Job Hopping Motives: An Extension of the Unfolding Model of Voluntary Employee Turnover A PLAN B RESEARCH PROJECT SUBMITTED TO THE FACULTY OF UNIVERSITY OF MINNESOTA BY Alison Green Shrift IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS Christopher Lake, Ph.D., Chair Alexandra Luong, Ph.D. Jannifer David, Ph.D. May 2016 © Alison Green Shrift 2016 i Acknowledgements I would like to personally thank my advisor, Dr. Christopher Lake, for all of the help and mentoring that he has provided to me during the past two years. He has made my graduate education at the University of Minnesota-Duluth very rewarding and exciting. I believe he really helped me grow as a student and always highlighted areas for improvement. I would also like to thank my family for being so supportive of my decision to attend graduate school. They have supported me emotionally and financially for the past six years and I will be forever grateful for them. Furthermore, my Grandparents have always been extremely involved in my life and have inspired me to push myself to be the best that I can. ii Abstract Job hopping motives were compared to the components of the unfolding model of voluntary turnover (Lee & Mitchell, 1994). The advancement and escape job hopping motives of 708 people were examined, and incorporated into the current model paths. Image violation was related to advancement motives but unrelated to escape motives (r = -.13, p < .01). Scripts were related to escape motives but not related to advancement motives (r = .18, p < .01). Six scales were created based on questions used by Lee and Mitchell (1994) in the creation of the model. Prediction of future quitting intentions (R2 = .086, R Square Change = .035, F Change = 13.440, Sig. F Change = .000), and past quitting behaviors (R2 = .066, R Square Change = .017, F Change = 6.375, Sig. F Change = .002) was improved using the original model components and job hopping motives. Two additional paths were proposed to describe advancement and escape job hopping profiles. Structural equation modeling was used to compare path fit compared to participants with high and low job hopping motives. Further research is needed to better understand how job hoppers fit into traditional turnover models. iii Table of Contents List of Tables…………………………………………………………………….………...…iv List of Figures…………………………………………………………………………..……..v Recent Employment Trends………….…………..….……………………..………………….1 Importance of Turnover for Organizations ………………………………...…………………2 Turnover Perspective for Employees ………………………………..…..……..……………..3 The Job Hoppers………………….…..……………………………………………………….3 Job Hopping Motives.…………..……………………………………………………….…….4 Early Conceptual Turnover Models….…………………………………..…….…………..….6 The Unfolding Model of Voluntary Turnover………….…..……………………..…………..7 The Unfolding Model of Voluntary Turnover Components……………….……………….…8 The Unfolding Model of Voluntary Turnover Paths...……………………………..………...11 The Proposed Study and Contributions………………………………………..……………..14 Hypotheses and Research Questions………..………………………………………………..15 Method………...………………………………………………………………..…………....22 Results...…...…………………………………..…………………………………………......27 Discussion………………………………………………………………………………..…..32 References..…...…………………………..………………………………………….............38 Appendices..…...…………………………..…………………………………………............42 iv List of Tables Table 1………………………………………………….. Original Interview and Survey Items Page 63 Table 2…………………………………………………………………….... Coefficient Alpha Page 64 Table 3………….…………..…..…………………………………………….……Correlations Page 65 Table 4.…………..…..…………………………………...…Regression of Quitting Intentions Page 66 Table 5 ……………………………………………...…Regression of Past Quitting Behaviors Page 68 Table 6 ……………………………………………...............……...Advancement Path Profile Page 70 Table 7 …………………………………………….................................…Escape Path Profile Page 71 v List of Figures Figure 1………………………………………………………………...…Current Model Paths Page 72 Figure 2………………………………………………………………Job Hopper Profile Paths Page 73 JOB HOPPING AND THE UNFOLDING MODEL 1 Job Hopping Motives: An Extension of the Unfolding Model of Voluntary Employee Turnover Recent Employment Trends A quit is defined as a voluntary separation between the employee and the employer, which is always initiated by the employee. Measuring the number of quits can be indicative of voluntary turnover trends in the labor market. The Job Openings and Labor Turnover Survey (U.S. Bureau of Labor Statistics) revealed a total of 4.9 million separations occurred between January and October, 2015. Voluntary quits accounted for the majority (2.8 million) of the separations, compared to involuntary layoffs and discharges combined (1.7 million). Further, the Bureau of Labor Statistics annually reports that the national voluntary turnover rate in the United States typically approaches 25% (Allen, Bryant & Vardaman, 2010), which can be alarming for organizations. Increased quitting frequency results in employees having a higher number of jobs with shorter tenure. A transition from long tenure to short tenure is becoming more evident and acceptable among younger workers. In recent times, an employee is more likely to have many short-term jobs than a small number of stable long-term jobs (Pfeffer & Baron, 1988). Workers ages 55 to 64 had an average tenure of 10.4 years and workers ages 25 to 34 had an average tenure of 3.0 years; the older workers had an average tenure more than three times longer than the younger workers. Furthermore, a larger proportion of older workers had 10 or more years of tenure with one company compared to younger workers (U.S. Bureau of Labor Statistics, Employee Tenure Summary). This trend has raised awareness of the academic and practical significance of voluntary turnover; turnover has become the subject of intense inquiry, generating a promising body of work (Hom & Zedeck, 2011). Better understanding of this JOB HOPPING AND THE UNFOLDING MODEL 2 phenomena has practical significance for organizations as they aim to select and retain talented employees. Importance of Turnover for Organizations Employees who frequently hop from job to job may appear to be a liability for organizations as turnover can drain money and resources (Lake & Highhouse, 2014). Some amount of turnover is arguably beneficial for organizations, as it allows for new prospective employees to be hired. However, frequently losing talented employees can be detrimental for organizations financially. Turnover incurs financial costs as recruiting and training new employees may cost 90% to 200% of that employee's annual pay (Allen, Bryant, & Vardaman, 2010). Not only is turnover expensive for organizations, there is evidence that links turnover rates to organization-level performance (Allen et al., 2010) such as decreased sales and employee morale. Turnover disrupts business operations (Ton & Huckman, 2008), increases accident rates, and decreases customer service (Shaw, Gupta, & Delery, 2005). Organizations have a competitive advantage if they have managers that are knowledgeable in regard to turnover and the negative effects of frequently losing employees. Research has shed some light on main turnover antecedents, but there is still a large portion of turnover variance that remains undiscovered. Reasons for leaving differ from employee to employee; an employee’s motivation for leaving a job