, a number of these respondents did not complete one of the first two 
questionnaires completely.  As a result there were only 121 (42%) participants for which 
growth trajectories could be calculated using all three time points. 
The control group consisted of a non-equivalent convenience sample of students 
currently enrolled in Recreation, Sports Management, and Physical Activity Program 
classes at the University of Minnesota.  Control group participants either completed the 
questionnaire during scheduled classes or they completed them at home and returned 
them to the course instructor.  A total of 307 questionnaires were collected from control 
group participants.  The final questionnaire was distributed during the fifth week after the 
conclusion of the service-learning trip.  The control group completed a questionnaire that 
was identical to the pre-trip measurement.  Of the original 307 surveys collected, 209 
(68%) could be matched for comparison of pre-trip and follow-up measurements.  This 
67 
 
moderate response rate was due to sporadic class attendance at the end of the semester 
and the voluntary nature of the study.   
Supplemental Items 
To determine which aspects of the experience were most salient for growth in 
each Wisdom domain, a supplemental survey was added to the post-trip questionnaire for 
the treatment group.  This section was developed using an expertise-based approach 
(Russ-Eft & Preskill, 2001) .  A group of 18 STLF leaders, all of whom had been 
participants on previous PIF tours, listed and prioritized the elements of the tour that they 
deemed most valuable for growth.  After narrowing these elements down to a list of 15 
items, these items were then clustered into conceptually appropriate factors.  These items 
and their factors can be seen in Table 4.1.  Each factor was represented by at least two 
items, and the score for each factor was determined by the average of the indicators for 
that factor.   
Table 4.1  
Items and Factors Included on the Supplemental Post-Test Questionnaire. 
Item Factor 
Diversity of Trip Participants 
Informal Conversations with Participants Relationships 
Conversations with Community Members 
Opportunity to take charge of projects 
Working with others toward a common goal Service 
Helping others in need 
Being away from my daily routine Travel 
Seeing new places 
Personal quiet times 
Group discussions Reflection 
Structured reflection activities 
The direction of the STLF leaders 
68 
 
Personal discussions with STLF leaders Leaders 
The example set by the STLF leaders 
Other: (Please Specify) 
 
Response Bias 
 In order to determine if the treatment group participants who responded for all 
three time points represented a unique group within the population, an independent 
samples t-test was performed on demographics and variables of interest as measured by 
the pre-test questionnaire.  The only significant difference was for gender (p = .005), 
indicating that more females were measured for all time points than males.  While gender 
is related to other variables in the study, there is no reason to believe that a difference in 
this variable alone would introduce undue bias regarding the longitudinal results. 
Selection Bias 
Since the treatment and control groups were not randomized and were non-
equivalent, an independent t-test was conducted on these groups using the pre-test 
measurement (see Table 4.2).  The groups did differ significantly on most measures.  In 
order to determine if these differences introduced bias into the quasi-experimental design, 
the analyses designed to compare these two groups were conducted systematically on a 
sample of cases.  First, a MANOVA was conducted on Wisdom, Openness, and Civic 
Attitude using all cases from the treatment and control groups.  Next, a random sample of 
the control group was selected from all of the existing cases using SPSS.  A separate 
MANOVA was conducted using this sample as the control group.  Finally, a purposive 
sample of the control group cases was selected so that the control and treatment groups 
did not differ significantly on Wisdom scores for the first measurement.  A final 
69 
 
MANOVA was conducted using the purposive sample as the control group.  The results 
were comparable regardless of the sampling method used.  These findings indicate that, 
though the comparison groups were different on most measures, these differences have 
little effect on longitudinal measures for the dependent variables (see Table 4.3).  All 
subsequent analyses were conducted using the entire control group.    
Table 4.2 
Independent t-test on Initial Status of Treatment and Control Groups. 
Mean 
Difference 
Sig.          
(2-tailed) 
Affective -0.296 <0.001 
Cognitive -0.251 <0.001 
Reflective -0.208 <0.001 
Civic Attitude -0.339 <0.001 
Openness -0.308 <0.001 
Social Environment 0.028 0.628 
Social Relationships -0.207 0.001 
Gender -0.219 <0.001 
Age 0.982 0.002 
 
Table 4.3 
F-Statistic Values for Outcome Variables Using Various Sampling Methods. 
  Affective  Cognitive  Reflective  Openness  Civic Attitude  
All Participants 1.53 1.52 4.06* 7.88** 1.96 
Random Control 1.86 1.77 4.62* 4.50* 1.73 
Purposive Control 4.44* 3.64 5.38* 11.49** 0.30 
 
Construct Validity and Reliability 
 Given a cutoff value of 0.70 (Howell, 2002), Cronbach’s alpha was acceptable for 
all five of the psychometric measurement scales (Table 4.4).  Kline (2005) suggests that 
constructs utilized in SEM should not have a skew in excess of + 3 nor a kurtosis 
70 
 
exceeding + 10.  Skewness and kurtosis levels for the five dependent variables were 
within acceptable limits. 
Table 4.4   
Cronbach’s Alpha, Skewness & Kurtosis Levels for Dependent Variables. 
Cronbach's 
Alpha Skewness Kurtosis 
Civic Attitude 0.88 -1.54 3.980 
Openness 0.79 -0.223 0.555 
Affective 0.75 0.011 0.381 
Cognitive 0.80 -0.163 0.292 
Reflective 0.77 -0.185 0.819 
Wisdom * -0.085 0.817 
    *Latent Construct (Measured via Confirmatory Factor Analysis) 
 In order to measure the validity of the latent Wisdom construct, a CFA was 
conducted using the mean scores for the Cognitive, Affective, and Reflective domains as 
observed indicators.  In order to conduct a CFA on a construct with only three indicators, 
the researcher must apply constraints in order for the model to be identified (Kline, 
2005).  Ardelt (2003) suggests that the Cognitive and Affective domains can be 
constrained so that they load equally on the Wisdom construct, because there is no 
conceptual argument for these domains to differ from one another.  The Reflective 
parameter should be freely estimated because it is reasoned that this domain is the “hub” 
of Wisdom.  For this reason, the Reflective domain is expected to have a higher factor 
loading on the Wisdom construct.  In coherence with Ardelt’s (2003) model, the CFA was 
conducted with the Affective and Cognitive loadings constrained to be equal and the 
Reflective loading freely estimated.  The model showed a strong fit for the data (Table 
4.5), indicating that the structure of the Wisdom construct is valid for this sample.   
 
71 
 
Table 4.5  
Fit Indices for the CFA of the Wisdom Construct. 
CMIN p TLI CFI RMSEA 
Wisdom CFA 6.968 0.031 0.987 0.996 0.045 
 
To determine whether or not the Reflective domain indeed factored higher on the 
Wisdom construct in this sample, the three indicator domains were respecified to covary 
without the addition of the latent Wisdom factor.  This model is mathematically 
equivalent to the previous model with the exception that the domains are related through 
their error variances, and their covariances are all allowed to be freely estimated (Byrne, 
2001).  Using this technique, one can determine which domain has the highest factor 
loading by examining the Critical Ratio (C.R.) associated with each covariance.  As 
shown in Table 4.6, it is evident that the two highest C.R. values are associated with the 
Affective domain, indicating that, in this sample, the Affective domain was the “hub” of 
the Wisdom construct. 
Table 4.6   
Inter-domain Covariance Within the Wisdom Construct 
Covariance C.R. 
Cognitive <--> Affective 12.019 
Cognitive <--> Reflective 11.651 
Affective <--> Reflective 13.290 
 
72 
 
MANOVA and t-Test Analyses 
 Hypotheses 1a and 3c were assessed through the use of two separate MANOVAs 
and Hypothesis 3d was assessed using a paired-samples t-test.  To assess the difference in 
Wisdom gains between the control and treatment groups, a repeated-measures MANOVA 
was conducted.  The pre and follow-up measurements for the three Wisdom sub-domains 
were defined as the dependent variables and group status as the between subjects 
variable.   Due to the fact that the control and treatment groups differed significantly on 
all measures before the trip, it was difficult to discern from the omnibus F-statistic 
whether the significance was due to the difference in the groups’ initial status or to the 
change in scores over time.  For this reason, the MANOVAs were conducted on 
“difference” scores to assess the difference in growth for each sub-domain. 
Hypothesis 1a.   
This hypothesis stated that those participating in a service-learning trip will 
demonstrate no higher gains in Wisdom domains than the control group.  To determine if 
the treatment group demonstrated significantly higher gains in Wisdom domains over the 
course of five weeks, a new variable was created for each of the three sub-domains.  This 
variable was created by subtracting the score on the initial survey from that of the follow-
up measurement.  A MANOVA was then conducted on the new variable for each sub-
domain with group status as the between subjects variable.  The results revealed that the 
two groups showed significantly different growth patterns in the Affective and Reflective 
domains.  The Mean difference for the control group was negative on all three sub-
domains, indicating that students who did not participate in the program declined in 
Wisdom over the course of five weeks.  The Mean difference scores on the Affective and 
73 
 
Reflective domains for the treatment group increased, while their scores on the Cognitive 
domain decreased over the course of five weeks.  This null hypothesis can be rejected.  
The results for this analysis can be seen in Table 4.7 and are illustrated in Figure 4.1. 
Table 4.7  
MA/OVA for Difference Scores of Wisdom Sub-Domains. 
(/)   
Control 
(/) 
Treatment 
Mean Diff 
Control 
Mean Diff 
Treatment F 
Sig.       
2-tailed 
Affective 192 121 -.088 .022 6.270 .013 
Cognitive 192 121 -.092 -.067 .260 .610 
Reflective 192 121 -.041 .078 7.591 .006 
 
 
 
Figure 4.1  Mean Difference scores for Control and Treatment Groups on Wisdom 
Sub-domains. 
Hypothesis 3c  
This hypothesis stated that service-learning participants would demonstrate no 
higher gains in civic attitude and openness than the control group.  To test this 
hypothesis, a new variable was created for civic attitudes and openness by subtracting the 
74 
 
score on the pre-test measurement from that of the follow-up measurement.  A 
MANOVA was then conducted on this new variable to determine if the treatment and 
control groups showed significantly different gains on these variables.  There was no 
significant difference for the two groups on measures of civic attitude, as both groups 
demonstrated positive gains.  However, there was a significant difference for the 
openness variable.  The treatment group demonstrated a Mean increase in scores for 
openness, while the control group showed a decline in openness scores.  Therefore, the 
null hypothesis was rejected (see Table 4.8 and Figure 4.2).   
Table 4.8   
MA/OVA for Difference Scores of Civic Attitude and Openness. 
(/)   
Control 
(/) 
Treatment 
Mean Diff 
Control 
Mean Diff 
Treatment F 
Sig.       
2-tailed 
Civic Attitude 192 121 .043 .166 2.080 .150 
Openness 192 121 -.090 .068 10.518 .001 
 
 
Figure 4.2.  Mean Difference Scores for Control and Treatment Groups on Civic 
Attidude and Openness. 
 
75 
 
Hypothesis 3d   
This hypothesis stated that there would be no difference in pre and post-trip 
scores for civic attitude and openness-to-experience for service-learning participants.  To 
test this hypothesis a paired samples t-test was conducted on the pre and post-trip 
treatment group scores for civic attitude and openness.  Given that the control group was 
not administered a post-trip measurement, no comparisons were made across groups.  As 
shown in Table 4.9, the treatment group demonstrated significant gains in civic attitude 
and openness to experience over the course of the service-learning experience.  Even 
when adjusting for the multiple dependent variables using the Bonferroni method (p< 
.025), the p-values are well below significance levels.  As such, hypothesis 3d was 
rejected. 
Table 4.9   
Paired t-Test for Pre and Post-Test Civic Attitude and Openness Scores. 
t df Sig. (2-
tailed) 
CivAtt1 - CivAtt2 -4.59 179 <.001 
Openness1 - Openness2 -4.33 179 <.001 
 
Service-Learning Component Ranking 
Research Question 1b   
To determine which elements of the service-learning program were most salient 
for growth in each Wisdom domain, the mean scores for each factor were ranked under 
the three sub-domains.  Though these items were clustered under each factor based on 
their conceptual relevance, reliability measures indicate that these items were perceived 
as similar by the participants as well (see Table 4.10).  Cronbach’s alpha for each factor 
76 
 
measured across all three Wisdom sub-domains was above the recommended level of .70 
(Howell, 2002).   
Table 4.10   
Reliability Measures for Trip Component Factors. 
Factor # Items Cronbach's alpha 
Relationships 9 0.83 
Service 9 0.85 
Travel 6 0.83 
Reflection 9 0.88 
Leaders 9 0.96 
 
The relationships formed on the trip were perceived as the most important aspect 
for growth in the Reflective domain (see Table 4.11).  The most important component for 
growth in the Affective and Cognitive domains was the “service activities” (see Tables 
4.12 & 4.13).  Both of these domains increased significantly during the trip, then returned 
to pre-trip levels at the final measurement.  All three domains were most influenced by 
the service activities and the relationships formed on the trip.  The reflective activities 
were ranked third in every domain, with the “travel” and “leader” factors ranking 4
th
 and 
5th across all domains.   
Table 4.11  
Component Ranking for the Reflective Domain. 
N Minimum Maximum Mean Std. 
Deviation 
Relationships 188.00 2.00 5.00 4.34 0.61 
Service 187.00 2.00 5.00 4.24 0.69 
Reflection 188.00 1.00 5.00 4.19 0.89 
Travel 188.00 1.00 5.00 4.16 0.86 
Leaders 188.00 1.00 5.00 4.03 0.99 
 
77 
 
Table 4.12   
Component Ranking for the Affective Domain. 
N Minimum Maximum Mean Std. 
Deviation 
Service 188.00 2.00 5.00 4.35 0.55 
Relationships 187.00 2.00 5.00 4.29 0.60 
Reflection 187.00 1.00 5.00 4.13 0.92 
Leaders 187.00 1.00 5.00 3.99 1.03 
Travel 187.00 1.00 5.00 3.70 1.05 
 
Table 4.13   
Component Ranking for the Cognitive Domain. 
N Minimum Maximum Mean Std. 
Deviation 
Service 187.00 1.00 5.00 4.15 0.70 
Relationships 188.00 2.00 5.00 4.14 0.68 
Reflection 188.00 1.00 5.00 4.07 0.92 
Travel 187.00 1.00 5.00 3.96 0.89 
Leaders 188.00 1.00 5.00 3.87 1.07 
 
 In addition to analyzing the trip component factors for each Wisdom domain, an 
examination of the means for each individual component across all domains helps to 
clarify the role each component plays in the cultivation of Wisdom.  A total score for each 
trip component was attained by adding the means for each sub-domain (see Table 4.14).  
The six highest of these component rankings represent a mixture of service and relational 
elements.  Lower ratings were given to the structured program components, the travelling 
elements, and leadership. 
 