is a combination of external and internal forces (Lee & Mitchell, 1994). The current job market and alternative job openings are external forces to the employee that may impact turnover rates; whereas, employee attitudes and traits are internal forces that may increase or decrease turnover (Lee & Mitchell, 1994). Since employees are becoming highly mobile, JOB HOPPING AND THE UNFOLDING MODEL 3 organizations have spent a great deal of time and money attempting to predict and prevent the loss of talented employees. Turnover Perspective for Employees From the employee perspective, voluntarily changing from job to job may be the new normal. Job hopping behavior may be a necessary means of finding a tolerable work setting or getting ahead in one’s career (Lake & Highhouse, 2014). Increased demands placed on workers and extra job stress could drive an employee to evade a current inhospitable work environment. Meanwhile, a lack of internal promotion systems and increased external hiring could force employees to actively manage their own career by moving from organization to organization (Lake & Highhouse, 2014). It is becoming more acceptable for employees to manage their own employment status and employment opportunities (Strauss, Griffin & Parker, 2012). As a result, moving through jobs quickly is becoming the norm compared to the exception. Organizations find it challenging to predict and prepare for this trend. Job hopping employees may not follow many of the traditional turnover steps which normally include: thoughts of quitting, searching for a job, evaluation and comparison of alternative opportunities, and eventually voluntary turnover (Allen, Bryant & Vardaman, 2010). Careers are becoming more discontinuous (Arthur & Rousseau, 1996), resulting in highly mobile employees who choose to make frequent career changes based on different antecedents. The rise in non-traditional careers has led scholars to pay greater attention to these employees who actively shape their employment status (Tharenou & Terry, 1998), colloquially referred to as job hoppers. The Job Hoppers JOB HOPPING AND THE UNFOLDING MODEL 4 The practice of frequently hopping from job to job has been referred to as job hopping. This behavior has a negative connotation since Ghiselli (1974) coined the term hobo syndrome to describe job hopping behavior. Employees who exhibited this syndrome demonstrated frequent voluntary job hopping behavior and positive attitudes about engaging in such behavior (Woo, 2011). In the current study, job hopping will be defined as the practice of making frequent voluntary job changes; to date, there is not an exact number of jobs an employee must have held to be qualified as a job hopper. A job hopping propensity ratio variable will be calculated to measure a participant’s number of jobs quit to number of jobs held. Employees who have a history of changing jobs in a particular manner are more likely to leave their next employer in a similar manner; employees who do not have a history of changing jobs are more likely to stay with their current or next employer (Judge & Watanabe, 1995). However, this does not mean that all job hopping employees are a bad investment for organizations. When referring to job hopping employees, the motives for frequently hopping jobs is used to describe and better understand this type of voluntary turnover. Job Hopping Motives Job hopping employees can be delineated into two categories based on the motives for changing jobs: advancement driven and escape driven (Lake & Highhouse, 2014). A job hopping motive is defined as the forces that compel an employee to make frequent voluntary job changes (Lake & Highhouse, 2014). Both types of job hopping motives were derived from examining factors such as: personality traits, impulsivity, proactivity, persistence, self-concept, career self- efficacy, growth need strength, previous quitting behaviors, and locus of control. Advancement and escape driven job hoppers may be quite different, but there are some similarities observed in past work history and turnover cognitions (Lake & Highhouse, 2014). Job hopping employees JOB HOPPING AND THE UNFOLDING MODEL 5 are likely to have quit many jobs and have positive thoughts about doing so, regardless of job hopping motive; therefore, advancement and escape motives are correlated due to the aforementioned similarities (r =.34, p < .001). Although these motives are related, they have unique differences that are worth considering. Escape driven job hoppers are referred to as impulsive escape artists (Lake & Highhouse, 2014) because they leave employers to avoid dissatisfaction with the job or coworkers. These employees quit as a means to avoid boredom, annoyance, or frustration with the current work environment (Ghiselli, 1974). Lake and Highhouse (2014) indicate that the escape driven job hopper may lack fortitude and persistence, or they may be very impulsive (Mobley et al., 1978). Traits associated with escape driven job hoppers include negative reactivity, impulsivity, lack of persistence and an external locus of control (Lake & Highhouse, 2014). Therefore, the escape driven job hopper is given this label because it describes the desire to immediately escape a work environment (Lake & Highhouse, 2014). Conversely, advancement driven job hoppers are employees who aim to enhance their career by hopping through jobs vertically (Lake & Highhouse, 2014). This type of job hopper is referred to as advancement driven because the employee desires to advance their own career by seeking jobs that will propel them towards their ultimate career goal. Traits associated with advancement driven job changes include positive proactivity, career confidence, and a desire to grow (Lake & Highhouse, 2014). Advancement driven job hoppers have been referred to as ambitious ladder climbers because some may use frequent job changes as a means to leverage a job offer from another company (Lake & Highhouse, 2014). Unlike escape driven job hoppers, the advancement driven job hoppers are more likely to make elaborate evaluations before making a final quitting decision. JOB HOPPING AND THE UNFOLDING MODEL 6 Simply measuring the number of jobs an employee held may not be as informative when evaluating turnover decisions. However, measuring job hopping motives may shed light on specific turnover antecedents that may be influential for certain types of employees. Understanding why and how an employee quit a job is more beneficial for organizations as they try to hire the most qualified and dependable employees. As suggested by Lee and Mitchell (1994), several dispositional and situational variables may influence quitting decisions for employees. Self-concept, which includes the variables used to delineate the two types of job hoppers, may make certain antecedents more influential on the quitting decision. Viewed as individual choice behavior, voluntary turnover has long captivated academicians attempting to validate popular turnover models, as well as employers seeking to manage a costly behavior (Campion, 1991). Early Conceptual Turnover Models Early research on the topic of turnover supposed that job dissatisfaction was the main antecedent for turnover. It was believed that if the employees felt the organization did not meet their expectations, turnover would result. Research is showing that job dissatisfaction might be the driving force in fewer than half of individual turnover decisions (Lee, Mitchell, Holtom, McDaniel & Hill, 1999). Employees who are dissatisfied with their current job may remain with their current employer for an extended period of time; many satisfied employees leave their current job for other opportunities. Many organizations have focused on improving employee satisfaction in hopes of retaining talented employees, yet turnover is still occurring due to other antecedents. The next school of thought in regard to voluntary turnover combined job satisfaction with perceived ease of movement in March and Simon’s (1958) model of organizational equilibrium. JOB HOPPING AND THE UNFOLDING MODEL 7 A combination of these two factors was believed to be the most predictive antecedent of voluntary turnover. March and Simon (1958) influenced contemporary turnover models by recognizing alternative employment opportunities can play a role in quitting decisions (Hom & Zedeck, 2011). Mobley (1977) expanded on this theory by focusing on how comparisons are made and allowed researchers to delineate multiple steps in the turnover process (Hom & Zedeck, 2011). Despite these early frameworks for predicting turnover, most models were only able to explain approximately 25% of voluntary turnover variance (Maertz & Campion, 1998). The process of quitting tends to be more complex than early turnover theorists hypothesized; not all employees follow the traditional turnover models aforementioned. As a result, the unfolding model was created (Lee & Mitchell, 1994) to better understand not only why but how employees quit. The Unfolding Model of Voluntary Turnover Early frameworks for turnover had many shortcomings, which motivated Lee and Mitchell (1994) to create the unfolding model of voluntary turnover (Hom & Zedeck, 2011). Their model has ignited considerable research about how and why employees leave an organization. To gain information as to why employees leave, organizations often conduct exit interviews in hopes of better understanding the turnover process (Branham, 2005). The unfolding model was built from semi-structured interview data obtained from exit interviews with employees. Next, data were obtained from the same employees a few weeks later using a follow- up survey. The questions in the exit interview and follow-up survey addressed many antecedents that were hypothesized to impact turnover decisions. Results revealed that employees took a wide variety of quitting patterns using six common turnover antecedents. These six antecedents JOB HOPPING AND THE UNFOLDING MODEL 8 were chosen as model components to create the model paths. The model assumes that employees experience each model component in a step-by-step fashion during the quitting process. Lee and Mitchell (1994) developed four unique paths (see Figure 1) to classify organizational leavers and to help organizations better understand common quitting processes. Each path includes a combination of these turnover antecedents and summarizes how employees interpret their work environments as well as how employees decide to quit (Lee & Mitchell, 1994). Although this model advanced turnover knowledge, there is still the possibility that an employee could leave an organization and not fit into one of the four established paths (Lee & Mitchell, 1999). Further research is needed to classify these unique organizational leavers which account for a proportion of unexplained turnover. The Unfolding Model of Voluntary Turnover Components As Figure 1 shows, the first component an employee might encounter is a shock. A shock is defined as a jarring event that often initiates the psychological analyses involved in quitting a job (Lee et al., 1999). The shock will raise awareness of the current employment situation; it generates information that has a meaning to the employee, and it must be interpreted and integrated into the employee’s system of beliefs and images (Lee & Mitchell, 1994). If the shock cannot be accepted or integrated, the employee may have thoughts of quitting or even quit immediately. In any case, a shock may be positive (e.g. job offer or promotion), negative (e.g. a demotion or getting fired), expected (such as a company merger), unexpected (e.g. death of boss), job related (e.g. required weekly work hours) or non-job related (e.g. a pregnancy). Regardless of the type of shock, it is a noticeable and jarring event that an employee experiences. JOB HOPPING AND THE UNFOLDING MODEL 9 The second component that an employee may encounter is a script, which is perhaps the most relevant model component to the job hopping employee. A script has previously been defined as a schematic knowledge structure held in memory that specifies behavior or event sequences that are appropriate for specific situations (Gioia & Poole, 1984). A script focuses on an existing plan of action or habit, which is based on past experience and social expectations (Lee et al., 1999). For example, an employee who has previously quit when they were not cohesive with their coworkers will likely look to that previous behavior for guidance if they encounter a similar situation in the future. Consistent with the script component and job hopping behavior (Judge & Wantabe, 1995), past quitting behavior predicts future instances of quitting behavior. This quitting behavior may become a habitual script that is enacted by employees (Gioia & Poole, 1984). Prior to the development of the unfolding model, scripts and habits were not typically considered by turnover theorists and researchers (Lee & Mitchell, 1994). Most quitting decisions were believed to be made consciously rather than in a routinized or scripted manner (Lee & Mitchell. 1994). Although the current model posits additional steps for some quitting paths, the model’s authors assumed that fast-paced scripted quitting decisions did not involve the satisfaction or image violation components (see Path 1, Figure 1). Currently, only Path 1 includes the script component. The third component that an employee may encounter is image violation, which evaluates how well the employee can find harmony between their personal values and goals and that of their current situation in the organization (Lee et al., 1999). An image is unique to each employee, and represents the employee’s ideal situation or goal. Individuals can hold images for every aspect of their life (e.g. social life, family life, academic life, or even career life). This component differs from job satisfaction as many employees endure jobs that they dislike because JOB HOPPING AND THE UNFOLDING MODEL 10 it fits their professional image. Image theory (Beach, 1990) assumes the employee will evaluate their current situation use this information and make a comparison between their current and ideal situations (Lee & Mitchell, 1994). If an employee enjoys working with others in a team- based setting, and their employer encourages such group interaction, image violation has not occurred. The employees’ and the employers’ images are cohesive. However, if the employer does not allow team-based work, image violation may occur because that employee enjoys working with others. These images may be more or less clear, easy or hard to articulate, and strong or weakly held (Lee & Mitchell, 1999). Decisions will be