 
 
78 
 
Table 4.14   
Ranking of Total Scores for Trip Components. 
Trip Components Total Score 
Helping Others in Need 13.63 
Working Toward a Common Goal 13.14 
Conversations with Community Members 12.93 
Group Discussions 12.75 
Informal Conversations with Participants 12.74 
Diversity of Trip Participants 12.63 
The Example Set by the Leaders 12.34 
Seeing New Places 12.26 
Structured Reflective Activities 12.05 
Direction of the Trip Leaders 11.69 
Informal Discussions with Leaders 11.64 
Service Projects 11.46 
Being Away from Normal Routines 11.36 
Quiet Time Alone 11.33 
Ice Breakers and Games 10.02 
  
Structural Equation Model 
Hypothesis 3a   
To determine if social environment and social relationships had a significant 
direct effect on Wisdom scores a structural equation model was constructed using Amos 
16.  This model is illustrated in Figure 4.3.  The mean scores for Social Environment (3 
items) and Social Relationships (6 items) were included as predictors of the three the 
Cognitive, Affective, and Reflective Wisdom domains.  Mean scores for the Civic 
Attitude and Openness scales were included in the model as additional predictors.  This 
analysis is similar to a single-step regression model, which controls for the effects of all 
predictors in the model.  A comparison of the factor loadings for each predictor can be 
seen in Table 4.15.  Social Relationships had a significant direct effect on all three 
79 
 
Wisdom domains, but Social Environment had no direct effect on any domain.  
Hypothesis 3a is partially supported by these results. 
Table 4.15 
Standardized Direct Effects of Predictors on Wisdom Domains. 
Affective Cognitive Reflective 
Social Environment 0.008 0.027 0.029 
Social Relationships 0.122** 0.132*** 0.125*** 
Openness 0.271** 0.551*** 0.289*** 
Civic Attitude   0.305*** 0.067* 0.250*** 
*p < .05,  **p < .01,  ***p < .001 
 
Hypothesis 3b  
This hypothesis stated that openness to experience and civic attitude would have a 
stronger direct effect on Wisdom scores than would social environment and social 
relationships. This hypothesis was tested using the information from the SEM described 
above.  As seen in Table 4.15, Openness to Experience had stronger direct effects on all 
three Wisdom domains than did Social Environment and Social Relationships.  Civic 
attitude had stronger direct effects on Wisdom domains than did Social Environment, as 
well as stronger direct effects on two of the three domains (Affective and Reflective) than 
did Social Relationships.  These results support Hypothesis 3b. 
 
80 
 
 
 
    
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 4.3.  Standardized Regression Weights for Predictors on Wisdom Domains. 
In order to test the full SEM proposed for this study, additional predictors were 
introduced into the model.  The new predictors included age, gender and GPA.  Social 
Environment, Social Relationships, age, and gender were specified as exogenous 
predictors which affected all other variables in the model.  Openness, Civic Attitude and 
GPA acted as mediators, being influenced by the aforementioned predictors and also 
influencing the three Wisdom domains.  The analysis proceeded in an exploratory 
manner, by first allowing all predictors to influence all mediators and dependent 
.067* 
.305** 
.250** 
.551** .289** .271** 
.122** 
.132** 
.125** 
.029 
.008 
.027 
*p< .05,  **p< .001 
81 
 
variables.  After each consecutive analysis, a single non-significant parameter was 
deleted from the model (Byrne, 2001).  This process continued until every non-significant 
parameter was deleted from the model. 
  After dropping all non-significant parameters, the proposed model demonstrated 
a strong fit.  However, as suggested by Kline (2005) a good model fit does not indicate 
that the best model has been found.  First, a confirmatory factor analysis (CFA) should be 
performed in order to determine that the proposed model fits the data for multiple 
samples.  To do this, a multiple-group CFA was conducted using Amos 16 using the data 
for treatment and control groups.  First, the model was tested using only the data from the 
control group.  The control group data fit the proposed model well.  Next, the model was 
tested using the data from the treatment group, while constraining certain parameters.  All 
covariances, factor loadings and error variances for the treatment group model were 
constrained to be equal to those of the control group.  In this way, the highly restricted 
model could be verified by a set of data from an independent population (Byrne, 2001).  
The restricted model showed a strong fit when applied to the treatment group data (see 
Table 4.16), validating the Confirmatory Factor Analysis.   
Table 4.16  
  
Fit Indices for the Multi-group Confirmatory Factor Analysis. 
Fit Indices 
CMIN 
(X2) DF P TLI CFI RMSEA      ECVI 
Proposed Model 126.88 76 .000 0.937 0.957 0.033 0.388 
Saturated Model             0.439 
 
The details of each fit index are described in Chapter three, but a brief review will 
aid with interpretation.  The X
2 
statistic should not be significant, but this statistic is 
influenced by sample size, making it an unreliable index in studies involving more than 
82 
 
200 participants (Kline, 2005).  Hu & Bentler (2001) recommend a TLI and CFI of <.90 
along with a RMSEA of <.05.  Finally, while there are not standardized cut-off values for 
the AIC and ECVI, Byrne (2001) contends that the values for the proposed model should 
be smaller than those for the “saturated” model.   
For the final SEM analysis, all data for the treatment and control groups were 
utilized.  The proposed model demonstrated a strong fit for both groups, with all 
parameters remaining significant (see Table 4.17).  One of the most common errors when 
using SEM is to neglect the testing of the proposed model against alternative models 
using the same variables (Kline, 2005).  In order to determine that the proposed model 
was the best model for the data, three alternative models were tested: 1) The Reverse 
Model, in which Wisdom was a mediator for openness and civic attitude; 2) The 
Covariance Model, in which all variables were specified to be related, but with no causal 
inferences; and 3) The Two-Stage model, in which openness and civic attitude were 
treated as predictors instead of mediators.  Comparison of multiple fit indices reveals that 
the proposed model fits the data better than the alternative models (see Table 4.17).   
Table 4.17 
 
Comparison of Fit Indices for the Proposed Model and Alternative Models. 
 
CMIN 
(X2) DF P TLI CFI RMSEA ECVI 
Proposed Model 82.918 42 0 0.942 0.978 0.033 
0.287 
Saturated Model     0.288 
Reverse Model 128.643 29 0 0.855 0.923 0.075 0.326 
Saturated Model     0.211 
Covariance Model 203.925 40 0 0.84 0.929 0.058 0.276 
Saturated Model     0.175 
2-Stage Predictors 404.104 44 0 0.654 0.862 0.082 0.468 
Saturated Model             0.211 
83 
 
  
 
 
 
   *p < .05,  **p < .001 
 
Figure 4.4.  Full Final Model. 
 
 
 Table 4.18 displays the amount of variance explained for each variable in the 
model.  As exogenous variables, age, social environment and social relationships were 
not predicted.  Social environment accounted for a significant, albeit small, amount of the 
variance in GPA.  Age and gender combined account for only three percent of the 
variance in openness, while gender and social relationships account for a similar amount 
of the variance in civic attitude.  All of the predictors in the model account for about one 
quarter of the variance in the Affective and Reflective domains and about two-fifths of 
the variance in the Cognitive domain. 
 
.291** 
.156** .301** 
.066* 
.292** 
.161** 
.116** 
.103** 
.135** 
.086* 
.552** 
.264** 
.249** 
.129** 
.160** 
.135** 
84 
 
Table 4.18 
  
Percent of Variance Explained in Each Item. 
 
Item %  
Social Environment * 
Age * 
Social Relationships * 
GPA 9 
Openness 3 
Civic Attitude 5 
Affect 26 
Reflective 22 
Cognitive 42 
 
 Not shown in Figure 4.4 are the relationships between the predictor variables.  As 
with a multiple regression analysis, these predictors often share a considerable amount of 
variance with one another.  This shared variance affects the power of each predictor to 
explain the difference in the dependent variable.  All non-significant relationships were 
dropped in the process of establishing the best fit for the final model.  Table 4.19 displays 
all of the significant relationships between the predictor variables.  Social relationships 
were negatively related to age indicating that older students tended to be less socially 
active.  Those students from a positive social environment were involved in more social 
groups.  The two mediator variables, civic attitude and openness, demonstrated a strong 
positive relationship, and both were also positively related to GPA.  Finally, the three 
dimensions of the Wisdom construct were highly related. 
 
 
 
 
85 
 
Table 4.19 
Correlations of SEM Variables. 
    Estimate 
Social Relationships <--> Age -0.123* 
Social Relationships <--> Social Environment 0.134** 
Civic Attitude <--> Openness 0.322** 
Openness <--> GPA 0.244** 
Civic Attitude <--> GPA 0.266** 
Affective <--> Cognitive 0.407** 
Cognitive <--> Reflective 0.379** 
Affective <--> Reflective 0.520** 
 
 
Latent Growth Analysis 
Hypothesis 2a & 2b 
To assess the differences in pre- and post-trip Wisdom scores, and in pre- and 
follow-up Wisdom scores for service-learning participants a mean structure analysis was 
conducted using the treatment group data to compare Wisdom scores over three time 
points (see Figure 4.5).  Given that Wisdom is considered a latent variable that cannot be 
directly observed, the mean of this variable is estimated with the use of the three indicator 
domains (Affective, Reflective, and Cognitive).  The mean for the latent Wisdom variable 
for the pre-test (wise) was constrained to zero in order to provide a baseline from which 
to compare the post and follow-up measures.  The means for the Wisdom variable on the 
post-test (wise2) and the follow-up test (wise3) were allowed to be freely estimated. The 
results for this analysis are presented in Table 4.20.  
  
 
86 
 
 
 
 
 
Figure 4.5.  Mean Structure Analysis for Pre, Post, and Follow up Wisdom Scores. 
 
The results indicate that the mean Wisdom score on the post-test was significantly 
different from that of the pre-test.  The parameter of interest is the Critical Ratio (C.R.), 
which must be greater than + 1.96 to be considered significant at an alpha level of p < .05 
(Byrne, 2001).  Since the C.R. is greater than 1.96 and positive, the mean Wisdom score 
for the post-test instrument (wise2) is significantly higher than that of the pre-test.  While 
the C.R. value is smaller for the follow-up measurement (wise3), the results indicate that 
the increase in overall Wisdom scores remained significant one-month after the 
experience.  Hypotheses 2a and 2b can be rejected. 
 
 
 
 
 
87 
 
Table 4.20 
  
Chi-Square test for the Difference in Pre and Post and Follow-up 
Wisdom Scores. 
 
Estimate C.R. p 
wise2 0.14 6.28 <.001 
wise3 0.072 2.996 0.003 
 
 
 A unique facet of Latent Growth Analysis is the ability to measure individual 
status and individual growth trajectories over multiple time points.  While an ANOVA 
allows the researcher to determine the trajectory of the group mean, LGA enables the 
analyst to investigate individual differences in trajectories, and even test the predictive 
value of independent variables on individual entry status and growth trajectories.  Figure 
4.6 shows a comparison of the group mean trajectory and the individual trajectories of the 
first 30 study participants.  It is evident from this illustration that much valuable 
information would be missed if the mean group trajectory was the only parameter of 
interest.  In order to determine the influence of predictor variables on initial status 
(Intercept) and individual trajectories (Slope), all predictors were entered into the model 
in a single step. 
 To assess the impact of each predictor on the original status (intercept) and the 
growth trajectories (slope) of individual participants, the following predictors were added 
to the Latent Growth Curve model in one step: Social Relationships, Social Environment, 
Age, Gender, Grade Point Average (GPA), and weekly time spent in personal reflection 
(PerRef).  In order to understand the contribution of each predictor to each Wisdom sub-
domain, a separate LGA was conducted on each sub-domain as well as on the second-
order latent Wisdom construct.  These results can be seen in Table 4.21.   
88 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
Figure 4.6.  Comparison of Group Mean Trajectory with Individual 
Trajectories. 
 