made easier and quicker to the extent the former conditions hold (Lee & Mitchell, 1999). The fourth component that an employee may encounter is job satisfaction. Low levels of job satisfaction may occur when employees come to feel that their job no longer provides the intellectual, emotional, or financial benefits they desire (Lee et al., 1999). In the current model, this is expected to occur over an extended period of time due to experiences that occur on-the- job. Although previously thought to be very important, job satisfaction seems to be a weak antecedent to predict turnover (Hom et al., 2011). Researchers have found that decreased job satisfaction may be more predictive of counterproductive work behaviors than turnover. Many unsatisfied employees continue to work at the same company for a very long period of time, while many satisfied employees leave satisfying jobs to explore alternative job options (Holtom, Mitchell, Lee, 2005). The fifth component that an employee may engage in is searching for alternatives. This component includes the activities involved with looking for alternative employment opportunities and evaluating the current job openings. Searching for alternatives may take the form of internet job searches, improving resumes and cover letters, or even applying to other JOB HOPPING AND THE UNFOLDING MODEL 11 organizations. An employee may have engaged in searching for alternatives prior to quitting their job, but not always (Lee et al., 1999). As Mobley suggested (1977), impulsive quitting may not include searching for alternative before making the final quit-decision because the withdrawal period is much quicker. More deliberate quitting decisions may take more time, and therefore, may include more time to search for alternative employment. The sixth component than an employee may encounter is a likely offer, meaning that the employee is likely to receive an offer, or has an offer of employment with another organization. Although this component is relatively simple compared to the others, it is an important antecedent for turnover. Having another job offer may persuade an employee to quit their current job. Some employees may wait for a job offer before deciding to quit their current job, while others may quit without having another job offer at hand. These six components have been combined to create the four pathways in the unfolding model; each path utilizes a different combination of model components and is briefly described below (Lee & Mitchell, 1994). The Unfolding Model of Voluntary Turnover Paths Path 1- Shock to the System and a Memory Probe Resulting in a Match: A Script-Driven Decision (Lee & Mitchell, 1999); Following a Plan (Hom, Zedeck & Sheldon, 2011). A shock occurs which initiates thoughts of quitting. This path is very different from traditional turnover models (Lee & Mitchell, 1994), as it incorporates the shock and script components. Next, the employee searches their memory to find a script, which often results in the employee acting out the scripted behavior as they have in the past. If the employee finds that they had past work experience that was very similar, they will often react accordingly based on that previous situation. Path 1 is the only path that incorporates scripted decision making. In this path, a quitting response is ready, available, and used with minimal deliberations, evaluation of JOB HOPPING AND THE UNFOLDING MODEL 12 the job, or other job alternatives (Lee & Mitchell, 1994). Therefore, Path 1 is the fastest path in the model for employees to move through as it assumes other model components, such as image violation and satisfaction, are unimportant (Lee & Mitchell, 1999). An example of Path 1 would be an employee who did not receive a promotion. In previous situations that a promotion was not received, the employee has quit. The script is enacted and the employee decides to quit rather quickly. Path 2- Shock to the System, No Match, and No Specific Job Alternatives: A Push Decision (Lee & Mitchell, 1999); I’m out of Here (Hom, Zedeck & Sheldon, 2011). A shock occurs, which initiates thoughts of quitting for an employee. However, the employee does not have a pre-existing plan of action to leave, a similar past work experience, or social expectations regarding the shock. Next, the employee evaluates how well they can adapt to the shock. Employees use image theory for integration of the shock, which states that the employee will evaluate how compatible they feel with the job after the shock has occurred. They may also consider if they feel a different organization would be a better fit. It is important to note that this path is similar to Path 1, as it includes a shock. The major difference is that Path 2 lacks a scripted decision and requires more mental deliberations. Typically, a negative shock occurs on this path and the employee does not think they can adapt, which results in image violation and turnover. This pathway does not include the search for alternative employment, implying that the quitting decision could be made impulsively (Donnelly & Quirin, 2006). An example of Path 2 would be an employee who left immediately when the hospital shifted from individualized patient care, the employees preferred nursing philosophy, to team-based nursing (Hom & Zedeck, 2011). JOB HOPPING AND THE UNFOLDING MODEL 13 Path 3- Shock to the System, No Match and Presence of Specific Job Alternatives: A Pull Decision (Lee & Mitchell, 1999); Leaving for Something Better (Hom, Zedeck & Sheldon, 2011). This path is very similar to the previous paths as it begins with a shock. There is no pre- existing plan of action or scripted decision. Next, the employee evaluates how well they could integrate to the shock by comparing the jarring event with personal and career images. However, this path includes the search for alternative employment which is different from Path 2. If the employee can find an alternative employment option that may be a better fit, they quit. However, if the employee believes the current employer is the better fit, they remain employed regardless of the shock (Donnelly & Quirin, 2006). Instead of focusing an employee to reassess the commitment to the current company (as in Path 2), the employee assesses alternative options for employment and the likelihood they could obtain alternative employment (Donnelly & Quirin, 2006). If so, a quit is likely to occur (Lee & Mitchell, 1994). An example of Path 3 would be an employee who searches for other jobs when a promotion was not received by their current employer. This shock initiated the employee to compare alternatives with the current job which may motivate them to pursue and evaluate other jobs (Hom & Zedeck, 2011). Path 4a and 4b- No Shock to the System: Affect Initiated (Lee & Mitchell, 1999); Leaving Dissatisfying Job (Hom, Zedeck & Sheldon, 2011). Unlike the other three paths, Path 4a and 4b do not include a shock. Rather, the main antecedent of these paths is decreased job satisfaction that unfolds over time. According to Lee and Mitchell (1994), no singular event jars mental deliberations toward recognition of prior shocks (Path 1), reassessment of an individual's basic commitment to the current organization (Path 2), or assessment of the likelihood of commitment to another organization (Path 3). These paths are similar to the traditional models of turnover, which presumed that most decisions were JOB HOPPING AND THE UNFOLDING MODEL 14 dissatisfaction-driven. Path 4a classifies employees who become dissatisfied and quit their job without searching for alternative employment. Path 4b includes a search for alternatives as a result of boredom or dissatisfaction with the current employer (Donnelly & Quirin, 2006). Note that both Path 4a and 4b have some similar components to the other three paths, but they lack the shock component. At this point, it is unclear where job hopping employees may fit into the existing paths. Based on traits associated with each job hopping motive, certain model components may be influential on their turnover decisions; some model components may not be influential of job hopping turnover. Further research is needed to better understand how and why job hoppers voluntarily withdraw from an organization. The Proposed Study and Contributions Extending this model to incorporate job hopping motives has many benefits. First, the current study will contribute to the turnover literature by evaluating the relationship between job hopping motives (Lake & Highhouse, 2014) and unfolding model of voluntary turnover components (Lee & Mitchell, 1994). The relationship between job hopping motives and model components has never been evaluated and will allow us to better understand how the personality profiles are related to certain turnover antecedents. The relationships between job hopping motives, which describe the propensity of voluntary turnover, and the model components, which are specific to a single quitting experience, have never been evaluated. Second, the current study will create six scales to measure each component of the unfolding model (Lee & Mitchell, 1994). The scale items will be derived and modified from the original exit interview and the follow-up survey questions (Table 1) used by Lee and Mitchell (1994). Each of the six scales will be designed to replicate the original questions; as needed, JOB HOPPING AND THE UNFOLDING MODEL 15 open ended questions will be transformed into statements to allow participants to respond on a 5- point Likert scale. Third, the prediction of future quitting intentions and past quitting behaviors will be assessed using demographic variables, the original model components, and job hopping motives. It is expected that job hopping motives are accounting for a different portion of turnover variance compared to traditional turnover antecedents. It is unknown if job hopping motives measure the same portion of variance as the original unfolding model of voluntary turnover components. Regression analyses will be used to see if job hopping motives, when added to the original model components, can better predict future quitting intentions and past quitting behaviors. Fourth, structural equation modeling will be used to assess path fit for advancement and escape job hopping motives. Two additional paths were proposed to describe the advancement and escape job hopping profiles. Participants will be divided into high and low advancement to assess if the path better fit employees who strongly demonstrated the advancement profile. Next, participants will be divided into high and low escape motives to assess if the proposed path better fit employees who strongly demonstrated the escape profile. Hypotheses and Research Questions Shocks The relationship between job hopping motives and the shock component is unknown. Job hopping employees may not wait for a jarring event, such as a shock, to initiate thoughts of quitting. Although external work shocks may influence quitting behavior for some employees, it may not be the most influential factor for employees with strong advancement or escape motives as these values are held internal to the employee. However, research suggests that the escape motive may attempt to exit a negative work environment or react impulsively to a negative work JOB HOPPING AND THE UNFOLDING MODEL 16 situation, such as a shock. The shock component may initiate quitting for some employees, but employees with high escape or advancement motives may or may not wait for the shock component to occur before initiating a quit. The relationship between job hopping motives and the shock component of the unfolding model (Lee & Mitchell, 1994) warrants further investigation. Research Question 1a: Does a significant relationship exist between advancement motives and the shock component? Research Question 1b: Does a significant relationship exist between escape motives and the shock component? Scripts The script component is a cognitive plan which may automate turnover behavior in well- known or commonly experienced situations (Hom et al., 2011). This component is likely related to employees with high advancement and escape job hopping motives as it is assumed to develop out of routine behavior. Employees higher on the advancement or escape spectrum are expected to have quit many jobs which may result in a script. Lee and Mitchell (1999) suggest that individuals with a history of voluntarily leaving many organizations might be more likely to hold pre-existing scripts about when to leave an organization in the future. Based on the fact that advancement and escape driven job hoppers are likely to have quit many jobs and have positive opinions on doing so, both job hopping motives are hypothesized to correlate positively with preexisting scripts (Lee & Mitchell, 1994). Employees with this preexisting script will have an idea of when and how to quit, making the withdrawal period faster than employees without a preexisting script. It is hypothesized that the script component will correlate positively with shorter withdrawal time. JOB HOPPING AND THE UNFOLDING MODEL 17 Hypothesis 2a: Advancement motives will correlate positively with the script component. Hypothesis 2b: Escape motives will correlate positively with the script component. Hypothesis 2c: The script component will correlate positively with a shorter withdrawal time. Image Violation Job hopping motives are strongly related to personality traits. Since both job hopper profiles are derived from personality traits (e.g. self-esteem, self-concept, self-efficacy, locus of control, growth need strength), they are expected to hold a strong personal and professional image. Personal images may hold different value and represent the decision-maker’s principles or morals. Since employees with high advancement or escape motives hold a strong self-image, it is likely that image violation will influence their quitting decision. When the environment is hindering progress toward the employee’s ideal image, violation has occurred. This strong self- image may allow job hopping employees to make faster quitting-decisions as suggested by Lee and Mitchell (1999). Specifically, personal and professional images may serve as a screening tool for employees to enhance the speed of decision-making. The scale is designed to measure person-organization fit or harmony. A high score on this scale indicates that an employee’s image is cohesive with the company; a low score on this scale indicates poor fit or image violation. A significant negative relationship between job hopping motives and the image component of the unfolding model (Lee & Mitchell, 1994) is expected. Further, the image component is expected to