 
 A significant relationship between the predictor and the intercept variable is 
equivalent to a cross-sectional regression analysis, since the intercept is a measure of the 
initial status on the pre-trip questionnaire (Bollen & Curran, 2006).  The strongest 
predictor of overall Wisdom was Gender, being bolstered by a high female score on the 
Affective domain.  It is noteworthy that Gender was not a good predictor of the 
Reflective domain.  Social Relationships were also a strong predictor of overall Wisdom, 
being significantly related to the Affective and Reflective domains.  Self-report measures 
89 
 
of GPA were significantly related to all three Wisdom domains, being a very strong 
predictor of the Cognitive domain. 
Table 4.21 
 
Predictors of Latent Intercepts and Growth Trajectories of Wisdom Domains. 
 
Affective Cognitive Reflective Wisdom 
Soc Rel Intercept 2.913** 1.696 3.459*** 3.321*** 
Slope -0.601 0.626 -0.859 -0.787 
Soc Envir Intercept 0.262 0.531 0.054 0.187 
Slope -0.132 -0.086 -0.084 0.119 
Age Intercept -1.499 0.862 0.269 -0.607 
Slope 0.651 0.136 -1.676 0.158 
Gender Intercept 3.962*** 2.466* 1.681 3.554*** 
Slope -1.346 -0.932 -1.448 -2.598** 
GPA Intercept 2.357* 5.487*** 2.149* 3.278*** 
Slope -0.160 -2.343* 1.447 0.276 
Per Ref Intercept 1.576 1.238 1.069 1.598 
  Slope 2.440* 2.410* 1.538 2.258* 
 * p < .05,  **p < .01,  *** p < .001 
 
 A significant relationship to the Slope variable indicates that the predictor had an 
effect on the latent growth trajectory of Wisdom.  While females showed a higher initial 
status on Wisdom domains, males demonstrated a higher rate of growth over the five 
week period.  The amount of time spent in personal reflection (PerRef) was marginally 
related to initial Wisdom scores, but did emerge as a significant predictor of the overall 
growth trajectory.  GPA, while initially a strong predictor of all three Wisdom domains, 
only had a significant effect on growth in the Cognitive domain.  Those with high GPA 
scores reported lower growth rates for cognition over the five week measurement period.   
 The final analysis for this study used a multivariate LGA including time-invariant 
and time-variant predictors of Wisdom.  The six predictors that were included in the 
90 
 
original LGA were considered time-invariant because they were measured only once and 
were not expected to change as a result of the service-learning program.  The mediating 
variables (Openness and Civic Attitude), however, were measured at all three time points 
with the assumption that they might show marked change.  For this reason, Openness and 
Civic Attitude were included in the model as time-variant predictors, each with its own 
latent curve trajectory.  The LGA for Wisdom, the LGA for Openness, and the LGA for 
Civic Attitude were all regressed onto the six time-invariant predictors.  In addition, the 
three latent growth curve models were allowed to covary, providing some indication of 
how each growth trajectory affected other trajectories.  The results for this analysis can 
be seen in Table 4.22.   
 
Table 4.22 
 
Time Invariant Predictors of LGA for Wisdom, Civic Attitude, and Openness. 
 
    Wisdom Civic Attitude Openness 
Soc Rel Intercept 3.123** 1.648 -1.106 
  Slope -0.295 0.215 -0.275 
Soc Envir Intercept 0.269 -1.501 -0.215 
  Slope 0.098 2.819** 0.918 
Age Intercept -0.105 -0.425 0.377 
  Slope -0.343 0.790 0.861 
Gender Intercept 3.275*** 3.167** 1.842 
  Slope -2.380* -1.712 -2.148* 
GPA Intercept 3.907*** 3.119** 3.203*** 
  Slope -0.773 -2.802* -0.597 
Per Ref Intercept 1.472 1.021 2.087* 
  Slope 2.628** 0.046 2.452** 
% Variance Intercept 20.5 14.6 8.7 
  Slope 63.4 19.1 19 
 * p < .05,  **p < .01,  *** p < .001 
 
91 
 
 Age was neither a significant predictor of initial status nor of the growth 
trajectory for any of the dependent variables.  All variables were affected by Gender, with 
females demonstrating a higher initial status and males showing a stronger growth 
trajectory over the five week period.  GPA was a significant predictor of initial status for 
Wisdom, Openness and Civic Attitude.  Those with a lower initial GPA demonstrated a 
stronger positive growth trajectory for Civic Attitude.  Those with a higher score for 
Social Environment also demonstrated a more positive growth trajectory for Civic 
Attitude.   
 The percentage of variance explained for each variable is also included in Table 
4.22.  These numbers represent the amount of variance accounted for by all six predictor 
variables.  The variance for the slope of Wisdom is of particular interest.  Though gender 
and personal reflection were the only significant predictors of this variable, 63% of the 
variance in the growth trajectory of Wisdom was explained by the predictors in the 
model.   
 Table 4.23 summarizes the relationships between initial status (intercept) and 
growth trajectories (slope) for Wisdom, Openness and Civic Attitude.  These results 
present no surprises for the three variables, since their positive relationship was already 
demonstrated in the SEM analysis.  In general, a high initial status for each variable is 
associated with a lower growth rate in all three variables.   
 
 
 
 
 
 
 
 
92 
 
Table 4.23 
 
Covariance of Time-Variant Predictors and the Wisdom Construct. 
 
Wisdom Openness Civic Attitude 
Intercept Slope Intercept Slope Intercept  Slope 
Wisdom Intercept   .597 6.05*** -0.562 6.176*** -2.296* 
Slope -2.279* 5.258*** -3.765*** 4.602*** 
Openness Intercept -1.143 3.976*** -2.278* 
Slope -2.554** 4.544*** 
Civic Attitude Intercept -4.341*** 
* p < .05,   **p < .01,   *** p < .001 
 
   
 
 
93 
 
CHAPTER V 
CONCLUSIONS 
 This chapter reviews the findings of the current study and the implications of the 
hypotheses and research question.  The conclusions proceed from a discussion of cross-
sectional results (SEM) to a discussion of the results from longitudinal analyses 
(MANOVA, LGA).  Course component rankings are included in the longitudinal 
discussion.  Comparisons to existing research findings are made within each respective 
section, along with the implications of said results.  This chapter concludes with 
recommendations for future research that may further the study of Wisdom, experiential 
education, and the field of human development.   
Cross-Sectional Conclusions 
 The results were obtained with the use of SEM using only the first measurement 
for treatment and control groups.  The final model was achieved through a systematic 
reduction of all non-significant parameters using only the treatment group data.  The 
model was then validated through a CFA using the control group data.  All SEM results 
include the data from all participants in the study (n= 607).  It is difficult to determine 
causation when utilizing cross-sectional data, and care must be taken not to extrapolate 
findings beyond the scope of this study.  Taking this precaution into account, these results 
proceed by first examining the contribution of the exogenous predictors in the model 
(Social Environment, Social Relationships, Demographics) and then discussing the 
effects of the endogenous predictors (Civic Attitude and Openness).  
 
 
94 
 
Social Environment and GPA 
 The SEM analyses provided many illustrative results in regards to predictors and 
mediators of the Wisdom construct, the foremost of which is the disproportionate 
contributions made by social environment and social relationships.  Ardelt (2000) 
reported that one’s social environment in early adulthood was the only significant 
predictor of Wisdom in later adulthood.  In this study, however, social environment 
played no direct role in the Wisdom of young adults.   
 It should be noted that, while conceptually similar, the measures used in this study 
were not identical to those used by Ardelt.  The main reason for this difference is the age 
of the sample.  Ardelt’s study was conducted retrospectively and used measures of 
familial Socio-Economic Status (SES) in childhood (i.e. education and occupation of 
parents), as well as the education and occupation of the participants in early adulthood.  
The current study also included measures of parents’ education.  However, given that the 
participants in the current study were still in school, their future environment was 
measured only by their educational intent.  While these measures tend to emphasize the 
“intellectual” environment, this domain is strongly related to other domains, including 
SES (Schneewind, 1995), which impact the quality of a child’s overall environment (i.e. 
school, neighborhood, activities, etc.).   
 As shown in Figure 4.3, one’s social environment does not account for a 
significant amount of unique variance in any of the three Wisdom domains.  In fact, GPA 
was the only item in the model uniquely predicted by one’s social environment.  Much 
research has been done to determine the predictors of student GPA, and SES is a 
consistent predictor in all of these studies (Robbins, Lauver, Le, Davis, Langley & 
95 
 
Carlstrom, 2004).  Such results seem to support the concepts of “cultural capital” and 
“human capital” in regards to educational achievement (Gordon, Bridglall, & Meroe, 
2005).  Essentially, those who come from a family that “speaks the language” of higher 
education and whose normal social networks include intelligent, well-educated 
individuals, tend to excel in the culture of higher education.  This educational 
achievement gap seems to be a self-perpetuating cycle, and has been found to exist in 
early childhood education and to grow exponentially as students age (Scales, 
Roehlkepartain, Neal, Kelsmeier, & Benson, 2006).   
 Also noteworthy is the fact that social environment was the only unique predictor 
of GPA, accounting for 9% of the variance.  While nine percent is not a large portion of 
variance, it is remarkable that nearly one-tenth of the variance in GPA can be accounted 
for by a non-malleable predictor.  Social relationships did not have a significant effect on 
GPA, but it should be noted that the measure of relationships in this study only included 
quantitative measures (i.e. number of close friends and hours spent in clubs).  These 
items may not capture the quality of these relationships or the values of those closest to 
the participant.  It is conceivable, for instance, that the GPA of those in one’s social 
network could predict one’s own GPA.   
 GPA, in turn, emerged only as a unique predictor of the Cognitive domain of 
Wisdom.  This finding validates the concerns of those who assert that traditional 
education over-emphasizes cognition while neglecting other important developmental 
assets (Gardner, 1983; Goleman, 1995; Sternberg, 2003).  GPA is often treated as a 
stand-alone measure of academic performance and has been shown to be a good predictor 
of future success in higher education (Robbins et al. 2004).  Though GPA is not a direct 
96 
 
measure of intelligence, it is highly correlated with IQ, SAT and ACT scores, all of 
which are common indicators of cognitive competence (Coyle & Pillow, 2008).  Several 
studies have linked other assets to academic achievement, including measures of 
emotional intelligence (Parker, et. al, 2002; Pritchard, 2003), and personality traits 
(Ridgell, 2004; Laidra, Pullman, & Allik, 2006), but cognitive measures commonly 
account for the largest measure of variance in GPA (Robbins, et. al, 2004).   
 The Cognitive domain of Wisdom, while conceptually related to intelligence, 
represents a broader measure of cognition.  According to Ardelt (2003), the Cognitive 
domain includes “knowledge of the positive and negative aspects of human nature, of the 
inherent limits of knowledge and of life’s unpredictability and uncertainties” (p. 278).  
This domain includes knowledge, but also the love of knowledge and a willingness to 
exert considerable effort toward understanding phenomena on a deeper level.  Other 
measures of Wisdom which are very closely associated with intelligence have found 
stronger empirical support for intelligence as a predictor (Baltes & Smith, 1990; 
Staudinger & Pasupathi, 2003).   Staudinger and Pasupathi (2003) defined Wisdom as 
“expert knowledge in the fundamental pragmatics of life” and found that intelligence 
predicted 10% of the variance in Wisdom performance.  In the current study, GPA 
predicted only 2.4% of the variance in the Cognitive domain. GPA was moderately 
related to Openness (r=.244) and Civic Attitude (r= .266), both of which made significant 
contributions to all three Wisdom domains.  GPA, then, would appear to be the least 
comprehensive of the three measurements. 
 This study found no unique association between GPA and the Affective and 
Reflective Wisdom domains.  In an effort to understand why this association was missing, 
97 
 
a bivariate correlation was conducted between GPA and the three Wisdom domains.  
Results indicate that GPA is related to the Affective and Reflective domains, but the 
relationship disappears when controlling for the effects of the Cognitive domain.  Thus, 
the need for incorporating broad, multi-dimensional constructs in social science research 
should be emphasized.  Studies that define Wisdom primarily as a cognitive function may 
be overlooking important related domains.  Likewise, those concerned mainly with 
affective competencies (i.e. emotional intelligence) may not recognize the inherent 
association of affect and cognition.  Therefore, a long-term view of human development 
could require the balance and integration of multiple domains. 
 The predictive power of GPA and other measures of cognitive performance in 
regards to success in adulthood are dubious at best.  The classic meta-analysis conducted 
by Bretz (1989) reported that GPA had no significant effect on success in adulthood; a 
result that still lacks a substantial rebuttal.  Goleman (1995) found that IQ accounted for 
approximately 4-10% of career success, and non-cognitive indicators often outperform 
intelligence in regards to leadership ability (Maulding, 2002) and life satisfaction 
(Livingstone & Day, 2005).  If the goal of education is the improvement of experience 
(Dewey, 1938) and long-term well-being for individuals and the community, then it 
seems prudent to incorporate educational objectives and methods that ensure the 
cultivation of all relevant assets.   
Social Relationships 
 The current status of one’s formal and informal social relationships had a 
significant influence on all three Wisdom domains and on one’s civic attitude with all 
significant predictors in the model.  This supports previous research (Bailey & Russell, 
98 
 