negatively correlate with a shorter withdrawal time because the image scale is measuring fit between the employee and the image. Therefore, a negative relationship indicates image violation or a lack of fit. JOB HOPPING AND THE UNFOLDING MODEL 18 Hypothesis 3a: Advancement motives will correlate negatively with the image violation component. Hypothesis 3b: Escape motives will correlate negatively with the image violation component. Hypothesis 3c: Image violation will correlate negatively with withdrawal time. Job Satisfaction Job satisfaction is currently included in the model, but the relationship with job hopping turnover is unknown. Mobley (1977) recognized that a more complete understanding of the psychology of the withdrawal decision process requires investigation beyond the satisfaction- turnover relationship. Further, the relationships between job satisfaction is consistent but not particularly strong (Mobley, 1977). Alternative forms of withdrawal (e.g. absenteeism, passive job behavior, counterproductive work behaviors) are more prevalent outcomes of low job satisfaction compared to turnover. Although satisfaction may be important for retention, it may not be a driving force influencing turnover decisions for advancement driven job hopping employees who may endure a disliked situation to propel their career. Conversely, the escape job hopping employee may be predisposed to disliking most work environments, making job satisfaction less impactful. Supporting previous research (Veiga, 1981) that job satisfaction may not be very influential on job turnover, the relationship between job hopping motives and job satisfaction remains unknown. Research Question 4a: Does a significant relationship exist between advancement motives and the satisfaction component? Research Question 4b: Does a significant relationship exist between escape motives and the satisfaction component? JOB HOPPING AND THE UNFOLDING MODEL 19 Searching for Alternatives and Likely Offer The search for alternatives component may be impacted by job hopping motives. Similar to the constructs described by Mobley (1977), impulsive quitting decisions describe employees who did not use great mental deliberations when deciding to leave their previous job. Research indicates that escape driven job hoppers often act impulsively due to high reactivity, low persistence, and an external locus of control (Lake & Highhouse, 2014). Conversely, traits associated with the advancement driven job hopper indicate they may have higher persistence, an internal locus of control, and a higher growth need strength (Lake & Highhouse, 2014). This type of employee may be more likely to take more time making decisions compared to the escape driven job hopper. It is hypothesized that the advancement motives will be related to searching for alternatives and likely job offer components, but the escape driven job hoppers will likely exit rather impulsively resulting in no search for alternatives or likely offer (Lee & Mitchell, 1994). Hypothesis 5a: Advancement motives will correlate positively with the search for alternatives component. Hypothesis 5b: Advancement motives will correlate positively with the likely offer component. Hypothesis 5c: Escape motives will correlate negatively with the search for alternatives component. Hypothesis 5d: Escape motives will correlate negatively with the likely offer component. Withdrawal Periods Although it is not directly a component of the model, the withdrawal time may be influenced by habitual job changing. Employees higher on advancement and escape motives JOB HOPPING AND THE UNFOLDING MODEL 20 (Lake & Highhouse, 2014) are expected to have a higher frequency of job changes. The duration of the withdrawal period may be indicative of how many components of the model impacted the quitting decision. Quick organizational exits may represent a strong desire to leave; conversely, an employee who takes a long time to quit may be a result of evaluating many alternative employment options. Longer withdrawal periods may imply that more deliberative decision making is taking place; thus suggesting that more of the model components are being used. Advancement-driven withdrawals are likely to take into consideration many different factors before making a final decision; whereas, escape-driven quitting decisions may be made rather impulsively without many considerations. It is hypothesized that the advancement-driven job hopping motives will not demonstrate a relationship with withdrawal time, but the escape-driven job hopping motives will demonstrate a relationship with shorter withdrawal time. Hypothesis 6a: Advancement motives will correlate negatively with withdrawal time. Hypothesis 6b: Escape motives will correlate positively with withdrawal duration. Predicting Future Quitting Intentions and Past Quitting Behavior Habitual hopping from job to job has been predictive of future quitting behavior (Judge & Watanabe, 1995). Reasons for this cycle are becoming a major interest for researchers even though most organizations simply accept the phenomenon and hope to retain talented employees. Researchers are interested in what turnover antecedents are influencing and predicting future quitting intentions. It is expected that employees demonstrating high advancement or escape motives will be more likely to hold intentions for quitting in the future. Traditional turnover antecedents are likely to predict some proportion of variance in quitting intentions, however, it is hypothesized that the addition of job hopping motives to the traditional turnover antecedents will increase the prediction of future turnover intentions. JOB HOPPING AND THE UNFOLDING MODEL 21 Job hopping motives may be accounting for a different proportion of voluntary turnover variance compared to traditional turnover antecedents. Therefore, the traditional turnover antecedents may account for a certain proportion of past turnover behavior. It is hypothesized that the addition of job hopping motives to traditional turnover antecedents will account for a greater proportion of past quitting behavior. To summarize, the addition of job hopping motives to traditional turnover antecedents is hypothesized to better predict future quitting intentions and account for more of the variance in past turnover behavior. Hypothesis 7a: Advancement and escape motives will significantly improve the prediction of future quitting intentions when added to the original Unfolding Model of Voluntary Turnover components. Hypothesis 7b: Advancement and escape motives will significantly improve the prediction of number of jobs quit when added to the original Unfolding Model of Voluntary Turnover components. Proposed Paths Based on the job hopping profiles described by Lake and Highhouse (2014) for advancement and escape driven job hoppers, two additional paths are proposed a priori (see Figure 2). Traits associated with escape driven job hoppers include negative reactivity, impulsivity, lack of persistence and an external locus of control (Lake & Highhouse, 2014). Traits associated with advancement driven job changes include positive proactivity, career confidence, and a desire to grow (Lake & Highhouse, 2014). Path 5a describes the path that advancement driven job hoppers would take to exit an organization; whereas, Path 5b describes the path that