2008a; 2008b) that was conducted using similar measures.  The items used to assess 
one’s social relationships addressed the level of involvement in voluntary social clubs, 
attendance at religious ceremonies, number of close friends, and average amount of time 
spent hanging out with friends at home and in public places.  These were measured by 
asking participants to indicate their level of involvement in each activity during the last 
12 months. 
 It is logical that those involved in more social clubs would score higher on 
measures of civic attitude.  A high level of civic attitude suggests the belief that adults 
should give back to their community and that one can make a difference in the world.  
Previous research has found that early involvement in voluntary organizations is a good 
predictor of continued service in adulthood (Astin, et. al, 1999), and service-oriented 
activities can increase self-efficacy (Lundy, 2007) and intentions toward future civic 
engagement (McGehee, 2002).  The results of this study contribute to the evidence 
reported by previous research. 
 Social relationships accounted for approximately 2% of the variance in each 
Wisdom domain.  While this is not a striking amount of predictive power, the balanced 
contribution across all domains is noteworthy.  All other predictors in the model “favor” 
one or two of the Wisdom domains.  Given that the new psycho-philosophy emphasizes 
the need for a balanced development of multiple personal domains (Sternberg, 1998), 
social engagement may be a valuable piece to the puzzle of holistic growth.   
 These findings could be interpreted within Ardelt’s (2003) framework of Wisdom, 
which identifies Reflection as the primary domain.  Through continual engagement with 
others, one would necessarily need to negotiate different views and values in an effort to 
99 
 
work toward a common goal.  In negotiating these worldviews, an individual would need 
to delay judgment of others and attempt to understand situations from their perspective 
(reflection).  Doing this successfully would then provide insight into the lives of others, 
generating empathy (affection), and perhaps motivating the individual to investigate their 
own beliefs in order to reconcile conflicting information (cognition).   This illustration 
also supports the dialectical process promoted by experiential educators (Kolb, 1984; 
Friere, 1993) who assert the need for interactive methods of learning. 
 An increase in diverse social encounters, however, does not guarantee positive 
outcomes.  Putnam (2007), for instance, found that residents of diverse neighborhoods 
tend to “hunker down”, resulting in lower levels of trust, altruism and cooperation.  It is 
possible that a wise disposition could be a mediator that gleans positive results from 
difficult circumstances.  It is also possible that wise individuals purposefully select their 
social environments, choosing to be involved in a variety of social contexts.  It is difficult 
to determine which construct (Wisdom or Social Relationships) precedes the other when 
using covariance-based analyses, but the proposed model does show a strong fit for the 
data.  
Demographics 
 Females scored higher on measures of openness, affection and civic attitude, 
accounting for approximately 2% of the variance in each variable.  Those who score 
highly on the Openness to Experience subscale of the NEO-PI are described as 
imaginative, curious, creative, and liberal (Costa & McCrae, 1985).  Results from 
previous studies regarding gender differences on the NEO-PI are mixed (Lippa, 2005).  
Additionally, few studies have reported significant gender differences specifically for the 
100 
 
Openness sub-scale.  With little empirical support for this finding, any justification falls 
into the category of speculation.  It is not conceptually difficult, however, to imagine that 
females would be more inclined toward aesthetics and imagination.   
 Mabry (1998) reported a higher score for females upon the initial testing of the 
Civic Attitude Scale.  A previous study (Bailey & Russell, 2008) also reported a strong 
association between females, pro-social values, and the Affective domain.  Given that 
civic attitude represents the belief that adults should give back to the community, this 
construct would likely overlap considerably with pro-social values.  This relationship is 
anecdotally supported by the nature of the Pay It Forward Tour clientele, two-thirds of 
whom are typically female.  The association of females to affective domains also garners 
support from Gilligan’s (1982) feminist theory of moral development, which asserts that 
women think in terms of caring and relationships as opposed to males, who favor rules 
and justice. 
 Age was a weak, but significant predictor of openness to experience.  The 
openness subscale has been shown to increase until about age 30, at which time it tends 
to stabilize (Staudinger & Pasupathi, 2003).  Also, of interest was a significant negative 
relationship between age and social relationships (r= -.123).  Due to the nature of the 
sample, the age range was only eight years (M= 21), which does not allow for a detailed 
analysis of the contribution attributable to age.  However, previous research indicates that 
wisdom-related knowledge increases sharply between the ages of 15 and 35 (Baltes, 
Gluck & Kunzmann, 2005).  While this study utilizes a different measure of Wisdom, it is 
surprising that an eight-year age difference within this vital period of development would 
101 
 
have no direct effect on Wisdom domains.  This is perhaps another point of contention 
regarding the conceptualization of the Wisdom construct (Ardelt, 2004; Baltes, 2004).   
Civic Attitude 
 Civic attitude was measured with the use of a five-item scale developed by Mabry 
(1998).  This is the only construct in the model not directly mentioned as a predictor of 
Wisdom in previous research.  It was included based upon its theoretical relevance to the 
findings reported by Kunzmann and Baltes (2003).  In a study of wisdom and leadership, 
they reported that wise leaders prefer a cooperative and democratic management style.  
Based upon these findings, it seemed reasonable that wise individuals may also value 
cooperative efforts to solve community problems.  Additionally, civic engagement and 
Wisdom both supposedly operate for the common good (Rhoads, 1998; Baltes, Gluck & 
Kunzmann, 2005). 
 Civic attitude demonstrated a relatively strong relationship to openness in this 
sample (r= .322).  Even with this shared variance, all three Wisdom domains were 
affected by one’s civic attitude.  Though the parameter was significant, civic attitude 
accounted for a very small portion of the variance in the Cognitive domain (.4%).  This is 
likely due to the effects of other predictors in the model which, when combined, account 
for over 40% of the variance in the Cognitive domain.  Nearly 9% of the variance in the 
Affective domain was accounted for by civic attitude.  As previously discussed, this is 
likely due to their mutual pro-social content.  Those who value helping others would 
likely demonstrate compassion (affection) through the medium of service (Bailey & 
Russell, 2008a).   
102 
 
 Finally, civic attitude accounted for over 6% of the variance in the Reflective 
domain.  Reflective individuals are more likely to take responsibility for their actions and 
their life circumstances (Ardelt, 2003).  The Civic Attitude Scale measures one’s 
“attitude toward the responsibility to help others and solve societal problems” (Bringle, 
Philips, & Hudson, 2004, p. 197).  The common factor here would appear to be 
responsibility.  Thus, those who acknowledge their role in the scheme of things and 
accept the task of working to improve circumstances for themselves and others, would 
likely score higher in both domains.  These constructs may also be related to Locus of 
Control and Self-Efficacy (Bandura, 1977), since they all hinge on the individual’s belief 
that they can make a difference in the world.  Future research could help determine their 
true association. 
Openness to Experience 
 Openness to new experiences, as measured by a 12-item subscale of the NEO-PI, 
was the strongest predictor of overall Wisdom in the current study.  Over 30% of the 
variance in the Cognitive domain was accounted for by openness to experience.  Previous 
research has identified openness as an antecedent to wisdom-related performance in 
adolescence (Staudinger & Pasupthi, 2003).  However, in this previous study, personality 
accounted for only 5% of the variance, while crystallized intelligence accounted for 9%.  
The methods utilized by Staudinger and Pasupathi (2003) emphasize cognition, which is 
evident in their definition of wisdom as “expert knowledge”.  In the current study, 
variance in the Affective (7%) and Reflective (8.5%) domains was also attributable to 
openness.   
103 
 
 The association of openness with measures of cognition is not new in the 
research.  Previous studies have linked openness with general knowledge and intelligence 
(Furnham, Swami, Arteche, & Chamorro-Preemuzic, 2008; Gignac, 2005).  Other 
research has linked openness to self-efficacy in regards to creative and intellectual 
pursuits (Hartman & Betz, 2007).  This relationship is likely due to a natural curiosity 
exhibited by these individuals, which motivates them to go beyond established norms and 
discover new ideas.  This seems logical, given those who are open to experiencing new 
things would have a broader reserve of experience to draw upon when faced with 
complex dilemmas.   
 Openness, however, has also been found to hinder emotional regulation and 
increase the likelihood of job burnout (Ghorpade, Lackritz, & Singh, 2007).  In addition, 
those who score high in openness may be predisposed to risky behaviors such as alcohol 
and drug abuse (Flory, Lynam, & Milich, 2002).  These activities would appear to be 
opposed to any conceptualization of Wisdom.  Emotional burnout, for instance, is akin to 
the “helpless helper” syndrome described by Ardelt (2003).  This refers to those who give 
unheedingly until they are no longer able to care for others or for themselves.  A linear 
increase in openness, then, would not likely lead to an equivalent increase in Wisdom.  
Therefore, it is suggested that openness must be balanced with other assets to ensure 
holistic development.  
 The strong association of openness to the three Wisdom domains demands serious 
reflection.  Other researchers (Webster, 2003) have included openness as a core Wisdom 
domain.  This line of reasoning is not unwarranted, as life experience is central to most 
definitions of Wisdom (Birren & Fisher, 1990).  Individuals who seek new and exciting 
104 
 
things are likely to amass quite a collection of experiences to draw upon.  Experience, 
though, may be a “necessary but not sufficient” factor in the development of Wisdom.  
Webster (2003) contends that these experiences must be challenging and morally 
profound.  Even so, a wise interpretation of these events is not guaranteed.  An individual 
with low self-efficacy, for example, may attribute suffering to bad luck and become 
resigned to the fact that life is against him/her.  Future research will be needed to 
determine which aspects of openness are central to Wisdom.  As measured by the NEO-
PI, openness is assessed with 40 items and includes six sub-domains: Fantasy, Aesthetics, 
Feelings, Actions, Ideas, and Values.  Given that this study used an abbreviated version 
of the Openness scale (12 items) these sub-domains could not be accurately measured.  It 
is likely that certain items in the Openness scale overlap considerably with items in the 
3DWS (Ardelt, 2003). 
Longitudinal Conclusions 
 Previous studies have reported significant increases in Wisdom domains for 
participants of the Pay It Forward Tour (Bailey & Russell, 2008a; 2008b).  These studies 
included only two assessments, immediately before and after the service-learning 
experience.  There were no follow-up measures, nor was there a control group for either 
of the previous studies.  The primary focus of the current study was to determine the 
long-term effects of the PIFT when compared with the growth trajectories of comparable 
non-participants.  In addition, this study investigated the effects of predictors on the 
growth trajectories of the PIFT participants over the course of five weeks.  Presented first 
is a discussion of the difference in growth patterns for the treatment and control groups.  
105 
 
This is followed by a discussion of participant growth trajectories and the variables 
identified as salient for positive growth in Wisdom, Openness, and Civic Attitude. 
Treatment and Control Group Comparison 
 As measured by the 3DWS, Wisdom is defined as the integration of affective, 
cognitive, and reflective personal traits (Ardelt, 2003).  There is much debate about 
constancy and change in personality research.  Freudian theorists argue that personality 
becomes set in early childhood (Fenichel, 1945), while others attest to its malleability 
even in the latest stages of life (Field & Millsap, 1991).  The prevailing opinion, based on 
a variety of empirical studies, is that personality traits are relatively flexible until 
advanced adulthood (about age 50), at which time they tend to stabilize (Caspi & 
Roberts, 1999).  The participants in this study are well below this presumed age of 
stabilization, lending support to the potential for change. 
 In addition, researchers have identified several mechanisms that may induce 
personality change, including: Environmental Contingencies (i.e. social punishments or 
reinforcements), Reflection on one’s own actions, Modeling by respected peers or 
leaders, and Feedback from others (Caspi & Roberts, 1999).  Reflection, modeling, and 
feedback are inherent components of the Pay It Forward Tour which are intentionally 
facilitated on a regular basis throughout the experience.  Environmental contingencies 
include new social roles and new social environments that would require adaptation.  A 
service-oriented experience incorporating a diverse group of students would seem to meet 
these criteria.  Given the age of this sample and the nature of the PIFT experience, it is 
logical to presume that a significant change in personal attributes could occur. 
106 
 
 Growth Toward Wisdom.  The two groups demonstrated different growth 
trajectories for two of the three Wisdom domains.  Scores for affection and reflection 
increased for the treatment group but decreased for the control group.  Both groups 
reported decreases in cognition over the course of five weeks.  There was no difference in 
results for control group participants based on their primary Spring Break activity (i.e., 
work, travel, rest).  No control group members indicated service activities as their 
primary activity over the break.  It should be noted that all three domains increased 
significantly from pre-test to post-test measurement for the treatment group.  The decline 
in cognition for the treatment group, then, would appear to be the result of circumstances 
encountered after the PIFT.   
 It is unclear what would have caused such a decrease in cognition for all students 
during this time period.  One explanation is that the first test was conducted immediately 
before Spring Break, while the follow-up measure was administered immediately before 
final exams.  Given that the Cognitive domain assesses one’s willingness to expend 
significant effort to understand phenomena on a deep level, this could be indicative of a 
decline in intellectual enthusiasm.  Theorists have criticized traditional methods of 
education which utilize extrinsic forms of motivation (i.e. grades on a test) to coax 
students into rote memorization (Atherton, 2005).  Such methods often invoke feelings of 
drudgery, the effects of which could be a decline in cognitive functioning. 
 The control group also exhibited substantial declines in affection and reflection.  
These findings are consistent with the results of the SEM analysis reported previously.  
Researchers have asserted for decades that traditional forms of education neglect the 
development of domains other than cognition (Gardner, 1989; Goleman, 1995).  The 
107 
 
SEM analysis verifies the lack of influence GPA on other domains, demonstrating that 
those with higher GPAs do not necessarily display more compassion or reflectivity.  
These longitudinal findings indicate that the circumstances a typical student encounters 
during the college semester may inhibit the growth of Wisdom.   
 The actual cause of these declines cannot be ascertained from the variables 
assessed in this study, but these results do merit further investigation.  Lundy (2007) 
reported similar results for students enrolled in an introductory Psychology course.  
Students assigned to a service-learning project demonstrated higher test scores, and 
greater empathy than students who were assigned an interview or a report project.  Those 
not in the service-learning group exhibited decreases in empathy scores.  The service 
projects for Lundy’s study required weekly engagements with community organizations 
(2 hours each) for the duration of the semester.  The interview assignment was essentially 
a qualitative research assignment requiring three, one-hour interviews to be conducted on 
individuals across the lifespan.  The interview and service projects both required personal 
interaction and reflective analysis.  Lundy offered no clear explanation for the divergent 
group trajectories, but it would appear that empathy could have been moderated by the 
duration and intent of the personal contact.   It is possible, for instance, that the nature of 
a service-oriented assignment could affect the attitude with which the student approaches 
the task.  Conducting an assignment in the “spirit of service” could influence the 
students’ perspectives.  This servant perspective would not likely be invoked through 
interview or archive-based research. 
 Additional insight can be gleaned from the post-trip instrument completed by the 
trip participants.  The supplemental survey designed with the input of STLF leaders asked 
108 
 
participants to rate each trip component as to its importance for development in affection, 
cognition, and reflection.  The “service activities” and “relationships” were rated highest 
across all three domains.  Services activities included helping others, working toward a 
common goal, and taking charge of work projects.  Relationships included informal 
conversations with diverse participants and community members.  This combination of 
serving others and interacting with diverse groups may be a recipe for stimulating 
progress toward Wisdom.   
 The average college student would likely interact with a variety of individuals in 
and out of the classroom.  This interaction, though, would be more akin to “bonding” 
social capital, which implies a strengthening of the bonds within one’s chosen social 
circles (Putnam, 2003).  Even if these students are assigned to work groups with 
individuals they particularly dislike, they only have to endure the interaction for a short 
period of time.  The Pay It Forward Tour places individuals in a social environment 
where intense interactions are the norm.  As expressed by one participant, “For the most 
part, you cannot say anything you want while on the trip because you have to live with 
these people for 8 days and you want to make sure that you do not step on any toes.”  In 
these situations, there is little choice but to address relational issues.  This negotiation of 
diverse values and opinions, more akin to “bridging” social capital (Putnam, 2003), may 
represent one of the key differences in the types of social engagement encountered by 
PIFT participants and the average college student. 
 As previously mentioned, significant increases were reported for all domains by 
PIFT participants immediately after the trip.  While the Affective and Cognitive domains 
scores dropped down to previous levels within a month of the experience, gains in the 
109 
 
Reflective domain remained constant.  Participants rated relationships as the strongest 
contributor to growth in reflection.  These relationships often continue well beyond the 
bounds of the trip experience.  One participant noted, “I meet up with people from my 
bus all of the time and definitely have made friends who are going to stay with me for a 
long time”.  Similar comments were expressed by many PIFT participants when asked 
about tangible outcomes of the trip one month after the experience.  These relationships, 
which are founded upon the common bond of service, would represent another key 
difference in the social milieu of trip participants as opposed to that of the typical college 
student.  This assertion is consistent with the results of the SEM analysis, which 
identified formal social relationships as a key indicator of all three Wisdom domains.  A 
change in quality and quantity of social engagement could explain a portion of the 
variance in retained outcomes. 
 Civic Attitude and Openness.  Both groups reported increases in Civic Attitude 
over the course of five weeks.  However, the gains reported by trip participants were 
three times as strong as those reported by the control group.  Increased civic values are a 
central goal expressed by most service-learning programs (Mabry, 1998), and are a key 
component of the STLF mission.  Participants commonly mentioned civic outcomes as 
the most important thing they learned from the trip experience.    A few examples of 
these perceived learning outcomes include: “Helping others can make a huge difference,” 
“Learning how rewarding it is to help other people,” and “The importance of giving 
back.”  These responses give support to the quantitative findings, which indicate that 
PIFT participants experienced positive changes in civic attitudes, values, and efficacy 
that endured for at least one month after the trip.   
110 
 
 Participants were not specifically asked which trip components contributed most 
to their growth in civic attitudes or openness, but the feedback from open-ended 
questions on the final instrument provides some insight.  Those who mentioned civic-
related outcomes as the most important thing learned seemed to attribute this learning to a 
heightened awareness and sense of efficacy in serving others.  One comment exemplifies 
this idea exceptionally well:  “I have done many volunteering activities throughout my 
life, but it wasn’t until we as an entire group pulled out 19 tons of garbage out of the 
creek in Washington D.C. that I realized how my individual actions helped contribute to 
the cause.  It was sort of a revelation, and now it is easy to see how I can and will change 
the world through my actions.”  Similar comments echoed this response, describing a 
new awareness of how small actions can make a huge difference and expressing a 
commitment to serve in the future.  
 Openness scores also increased for PIFT participants while decreasing for the 
control group.  This finding is of particular interest because of the high correlation (r= 
.552) between openness and cognition.  While cognitive scores decreased for both 
groups, the group trends for openness were divergent.  This indicates that, though 
openness predicts 30% of the variance in the Cognitive domain, the two constructs are 
unique enough to show opposing trends.  It would appear that something within the PIFT 
experience, or perhaps a lifestyle change after the experience, spurs growth in the unique 
portion of openness without contributing to growth in cognition.   It is difficult to 
determine what that unique portion may represent from the data collected for this study, 
but the open-ended responses gathered on the final instrument shed some light on the 
subject. 
111 
 
 In addition to intellectual curiosity, those who are open to new experiences also 
tend to be more trusting, more adventuresome, and more mindful (Flory, Lynam, & 
Milich, 2002).  An increase in these characteristics might be demonstrative of an increase 
in openness without an increase in cognition.  Several responses were given that represent 
characteristics of openness as the most important thing learned on the trip.  Some of these 
responses included: “How to be more open and outgoing with others,” “to be okay with 
being in new, uncomfortable situations,” and “to listen and pay attention”.  When 
describing how they learned these things, participants stated for example:  “the silly 
games as well as the expressive reflection activities aided in teaching me about being 
‘comfortable with being uncomfortable’,” and “I met some amazing people on the trip 
and to not open up myself to them and offer my friendship risking pride and 
embarrassment I would be unable to experience these wonderful friendships forever,”.  
One final quote clearly expresses this growth in openness: “The spontaneous events on 
the tour taught me that small obstacles or unplanned events aren’t always bad.  I normally 
planned my life out in lists, and now I trust others more and understand that a missed turn 
on the interstate is not a tragedy but instead an opportunity for an adventure and running 
into a great, small town grocery store.”   
 An increase in trust, adventurousness, and mindfulness would not necessarily 
affect the Cognitive domain.  The mechanism of learning is not easy to identify, as 
respondents attributed their increases to a variety of experiences.  It is clear, though, that 
these increases were attributed to experiences, encounters, or social norms which are a 
part of any typical PIFT.  Whether openness was learned from games, reflection 
activities, conversations on the bus, or logistical obstacles, the opportunity for growth 
112 
 
was provided by the novel experiences in a novel social environment.  These 
opportunities may not be provided in the course of a normal school semester, but are 
central to experiential programs (Walsh & Golins, 1976; Lakin & Mahoney, 2006) 
 In sum, the treatment group demonstrated marked improvement for three of the 
five outcome measures, while the control group decreased in all but one measure.  
Though it is difficult to attribute these outcomes solely to the influence of the PIFT, this 
study provides strong evidence for the positive influence of the service-learning program.  
Participants reported lasting, significant increases in Reflection, Openness, and Civic 
attitudes after the trip experience.  Future research could help determine the reason that 
Cognitive and Affective domains did not remain significant for the treatment group, and 
why a control group of comparable college students would report declines in Wisdom and 
Openness.   
Latent Growth Analysis (LGA). 
 Having measured the PIFT participants on three occasions, it was possible to 
conduct a LGA to determine which predictors were most influential in growth trajectories 
for Wisdom, Civic Attitudes, and Openness over the course of five weeks.  The 
relationship of the predictors to the “Intercept” variable is equivalent to a cross-sectional 
analysis, which has already been discussed using the SEM analysis.  This section 
addresses the influence of each predictor on the “Slope” variable and the relationship of 
the Intercept and Slope for each dependent variable.  
 Upon inspection of Table 4.21, one can observe that, in general, those who began 
with a higher score for each independent variable also demonstrated lower growth 
trajectories for each outcome variable.  Social relationships, for example, were positively 
113 
 
related with the intercept for the Affective domain.  This indicates that those who began 
the program with higher levels of social interaction also reported higher levels of 
affection.  However, social relationships were negatively related to the slope for the 
Affective domain, indicating that those who began with higher levels of social interaction 
demonstrated lower levels of growth in affection.  This phenomenon may be referred to 
as a “regression to the mean”; a leveling of the playing field that is not uncommon in 
educational interventions, and is often the expressed intent of such programs.  This is the 
case for most predictors in Table 4.21, as illustrated by a positive relationship to the 
Intercept and a negative relationship to the Slope of each dependent variable (or visa 
versa).  
 Of particular interest is the result of Personal Reflection habits.  This predictor 
was measured with a single item which asked the participant how often they take 
intentional time for personal reflection on a weekly basis.  Those with the highest scores 
take time out daily to journal, meditate, or participate in some other type of quiet 
reflection.  This predictor was included in the model as an indicator of one’s tendency 
toward life review, a characteristic identified as central to wisdom-related performance 
(Baltes & Smith, 1990).   
  Those who spent more time in Personal Reflection on a regular basis reported 
higher initial scores in all Wisdom domains, though this relationship only achieved 
marginal significance.  Those with higher levels of personal reflection also demonstrated 
stronger growth in Wisdom over the course of five weeks.  There seems to be something 
unique about those who set aside daily reflection time that generates Wisdom in daily life 
and enables them to maintain the outcomes of significant experiences for a longer time.  
114 
 
In addition, Personal Reflection was positively related to the initial status and growth rate 
of participants for Openness to Experience.  Gender and Personal Reflection time were 
the only significant predictors of Wisdom and Openness, and accounted for a substantial 
amount of variance in the slope for each variable (63.4%, and 19% respectively). 
 This could have powerful implications in a world where time is becoming a 
scarce commodity (DeGraaf, 2003).  Educators and programmers should seriously 
consider how to encourage the simple activity of daily reflection.  The importance of 
reflection upon experience has been heralded by experiential educators (Dewey, 1938; 
Friere, 1993; Kolb, 1984), but reflection is often described as a very abstract mental 
technique and illustrated by all manner of complicated models.  It may be more beneficial 
to encourage the simple habit of reflecting daily, by any means, in order to cultivate a 
thoughtful awareness of oneself and their experience. 
 Demographic effects could be important in regards to future program design.  
There was no effect for age on any of the dependent variables in the current study.  The 
PIFT was equally effective for 18 year old participants as it was for 26 year olds.  Males 
demonstrated a higher level of growth in all outcomes, having started with lower scores 
than females.  It is unclear why males would report stronger levels of growth over the 
course of five weeks, but future research could explore the difference in outcomes due to 
gender.  Finally, those from a less supportive social environment demonstrated stronger 
growth in civic attitudes, perhaps an indicator of higher levels of efficacy as a result of 
seeing the difference a small group of students can make.     
115 
 