escape driven job hoppers would take to exit an organization. Advancement driven job hoppers are expected to use a script for making decisions, experience image violation, search JOB HOPPING AND THE UNFOLDING MODEL 22 for alternative employment, and have a likely offer. Escape driven job hoppers are expected to use a script for making decisions, experience image violation, and exit without searching for alternatives or having a likely offer. Neither the advancement or escape job hopper are expected to wait for a specific jarring event to occur as the job hopping motives are internal factors to the employee. Hypothesis 8a: Participants who score high on advancement motives will fit the Path 5a significantly better than participants who score low on advancement motives. Hypothesis 8b: Participants who score high on escape motives will fit the Path 5b significantly better than participants who score low on escape motives. Method Procedures All participants were recruited through Amazon’s MechanicalTurk database and paid $2.00 for participation in the study. The study consisted of one typed response under 400 words and nine survey scales. It was expected to take 15 minutes to complete the study. Eight of the scales used a 5-point Likert response scale ranging from “Strongly Disagree” to “Strongly Agree.” The demographic scale evaluated variables such as sex, age, and past work history. To begin the study, participants were provided with the definition of a quit and how it differs from being fired or laid-off. A quit is defined as a voluntary separation between an employee and an employer, which is always initiated by the employee. Being laid-off or fired is initiated by the employer. All participants were asked to recall a time they had voluntarily quit a job and explain the situation in detail by typing into a blank textbox. Participants were instructed to include events that occurred before, during, and after they quit a job. After that task was complete, participants were told to answer the remaining survey scales in regard to the specific quitting JOB HOPPING AND THE UNFOLDING MODEL 23 experience they described. Each scale had a brief description of the specific model component to familiarize the participants with the construct. After the qualitative summary and scales were completed, participants were given an end of survey message with a code for compensation. Participants The original sample included 1,000 participants. Due to an error in the survey setup and administration, specifically the end of survey message, some participants completed multiple surveys. These duplicate responses were an area for concern moving forward with the statistical analyses. Therefore, the data needed to be evaluated for duplicate as well as careless responses before proceeding (Meade & Craig, 2012). A total of 292 participant responses were removed from the original sample for (a) not taking enough time completing the survey (under 5 minutes), (b) not responding to the majority of the survey questions (over 80% of the questions), or (c) submitting duplicate surveys. Duplicates were identified by repeated IP Addresses. The error described above resulted in an abnormally large portion of participants to be removed from the study. A total of 708 participants remained after cleaning the data. The remaining sample contained 302 male participants (42%) and 405 female participants (56%). Of the 708 participants, 548 were Caucasian, 61 selected the Other or No Response option, 58 were African American, 38 were Hispanic, and 3 were Native American. Ages ranged from 18 years old to 69 years old with an average age of 34.48 years old. The majority of the sample was currently employed (n = 521; 72%) and unmarried (n = 431; 60%) without children (n = 428; 60%). The private job sector refers to all-for-profit companies that are not state or federally regulated. Conversely, the public job sector describes a job that is state or federally regulated, such as a non-profit organization. The sample was equally employed between the public (n = 310; 43%) JOB HOPPING AND THE UNFOLDING MODEL 24 and private (n = 382; 53%) sectors, with a very small portion choosing the No Response option (n = 16; 2.2%). Instruments When creating and measuring the current model paths, Lee and Mitchell (1994) conducted a semi-structured exit interview with nurses and then sent the same employees a follow-up survey which allowed them to evaluate the quit from two different perspectives. Original questions from the exit interview can be found on Appendix A. The authors cross- validated the survey items with the original interview items by conducting correlational analyses. The comparison between the original interview questions and follow-up survey questions can be seen in Table 1. As recommended by Lee and Mitchell (1994), appropriate methods with which to collect data on decision paths may be retrospective in nature, such as a cross-sectional survey design. Therefore, the current study created scales to measure the current turnover paths based on the original study questions. All of the newly developed scales contain questions adapted from Lee and Mitchell’s (1994) exit interview and follow-up survey. When possible, the exit interview questions were transformed from open-ended questions into statements, which allowed participant responses to be evaluated quantitatively. Extensive effort was placed on ensuring the newly created scales were similar in content and meaning to the original materials used in the creation of the unfolding model (Lee & Mitchell, 1994). The Job Shock Scale (Appendix B) was created to measure the shock component of the unfolding model (Lee & Mitchell, 1994). All of the participants read, “for some, but not all employees, an event occurs which makes them want to quit their job. Please refer to the quitting experience you previously described when answering the following questions,” before JOB HOPPING AND THE UNFOLDING MODEL 25 completing the scale. Example scale questions include, “a particular event made me think about quitting” and “there was a clearly distinguishable jarring event that provoked my leaving.” The Script Scale was created to measure the script component (Appendix C). All of the participants read, “many factors may impact an employee’s decision to quit. Some employees base decisions off of another experience that was very similar. For others, quitting is very spontaneous. Please refer to the quitting experience you described when answering the following questions,” before completing the scale. Questions on this scale were designed to evaluate routinized quitting behavior. Example scale questions include, “I had a pre-existing plan of action when I quit my last job,” and “choosing to leave my job is something I do frequently.” The Image Violation Scale was created to measure personal images and feelings of image violation (Appendix D). All of the participants read, “everybody holds personal images for different domains in their life. These images represent personal goals and desires that each person holds for themselves. For example, we hold images for work, family, friends, and recreation. Please refer to the quitting experience you described when answering the following questions,” before completing the scale. Example scale questions include, “I felt like I fit in with my