Conclusions and Implications 
 This study presents substantial evidence for the positive effects of a service-
learning tour designed as an alternative Spring Break experience for college students.  
The trip participants reported increases in overall Wisdom scores and in measures of 
Openness and Civic Attitude immediately after the experience.  These increases remained 
significant one month after the program.  A comparable group of college students who 
did not participate in such an experience demonstrated decreases in all variables with the 
exception of Civic Attitude.   
 These findings have important implications for the fields of education and 
recreation.  Experiential methods of education are often referred to as “supplemental” 
education (Cremin, 1990), implying that they are of secondary importance to formal 
methods.  As such, out-of-school programming may be supported “secondarily” in 
regards to policy and funding decisions.  In light of the psychological reform (i.e. positive 
psychology and prevention science) perhaps it is time to reconsider the aims of education 
as well.  Positive psychology requires a conceptualization of ideal human functioning to 
serve as a guide for optimal human development (Seligman & Csikszentmihalyi, 2000).  
No author has conceptualized ideal functioning as intelligence alone.  It is unfortunate, 
then, that public education is still driven mainly by cognitive-based assessments.  
Concerns over cognitive achievement gaps should be echoed by equal concerns over a 
lack of compassion and reflectivity among modern youth and emerging adults.   
 The techniques of experiential programs such as the PIFT should be valued and 
implemented into the educational mainstream.  Many colleges are beginning to include a 
community-service requirement for all degree programs.  This is a step in the right 
116 
 
direction.  However, community-service is often mandated on top of all other 
requirements, increasing the likelihood that it will be completed begrudgingly.  
Industrialized cultures advance at increasingly rapid rates, and there is no question that 
students must be brought up to speed on the latest discoveries.  The purpose of this study 
is by no means to undermine the value of formal education.  However, in light of the 
results of this research and of the many other studies described, it seems appropriate to 
include experiential education as a viable option, or even a substitute, for certain 
traditional courses.    
 Despite mounting evidence of the efficacy of experiential programs, educational 
policy-makers remain reluctant to mainstream these techniques.  Much of the problem 
may be attributed to a fragmented message and a lack of practical assessments.  The 
experiential programmer’s voice may be diffused by inefficient use of language.  The 
effects of experiential programs have been documented extensively, but the outcomes are 
often reported as a loosely connected set of assets that may or may not add up to positive 
long-term development.  Self-esteem and optimism are two examples of positive assets 
that are currently being critically investigated (Twenge, 2007; Puri & Robinson, 2007).  
Researchers assert that, though these are desirable traits, a linear increase in these traits 
alone could lead to negative outcomes such as narcissism and inappropriate risk-taking.  
The current emphasis placed upon evidence-based practice may exacerbate the problem 
by encouraging programs to report findings while neglecting to determine how these 
findings contribute to a long-term vision of human development. 
  Broader, multi-dimensional outcome measures may help to address some of these 
concerns.  Wisdom, for example, incorporates a broad set of assets that provide an 
117 
 
intrapersonal system of “checks and balances”.  An increase in cognition should be 
balanced by increases in other assets, such as affection and reflection.  The sub-domains 
of Wisdom are broad constructs as well, under which a variety of positive assets may be 
aligned.  As such, Wisdom could provide a balanced structural framework through which 
many other assets can be organized.  In this way, the fragmented list of outcomes 
associated with experiential education can be integrated into a framework of positive 
human development, and these outcomes can be substantively compared to other methods 
of education.   
 This united voice could provide the motivation to reconceptualize the aims of 
education.  Then the means of education could be informed by a clear picture of the 
desired end; an end measured by individual and social well-being, not merely by 
cognitive-based tests.  In the words of Peter F. Drucker, “There is nothing so useless as 
doing efficiently that which should not be done at all.”  Efficient methods of education do 
little good without a vision of their purpose.  Future research should be done to determine 
how published outcomes fit within this framework of human development.   
  Another barrier to this reconceptulization is the problem of assessing positive, 
long-term human development.  This study provides some insight into one possible 
method of measuring growth in critical domains.  Other methods of evaluating Wisdom 
involve highly controlled, lab-based experiments that are of limited valued to the greater 
public (Baltes & Staudinger, 1996; Staudinger & Pasupathi, 2003).  The results of this 
study verify the use of self-report measures by confirming the findings of those highly-
controlled studies.  Intelligence and openness to experience, for instance, emerged as 
important predictors of Wisdom in adolescents, substantiating the results reported by 
118 
 
Staudinger and Pasupathi (2003).  Future research could be done to determine the validity 
of repeated self-report measures, and investigate the use of other minimally-intrusive 
methods of assessing growth toward Wisdom.   
 This study also utilizes a new technique for the analysis of the Wisdom construct.  
Previous research has utilized linear regression, analysis of variance, and structural 
equation modeling to determine differences in Wisdom traits and performance (Baltes & 
Staudinger, 1996; Ardelt, 2000; Webster, 2003).  These analyses are mean-based, 
utilizing group averages to determine growth trajectories.  In addition, they do not 
account for dynamic growth within multi-dimensional constructs (Kline, 2005).  Using 
the technique of Latent Growth Analysis allows for the observation of factor-shifting 
within the Wisdom construct over time and enables the analysis of predictors on 
individual growth trajectories.  It was this technique that revealed the importance of 
personal reflection time for the growth of Wisdom and for the retaining of educational 
outcomes.  This type of analysis will become increasingly important as the use of multi-
dimensional constructs proliferates in the social sciences.  Future research could be done 
to determine other predictors of growth in Wisdom and to gain a better understanding of 
the dynamics within the Wisdom construct.  For example, one could investigate if 
reflection is truly the exclusive hub of Wisdom (Ardelt, 2003), or if other domains can be 
primary in various contexts. 
 Future research is needed to verify the comparative influence of one’s social 
environment and social relationships on Wisdom.  While the association of social capital 
and Wisdom has been established, the measures used have only assessed quantities of 
social engagement (Bailey & Russell, 2008a; 2008b).  It still remains to be determined 
119 
 
which types of engagement are most salient for each domain, and whether the quality of 
that engagement is a factor.  As the measures of social environment in this study mainly 
dealt with intellectual support, future research could determine if there are other factors in 
one’s family environment that influence Wisdom domains.  These other factors might 
include quality of family communication, daily routine (work & leisure habits), family 
structure (i.e. broken families, siblings), and the social habits of one’s parents. 
 There is much work to be done regarding the conceptualization and measurement 
of Wisdom, and the promise of experiential education in fostering its development.  This 
study provides strong evidence for the cultivation of Wisdom with the use of a service-
learning experience.  These results, though, are far from being the final word.  New 
techniques will surely arise that prove more appropriate, and the definition of Wisdom 
will not likely go unchanged.  It is vital that the search go on.  The measurement of 
intangible constructs is no easy task, but neglecting to incorporate them into the “model” 
is much more hazardous.  A future of meaning and fulfillment can only come through the 
continual search for and discovery of the truth behind the experience.  “As to the 
causation of the feeling of meaninglessness, one may say…that people have enough to 
live by but nothing to live for; they have the means but no meaning (Frankl, 2006, 
p.140).”  To provoke this search is perhaps an educator’s greatest task.   
“/o matter how ridiculous, the attempt to bridge between science and 
religion, between human greed and nature’s need, between what is and 
what ought to be, has to go on if we are to make sense of what is 
happening in such a way that humanity is to survive.”  
 (Csikszentmihalyi & Rathunde, 1990, p.43) 
 
120 
 
APPENDIX A. PARTICIPANT CONSENT FORM 
 
 By filling out this survey you are accepting the invitation to be involved in a 
research study on the Pay It Forward Tour conducted by Keith C. Russell, Ph.D., and 
Andrew Bailey, M.S., at the University of Minnesota.  Please read this form and direct 
any questions to the STLF leaders or call or email the investigators before completing the 
survey. 
 
Background Information:  The purpose of this study is to evaluate the psycho-social 
benefits associated with participation in the Pay It Forward Tour (PIFT).   
 
Procedures:  If you agree to be involved in this study, we would ask that you fill out this 
questionnaire before and after the PIFT.  These surveys will require approximately 20 
minutes to complete. 
 
Risks and Benefits of being in the study:  This study includes personal questions that may 
make some participants feel uncomfortable.  Participation in this study is entirely 
voluntary and you are free to neglect certain questions or terminate the study at any time. 
 
Examples of a personal question include: 
1. When I look back on what has happened to me, I can’t help feeling resentful. 
2. I often have not comforted another when he or she needed it. 
 
There are no direct benefits to participation in this study. 
 
Confidentiality:  The records of this study will be kept private and any reports that may 
be published will not include information that would make it possible to identify a 
participant.  Research records will be stored securely and only researchers will have 
access to them.  Participants will be identified by a random number identification and the 
names of subjects will not be attached to the identification number. 
 
Voluntary Nature of the Study:  Participation in this study is voluntary.  Refusal to 
participate will not affect your current relationship with the University of Minnesota or 
the STLF.  You may neglect certain questions or withdraw participation at any time. 
 
Contacts and Questions:  The researchers conducting this study are: Dr. Keith C. Russell 
and Andrew Bailey.  If you have any questions you are encouraged to contact them at the 
University of Minnesota, (612-626-4280). krussell@umn.edu, or baile270@umn.edu.  
 
Statement of Consent: 
I have read the above information and I consent to participate in this study. 
 
Name (Please Print clearly):_______________________________ Date:____________ 
121 
 
APPENDIX B.  PRE-TRIP QUESTIONNAIRE 
 
 
Please Circle the appropriate answer below. 
 
1. What is the highest level of education that your father obtained? 
 A) High School     B) Some College     C) B.S/B.A     D) Master’s    E) PhD or other 
 
2. What level of education did your mother obtain? 
 A) High School     B) Some College     C) B.S/B.A     D) Master’s    E) PhD or other 
 
3. In general, how many intellectual interests would you say your parents have? 
 A) Few   B) Some      C) Many       D) Very Many 
 
4. What level of education do you plan on obtaining? 
 A) Some college      B) B.S./ B.A.     C) Master’s      D) PhD or other terminal degree 
 
5. How many non-mandatory clubs/ groups/ organizations are you involved in?  
A) None          B) 1-2  C)  3-4            D)  5-6             E)  more than 6 
 
6. How many people in your life would you consider to be very important to you? 
 A)  0-1    B) 2-4 C) 5-8 D)  9-12 E) More than 12 
 
7. How many times in the past 12 months have you? 
 
 
 

 Volunteered?  
 
      A             B             C            D             E             F 

 Attended religious services?       A             B             C            D             E             F 

 Had friends over to your 
home?  
 
      A             B             C            D             E             F 

 Hung out with friends in a 
public place?  
 
      A             B             C            D             E             F 
To what extent do you agree or disagree with the following statements: 
 
 
 
 
 
Never       Once         2-4         5-9        10-15      >15 
                                  Times    Times    Times    Times 
122 
 
1. _____Adults should give some time for the good of their community or country. 
2. _____People, regardless of whether they’ve been successful or not, ought to help others. 
3. _____Individuals have a responsibility to help solve our social problems. 
4. _____I feel that I can make a difference in the world. 
5. _____It is important to help others even if you don’t get paid for it. 
6. _____In this complicated world of ours the only way we can know what’s going on is to 
                 rely on leaders or experts who can be trusted. 
7. _____I am annoyed by unhappy people who just feel sorry for themselves. 
8. _____Life is basically the same most of the time. 
9. _____People make too much of the feelings and sensitivity of animals. 
10. _____You can classify almost all people as either honest or crooked. 
11. _____I would feel much better if my present circumstances changed. 
12. _____There is only one right way to do anything. 
13. _____There are some people I know I would never like. 
14. _____It is better not to know too much about things that cannot be changed. 
15. _____Things often go wrong for me by no fault of my own. 
16. _____Ignorance is bliss. 
17. _____I can be comfortable with all kinds of people. 
18. _____A person either knows the answer to a question or he/she doesn’t. 
19. _____It’s not really my problem if others are in trouble and need help. 
20. _____People are either good or bad. 
21. _____I try to look at everybody’s side of a disagreement before I make a decision. 
22. _____If I see people in need, I try to help them one way or another. 
23. _____When I’m upset at someone, I usually try to “put myself in his or her shoes” for a 
                 while. 
24. _____There are certain people whom I dislike so much that I am inwardly pleased when 
                 they are caught and punished for something they have done. 
25. _____I always try to look at all sides of a problem. 
26. _____Sometimes I feel a real compassion for everyone. 
27. _____I try to anticipate and avoid situations where there is a likely chance I will have to 
                 think in depth about something. 
28. _____When I look back on what has happened to me, I can’t help feeling resentful. 
29. _____I often have not comforted another when he or she needed it. 
30. _____A problem has little attraction for me if I don’t think it has a solution. 
          1                            2                              3                             4                           5 
    Strongly                  Agree                     Neutral                   Disagree               Strongly  
     Agree                                                                                                           Disagree 
123 
 
 
 
31. _____I either get very angry or depressed if things go wrong. 
32. _____Sometimes I don’t feel very sorry for other people when they are having problems. 
33. _____I often do not understand people’s behavior. 
34. _____Sometimes I get so charged up emotionally that I am unable to consider many ways 
                 of dealing with my problems. 
35. _____Sometimes when people are talking to me, I find myself wishing that they would 
                 leave. 
36. _____I prefer just to let things happen rather than try to understand why they turned out 
                 that way. 
37. _____When I am confused by a problem, one of the first things I do is survey the  
                 situation and consider all the relevant pieces of information. 
38. _____I don’t like to get involved in listening to another person’s troubles. 
39. _____I am hesitant about making important decisions after thinking about them. 
40. _____Before criticizing somebody, I try to imagine how I would feel if I were in their 
                 place. 
41. _____I’m easily irritated by people who argue with me. 
42. _____When I look back on what’s happened to me, I feel cheated. 
43. _____Simply knowing the answer rather than understanding the reasons for the answer to 
                 a problem is fine with me. 
44. _____I sometimes find it difficult to see things from another person’s point of view.  
45. _____I don’t like to waste my time daydreaming. 
46. _____ Once I find the right way to do something, I stick to it. 
47. _____ I am intrigued by the patterns I find in art and nature. 
48. _____ I believe letting students hear controversial speakers can only confuse and  
             mislead them. 
49. _____ Poetry has little or no effect on me. 
50. _____ I often try new and foreign foods. 
51. _____ I seldom notice the moods or feelings that different environments produce. 
52. _____ I believe that we should look to our religious authorities for decisions on 
            moral issues. 
53. _____ Sometimes when I am reading poetry or looking at a work of art, I feel a  
          1                            2                              3                             4                           5 
    Strongly                  Agree                     Neutral                   Disagree               Strongly  
     Agree                                                                                                                Disagree 
124 
 
            Chill or wave of excitement. 
54. _____ I have little interest on the nature of the universe or the human condition. 
55. _____ I have a lot of intellectual curiosity. 
56. _____ I often enjoy playing with theories or abstract ideas. 
 
1. Gender.  Circle one:   M  F 
2. Age: _______ 
3. Which college are you currently attending?_______________________________ 
4. Year in school.  Please indicate Undergraduate or Graduate: 
___________________________ 
5. How many Leadership-related courses have you taken?  If you are currently 
enrolled, please count that as one course. Circle one: 
 0  1-2  3-5  6+ 
6. Have you had a significant service-related experience, (not STLF)?  Circle one: 
 YES NO 
a. If YES, please indicate the trip or program:_________________________ 
7. How many times have you traveled outside of North America?  Circle one: 
0  1-3  4-7  8+ 
8. How many non-STLF extra-curricular activities do you participate in on campus?   
0  1-3  4-7  8+ 
9. Of these activities, how many do you have leadership roles in?  Circle one: 
0  1-3  4-7  8+ 
10.  Are you an international student?  Circle one:   YES  NO 
11. Do you plan on pursuing advanced degrees?  Circle one: 
           No Way I’ll think about it  Probably  Of course 
12. Do you plan on working in a field related to education?  Circle one: 
No Way I’ll think about it  Probably  Of course 
125 
 
13. Do you plan on working in a field related to nonprofits?  Circle one: 
No Way I’ll think about it  Probably  Of course 
14. Would you like to learn more about AmeriCorps programs?  Circle one: 
 YES NO 
 
 
 
126 
 
APPENDIX C.  POST-TRIP SURVEY SUPPLEMENT 
 
The STLF program is made up of many different elements.  Please rate each of 
the elements below as to their importance in helping you grow in the following 
areas:  
1. Compassion:  Care and concern for others. 
2. Deep Understanding: A willingness to confront difficult issues and 
wrestle with them in order to achieve a deeper understanding. 
3. Introspection & Perspective-taking:  Self-awareness and the ability to 
view things from various perspectives. 
(Circle one for each column)       1= Not at all important     5= Extremely Important 
ELEMENTS 1) Compassion 
& Empathy 
2) Deep 
Understanding 
of life 
3) Introspection 
& perspective-
taking 
Diversity of Trip 
Participants 
1   2   3   4   5   1   2   3   4   5   1   2   3   4   5   
Informal Conversations 
with Participants 
1   2   3   4   5   1   2   3   4   5   1   2   3   4   5   
Conversations with 
Community Members 
1   2   3   4   5   1   2   3   4   5   1   2   3   4   5   
Ice-breakers and Games 1   2   3   4   5   1   2   3   4   5   1   2   3   4   5   
Opportunity to take charge 
of projects 
1   2   3   4   5   1   2   3   4   5   1   2   3   4   5   
Working with others for a 
common goal 
1   2   3   4   5   1   2   3   4   5   1   2   3   4   5   
Helping others in need 1   2   3   4   5   1   2   3   4   5   1   2   3   4   5   
Being away from my daily 
routine 
1   2   3   4   5   1   2   3   4   5   1   2   3   4   5   
Seeing new places 1   2   3   4   5   1   2   3   4   5   1   2   3   4   5   
Personal quiet times 1   2   3   4   5   1   2   3   4   5   1   2   3   4   5   
Group discussions 1   2   3   4   5   1   2   3   4   5   1   2   3   4   5   
Structured reflection 
activities  
1   2   3   4   5   1   2   3   4   5   1   2   3   4   5   
The direction of the STLF 
leaders 
1   2   3   4   5   1   2   3   4   5   1   2   3   4   5   
Personal discussions with 
STLF leaders 
1   2   3   4   5   1   2   3   4   5   1   2   3   4   5   
127 
 
The example set by the 
STLF leaders 
1   2   3   4   5   1   2   3   4   5   1   2   3   4   5   
Other: (Please Specify) 
 
 
1   2   3   4   5 1   2   3   4   5 1   2   3   4   5 
 
128 
 
BIBLIOGRAPHY 
Anderson, J. C., & Gerbing, D. W. (1988). Structural Equation Modeling in Practice: A 
Review and Recommended Two-Step Approach. Psychological Bulletin, 103(3), 
411-423.  
Ardelt, M. (2000). Antecedents and Effects of Wisdom in Old Age. Research on Aging, 
22(4), 360-394.  
Ardelt, M. (2003). Empirical Assessment of a Three-Dimensional Wisdom Scale. 
Research on Aging, 25(3), 275-324.  
Ardelt, M. (2004). Where Can Wisdom Be Found? Human Development, 47, 304-307.  
Ardelt, M. (2004). Wisdom as Expert Knowledge System: A Critical Review of a 
Contemporary Operationalization of an Ancient Concept. Human Development, 47, 
257-285.  
Assmann, A. (1994). Wholesome Knowledge: Concepts of Wisdom in a Historical and 
Cross-Cultural Perspective. In D. L. Featherman, R. M. Lerner & M. Perlmutter 
(Eds.), Life-Span Development and Behavior (pp. 188-222). Hillsdale, NJ: 
Lawrence Erlbaum Associates, Publishers.  
Astin, A. W., Sax, L. J., & Avalos, J. (1999). Long-Term Effects of Volunteerism During 
the Undergraduate Years. The Review of Higher Education, 22(2), 187-202.  
129 
 
Atherton, J. S. (2005). Learning and Teaching: Deep and Surface learning., 2 February 
2009-239.  
Bailey, A., & Russell, K. (2008). Psycho-Social Benefits of a Service-Learning 
Experience. Journal of Unconventional Parks, Tourism & Recreation Research, 
1(1), 9-16.  
Baltes, P. B., Gluck, J., & Kunzmann, U. (2005). Wisdom: Its Structure and Function in 
Regulating Successful life Span Development. In C. R. Snyder, & S. J. Lopez 
(Eds.), Handbook of Positive Psychology (pp. 327-347). New York, New York: 
Oxford University Press, Inc.  
Baltes, P. B., & Kunzmann, U. (2004). The Two Faces of Wisdom: Wisdom as a General 
Theory of Knowledge and Judgment about Excellence in Mind and Virtue vs. 
Wisdom as Everyday Realization in People and Products. Human Development, 
47, 290-299.  
Baltes, P. B., & Smith, J. (1990). Toward a psychology of wisdom and its ontogenesis. In 
R. J. Sternberg (Ed.), (pp. 87-120). New York, NY: Cambridge University Press.  
Baltes, P. B., & Staudinger, U. M. (2000). Wisdom: a Metaheuristic (Pragmatic) to 
Orchestrate Mind and Virtue Toward Excellence. American Psychologist, 55(1), 
122-136.  
Bandura, A. (1977). Self-efficacy: Toward a unifying theory for behavioral change. 
Psychological Review, , 191-215.  
130 
 
Batchelder, T. M., & Root, S. (1994). Effects of an undergraduate program to integrate 
academic learning and service: cognitive, prosocial cognitive, and identity 
outcomes. Journal of Adolescence, 17, 341-355.  
Berger, J. B., & Milem, J. F. (2002). The impact of community service involvement on 
three measures of undergraduate self-concept. /APSA Journal, 40(1), 85-103.  
Bollen, K. A., & Curran, P. J. (2006). Latent Curve Models: A Structural Equation 
Perspective. Hoboken, NJ: John Wiley & Sons, Inc.  
Bretz, R. D. (1989). College Grade Point Average as a Predictor of Adult Success: A 
Meta-Analytic Review and Some Additional Evidence. Public Personnel 
Management, 18(1), 11-22.  
Bringle, R., Phillips, M., & Hudson, M. (2004). The Measure of Service Learning. 
Washington, D.C.: American Psychological Association.  
Brown, S., & Greene, J. The Wisdom Development Scale: Translating the Conceptual to 
the Concrete. Journal of College Student Development, 47(1), 1-19.  
Byrne, B. M. (2001). Structural Equation Modeling with Amos. Mahwah, New Jersey: 
Lawrence Erlbaum Associates.  
Caspi, A., & Roberts, B. W. (1999). Personality Continuity and Change across the Life 
Course. In L. A. Pervin, & O. P. John (Eds.), Handbook ofpersonality: theory and 
research (2nd ed., pp. 300-326) Guilford Press.  
131 
 
Cattel, R. B. (1971). Abilities: Their structure, growth, and action. Boston, MA: 
Houghton Mifflin.  
Clayton, V. P., & Birren, J. E. (1980). The development of wisdom across the life span: a 
reexamination of an ancient topic. In P. B. Baltes, & O. G. Brim (Eds.), Life-Span 
Development and Behavior (pp. 103-135). New York: Academic Press.  
Coleman, J. (1990). Foundations of Social Theory. Cambridge, MA: Harvard UP.  
Coyle, T., & Pillow, D. (2008). SAT and ACT predict college GPA after removing g. 
Intelligence, 36(6), 719-729.  
Cremin, L. A. (1990). Public Education. New York: Basic Books, Inc.  
Creswell, J. W. (2003). Research Design. London: Sage Publications.  
Csikszentmihalyi, M. (1999). If We Are So Rich, Why Aren't We Happy? American 
Psychologist, 54(10), 821-827.  
Csikszentmihalyi, M., & Rathunde, K. (1990). The psychology of wisdom: an 
evolutionary interpretation. In R. J. Sternberg (Ed.), (pp. 25-51). New York, NY: 
Cambridge University Press.  
De Graaf, J. (2003). Take Back Your Time: Fighting overwork and time poverty in 
America. San Francisco: Berrett-Koehler Publishers, Inc.  
132 
 
Delve, C. I., Mintz, S. D., & Stewart, G. M. (1990). Promoting Values Development 
Through Community Service: A Design. /ew Directions for Student Services, 50, 
7-30.  
Dewey, J. (1916). Democracy and Education. NY: The Free Press.  
Dillman, D. A. (2000). Mail and Internet Surveys (2nd ed.). New York: John Wiley & 
Sons, Inc.  
Dyke, M. (2006). The role of the 'Other' in reflection, knowledge formation and action in 
late modernity. International Journal of Lifelong Education, 25(2), 105-123.  
Faul, F., Erdfelder, E., Lang, A. G., & Buchner, A. (2007). G*Power 3: A flexible 
statistical power analysis program for the social, behavioral, and biomedical 
sciences. Behavior Research Methods, 39, 175-191.  
Fenichel, O. (1945). The pschodynamic theory of neurosis. New York: Norton.  
Field, D., & Millsap, R. E. (1991). Personality in advanced old age: Continuity or 
Change? Journal of Gerontology, 46, 299-308.  
Flory, K., Lynam, D., & Milich, R. (2002). The relations among personality, symptoms 
of alcohol and marijuana abuse, and symptoms of comorbid psychopathology: 
Results from a community sample. Experimental and Clinical 
Psychopharmacology, 10, 425-434.  
133 
 
Flynn, J. R. (1998). WAIS-III and WISC-III gains in the United States from 1972 to 
1995: How to compensate for obsolete norms. Perceptual and Motor Skills, 86, 
1231-1239.  
Frankl, V. (2006). Man's Search for Meaning. Boston, MA: Beacon Press.  
Fredrickson, L. M., & Anderson, D. H. (1999). A Qualitative Exploration of the 
Wilderness Experience as a Source of Spiritual Inspiration. Journal of 
Environmental Psychology, 10, 21-39.  
Furnham, A., Swami, V., Arteche, A., & Chamorro-Premuzic, T. (2008). Cognitive 
Ability, Learning Approaches and Personality Correlates of General Knowledge. 
Educational Psychology, 28(4), 427-437.  
Gardner, H. (1983). Frames of Mind: The Theory of Multiple Intelligences. New York: 
Basic Books.  
Ghorpade, J., Lackritz, J., & Singh, G. (2007). Burnout and Personality: Evidence from 
Academia. Journal of Career Assessment, 15(2), 240-256.  
Gignac, G. E. (2005). Openness to Experience, General Intelligence and Crystallized 
Intelligence: A Methodological Extension. Intelligence, 33(2), 161-167.  
Gilligan, C. (1982). In a different voice: Psychological theory and women's development. 
New York: Cambridge University Press.  
134 
 
Goleman, D. (1995). Emotional Intelligence: Why It Can Matter More Than IQ. New 
York: Bantam Books.  
Gordon, E. W., Bridglall, B. L., & Meroe, A. S. (2005). Supplementary Education: The 
Hidden Curriculum of High Academic Achievement. New York: Rowman and 
Littlefield Publishers, Inc.  
Gottlieb, R. S. (2001). Spiritual Deep Ecology and World Religions: A Shared Fate, a 
Shared Task. In D. L. Barnhill, & R. S. Gottlieb (Eds.), Deep Ecology and World 
Religions (pp. 1-16). Albany, NY: State University of New York Press.  
Gray, D. E. (2005). Doing Research in the Real World. London: Sage Publications.  
Hancock, G. R., Kuo, W., & Lawrence, F. R. (2001). An Illustration of Second-Order 
Latent Growth Models. Structural Equation Modeling, 8(3), 470-489.  
Hartman, R. O., & Betz, N. E. (2007). The Five-Factor Model And Career Self-Efficacy: 
General and Domain-Specific Relationships. Journal of Career Assessment, 15(2), 
145-161.  
Hattie, J., Marsh, H. W., Neill, J. T., & Richards, G. E. (1997). Adventure education and 
outward bound: Out-of-class experiences that make a lasting difference. Review of 
Educational Research, 67(1), 43-87.  
Helson, R., & Srivastava, S. (2002). Creative and Wise People: Similarities, Differences 
and How They Develop. Personality and Social Psychology Bulletin, 28(10), 1430-
1440.  
135 
 
Holliday, S. G., & Chandler, M. J. (1986). Wisdom: Explorations in adult competence. In 
J. A. Meacham (Ed.), Contributions to human development (pp. 1-96). Basel, 
Switzerland: Karger.  
Howell, D. C. (2002). Statistical Methods for Psychology (5th ed.). Pacific Grove, CA: 
Duxbury.  
Hu, L., & Bentler, P. M. (1999). Cutoff Criteria for Fit Indexes in Covariance Structure 
Analysis: Conventional Criteria Versus New Alternatives. Structural Equation 
Modeling, 6(1), 1-55.  
Jacoby, B. (1996). Service-learning in today's higher education. In B. Jacoby (Ed.), 
Service-learing in higher education (pp. 2-25). San Francisco, CA: Jossey-Bass.  
James, W. (1984). Psychology: Briefer Course. Cambridge, Massachusetts: Harvard 
University Press.  
Jarvis, P., Halford, J., & Griffin, C. (1998). The Theory and Practice of Learning. 
Sterling, VA: Stylus Publishing, Inc,.  
Kekes, J. (1995). Moral wisdom and good lives. Ithaca, NY: Cornell University Press.  
Kerlinger, F. N., & Lee, H. B. (2000). Foundations of Behavioral Research (Fourth 
Edition ed.). New York: Harcourt College Publishers.  
Kline, R. B. (1998). Principles and Practice of Structural Equation Modeling. New 
York: The Guilford Press.  
136 
 
Kolb, D. (1984). Experiential Learning: Experience as the Source of Learning and 
Development. Englewood Cliffs, N.J.: Prentice Hall.  
Kolb, D. A., Boyatzis, R. E., & Mainemelis, C. (2000). Experiential Learning Theory: 
Previous Research and New Directions. In R. J. Stermberg, & L. F. Zhang (Eds.), 
Perspectives on cognitive learning and thinking styles (). NJ: Lawrence Erlbaum.  
Laidra, K., Pullman, H., & Allik, J. (2007). Personality and intelligence as predictors of 
academic achievement: A cross-sectional study from elementary to secondary 
school. Personality and Individual Differences, 42(3), 441-451.  
Lakin, R., & Mahoney, A. (2006). Empowering youth to change their world: Identifying 
key components of a community service program to promote positive 
development. Journal of School Psychology, 44, 513-531.  
Lippa, R. A. (2001). Gender, /ature, and /urture. New York: Routledge.  
Livingstone, H. A., & Day, A. L. (2005). Comparing the Construct and Criterion-Related 
Validity of Ability-Based and Mixed-Model Measures of Emotional Intelligence. 
Educational and Psychological Measurement, 65(5), 757-779.  
Lundy, B. L. (2007). Service-Learning in Life-Span Developmental Psychology: Higher 
Exam Scores and Increased Empathy. Teaching of Psychology, 34(1), 23-27.  
Mabry, J. B. (1998). Pedagogical variations in service-learning and student outcomes: 
How time, contact, and reflection matter. Michigan Journal of Community Service 
Learning, 5, 32-47.  
137 
 
Marcel, G. (1954). The Decline of Wisdom. London: The Harvill Press Ltd.  
Maulding, W. S. (2002). Emotional Intelligence and Successful Leadership. Annual 
Meeting of the Southern Regional Council on Educational Administration,  
Mayer, F. S., & Frantz, C. M. (2004). The connectedness to nature scale: A measure of 
individuals' feeling in community with nature. Journal of Environmental 
Psychology, 24, 503-515.  
McDonald, R. P., & Ho, M. R. (2002). Principles and Practice in Reporting Structural 
Equation Analyses. Psychological Methods, 7(1), 64-82.  
McGehee, N. (2002). Alternative Tourism and Social Movements. Annals of Tourism 
Research, 99(1), 124-143.  
McKenzie, M. (2003). Beyond "The Outward Bound Process:" Rethinking Student 
Learning. The Journal of Experiential Education, 26(1), 8-23.  
Mezirow, J. (1990). Fostering Critical Reflection in Adulthood. San Francisco, CA: Josey 
Bass.  
Moacir, G. (1994). Reading Paulo Freire: His life and work. Albany: State University of 
New York Press.  
Muthen, B. O., & Curran, P. J. (1997). General Longitudinal Modeling of Individual 
Differences in Experimental Designs: A Latent Variable Framework for Analysis 
and Power Estimation. Psychological Methods, 2(4), 371-402.  
138 
 
Pacione, M. (1999). The Relevance of Religion for a Relevant Human Geography. 
Scottish Geography Journal, 115(2), 117-131.  
Parker, J. D. A., Creque, R., Harris, J., Majeski, S. A., Wood, L. M., & Hogan, M. J. 
(2004). Academic Success in High School: Does Emotional Intelligence Matter? 
Personality and Individual Differences, 37(7), 1321-1330.  
Pascual-Leone, J. (1990). An essay on wisdom: toward organismic processes that make it 
possible. In R. J. Sternberg (Ed.), (pp. 244-279). New York, NY: Cambridge 
University Press.  
Pasupathi, M., Staudinger, U. M., & Baltes, P. B. (2001). Seeds of Wisdom: Adolescents' 
Knowledge and Judgment About Difficult Life Problems. Developmental 
Psychology, 37(3), 351-361.  
Piaget, J. (1997). Jean Piaget: Selected Works. New York: Routledge.  
Pritchard, M., & Wilson, G. S. (2003). Using Emotional and Social Factors to Predict 
Student Success. Journal of College Student Development, 44(1), 18-28.  
Puri, M., & Robinson, D. T. (2007). Optimism and economic choice. Journal of 
Financial Economics, 86(1), 71-99.  
Putnam, R. (2007). E Pluribus Unum: Diversity and Community in the Twenty-first 
Century. Scandinavian Political Studies, 30(2), 137-174.  
139 
 
Rhoads, R. A. (1998). In the Service of Citizenship: A Study of Student Involvment in 
Community Service. The Journal of Higher Education, 69(3), 277-297.  
Richardson, M. J., & Pasupathi, M. Young and Growing Wiser: Wisdom during 
Adolescence and Young Adulthood. A Handbook of Wisdom (pp. 139-159)  
Ridgell, S. D., & Lounsbury, J. W. (2004).  
Predicting academic success: general intelligence, "Big Five" personality traits, and 
work drive. College Student Journal, 38(4), 1231-1239.  
Russ-Eft, D., & Preskill, H. (2001). Evaluation in Organizations. New York: Basic 
Books.  
Saris, W. E., & Satorra, A. (1993). Power Evaluations in Structural Equation Modeling. 
In K. A. Bollen, & J. S. Long (Eds.), Testing Structural Equation Models (pp. 181-
204). Newbury Park, CA: Sage.  
Scales, P. C., Roehlkepartain, E. C., Neal, M., Kielsmeier, J. C., & Benson, P. L. (2006). 
Reducing Academic Achievement Gaps: The Role of Community Service and 
Service-Learning. Journal of Experiential Education, 29(1), 38-60.  
Schmidt, J. A., Shumow, L., & Kackar, H. (2007). Adolescent's Participation in Service 
Activities and Its Impact on Academic, Behavioral, and Civic Outcomes. Journal 
of Youth Adolescence, 36, 127-140.  
140 
 
Schneewind, K. (1995). Impact of Family Processes on Control Beliefs. In A. Bandura 
(Ed.), Self-Efficacy in Changing Societies (pp. 114-148). New York: Cambridge 
University Press.  
Schultz, P. W., & Zelezny, L. (1999). Values as predictors of environmental attitudes: 
Evidence for consistency across 14 countries. Journal of Environmental 
Psychology, 19, 255-265.  
Schumacher, E. F. (1977). A Guide for the Perplexed. New York, NY: Harper & Row 
Publishers, Inc.  
Schwartz, B. (2000). Self-Determination: The Tyranney of Freedom. American 
Psychologist, 55(1), 79-88.  
Seligman, M.E.P, & Csikszentmihalyi, M. (2000). Positive Psychology: An introduction. 
American Psychologist, 55, 5-14.  
Sibthorp, J. (2003). An Empirical Look at Walsh and Golins' Adventure Education 
Process Model: Relationships between Antecedent Factors, Perceptions of 
Characteristics of an Adventure Education Experience, and Changes in Self-
Efficacy. Journal of Leisure Research, 35(1), 80-106.  
Spring, Dietz, & Grimm. (2006). Educating for Active Citizenship: Service-Learning, 
School-Based Service and Civic Eingagement. Washington, DC: Corporation for 
National and Community Service.  
141 
 
Staudinger, U. M., & Pasupathi, M. (2003). Correlates of Wisdom-Related Performance 
in Adolescence and Adulthood: Age-Graded Differences in "Paths" Toward 
Desirable Development. Journal of Research on Adolescents, 13(3), 239-268.  
Staudinger, U. M., & Baltes, P. B. (1996). Interactive Minds: A Facilitative Setting for 
Wisdom-Related Performance? Journal of Personality and Social Psychology, 
71(4), 746-762.  
Staudinger, U. M., Maciel, A. G., Smith, J., & Baltes, P. (1998). What Predicts Wisdom-
Related Performance? A First Look at Personality, Intelligence, and Facilitative 
Experiential Contexts. European Journal of Personality, 12, 1-17.  
Sternberg, R. J. (2003). Wisdom, Intelligence, and Creativity Synthesized. New York: 
Cambridge University Press.  
Sternberg, R. J., & and Jordan, J. (Eds.). (2005). A Handbook of Wisdom: Psychological 
Perspectives. New York: Cambridge University Press.  
Sternberg, R. J. (1990). Wisdom and its relations to intelligence and creativity. In R. J. 
Sternberg (Ed.), (pp. 142-159). New York, NY: Cambridge University Press.  
Sternberg, R. J. (1998). A Balance Theory of Wisdom. Review of General Psychology, 
2(4), 347-365.  
Stewart, G. M. (1990). Learning Styles as a Filter for Developing Service-Learning 
Interventions. In C. I. Delve, S. D. Mintz & G. M. Stewart (Eds.), Community 
Service as Values Education (pp. 31-42). San Francisco: Jossey-Bass, Inc.  
142 
 
Takahashi, M., & Overton, W. F. (2002). Cultural Foundations of Wisdom. International 
Journal of Behavioral Development, 26, 267-277.  
Twenge, J. (2006). Generation Me: Why Today's Young Are More Confident, Assertive, 
Entitled-- And More Miserable Than Ever Before. New York: Simon and Schuster.  
Waldstein, F. A., & Reiher, T. C. (2001). Service-Learning and Students' Personal and 
Civic Development. The Journal of Experiential Education, 24(1), 7-13.  
Walsh, V., & Golins, G. (1976). The exploration of the Outward Bound Process. Denver: 
Colorado Outward Bound School.  
Webster, J. D. (2003). An Exploratory Analysis of a Self-Assessed Wisdom Scale. 
Journal of Adult Development, 10(1), 13-22.