previous employer,” and “my personal values were very similar to my previous employer’s values.” Job satisfaction (Appendix E) was evaluated using the Generic Job Satisfaction Scale (Macdonald & MacIntyre, 1997). Lee and Mitchell (1994) did not ask any questions regarding satisfaction in their exit interview or follow-up survey; therefore, this scale was chosen to evaluate the job satisfaction component. Initially this scale included 44-items, but the authors used factor analysis to reduce the scale down to 10-items. Cronbach’s alpha of .77 was calculated for the shortened version of the scale (Macdonald & MacIntyre, 1997). JOB HOPPING AND THE UNFOLDING MODEL 26 The Search for Alternatives Scale (Appendix F) was created to evaluate the search for alternatives component. All of the participants read, “prior to quitting an employee has the option of searching for alternative employment opportunities. Some quitting experience are more sudden than others and leave less time to search for alternatives. However, some quitting experiences are prolonged and well thought-out. Please refer to the quitting experience you described when answering the following questions,” before completing the scale. Example scale questions include, “I considered other jobs when I decided to quit,” and “I searched for other jobs before I quit my job.” The Likely Job Offer Scale was created to measure the likely offer component (Appendix G). All of the participants read, “please refer to the quitting experience you previously described when answering the following questions,” before completing the scale. Example scale questions include, “I had a likely offer at the time of my last quitting experience,” and “I felt confident that I would get a job offer elsewhere before I quit my last job.” The Withdrawal Period Scale was created to measure the duration of the quitting process (Appendix H). All of the participants read, “please refer to the quitting experience you previously described when answering the following questions,” before completing the scale. Example scale questions include, “My decision to quit was very quick,” and “I didn’t take much time considering the quit.” Job hopping motives were evaluated using the Job Hopping Scale (Lake & Highhouse, 2014), which delineates advancement and escape motives (Appendix I). Example escape motive questions include, “when a person discovers they dislike their coworkers, they should move to another job, and keep switching jobs until they finally find a good place to work,” and “because working for one company tends to create boredom, people should move from company to JOB HOPPING AND THE UNFOLDING MODEL 27 company often.” Example advancement motive questions include, “people should be willing to change jobs as many times as necessary to get the best job possible,” and “frequently moving between jobs is perfectly justified when each job change leads to a more prestigious job.” Demographic Information (Appendix J) will be collected for each participant. This self- report data will examine many factors such as age, sex, race, and public or private job sector. Results All of the six newly created scales were internally consistent (Cronbach’s alpha > .70). The Script Scale had a low alpha level initially (α = .64), which was a cause for concern. A factor analysis revealed that there were two different factors: past scripted quitting behavior and future quitting intentions. After editing the scale to evaluate only past scripted behavior, the alpha level of the scale increased to .73, which is stronger and resolved the cause for concern. All other scales met traditional guidelines for internal consistency reliability (see Table 2). Correlations were calculated to evaluate the relationship between job hopping motives and the model components (see Table 3). Neither the advance (r = -.03, p = .45) or escape (r = - .04, p = .30) motives were significantly correlated with the shock component, answering Research Question 1a and 1b. A significant positive relationship was not observed between the script component and advancement motives (r = -.03, p = .49) which fails to support Hypothesis 2a. Escape motives were positively correlated with the script component (r =.18, p < .001) which supports Hypothesis 2b. Further, a significant positive relationship was not observed between the script component and a shorter withdrawal time (r = .03, p = .45) which fails to support Hypothesis 2c. Since the use of a script did not relate to shorter withdrawal time, this is inconsistent with Lee & Mitchell’s Path 1 which ignored some model components to result in a quicker withdrawal. Note that an increase in image violation indicates better fit with an JOB HOPPING AND THE UNFOLDING MODEL 28 organization; the lowest end of the image violation scale indicates poor fit or image violation. Advancement motives were negatively correlated with image violation (r = -.13, p = .001), supporting Hypothesis 3a. Escape motives did not correlate negatively with image component (r = .04, p = .35), which fails to support Hypothesis 3b. However, image violation component was negatively correlated (r = -.16, p < .001) with withdrawal time, supporting Hypothesis 3c; advancement driven job hopping motives tended to be related with image violation, which supports the idea that this type of job hopper may be dissatisfied with their current job which results in hopping to a better organization. Further, image violation was related to a shorter withdrawal time indicating that stronger self-images were related to faster quitting decisions. Satisfaction was not correlated with the advancement (r = -.02, p = .67) or escape (r = .02, p = .62) job hopping motives, answering Research Questions 4a and 4b. Therefore, the internal forces that compel job hoppers to quit are not significantly related to job satisfaction. It was expected that advancement motives would positively correlate with searching for alternative employment and having a likely offer. No significant positive relationship was observed between searching for alternatives (r = .04, p = .30) or likely offer (r = .06, p = .11) which fails to support Hypothesis 5a and 5b. This result was surprising based on the profiles of advancement job hopper profile traits. Although negatively related, the correlation between escape motives and searching for alterative employment (r = -.02, p = .68) or likely offer (r = .06, p = .11) was nonsignificant, which fails to support Hypothesis 5c and 5d. Withdrawal time was positively correlated to escape motives (r = .08, p < .05) which indicates a shorter quit duration which supports Hypothesis 6b; since escape motives are related to impulsivity and high reactivity, it was expected that withdrawal time would correlate with escape motives, but not advancement JOB HOPPING AND THE UNFOLDING MODEL 29 motives. A significant negative relationship was not observed between the advancement motives and shorter withdrawal time (r = .02, p = .66), which fails to support Hypothesis 6a. Other correlations were calculated to evaluate relationships between model components (see Table 3). The shock component was negatively correlated with searching for alternatives (r = -.17, p < .01) and likely offers (r = -.17, p < .01), and positively correlated with a shorter withdrawal time (r = .32, p < .01). The script component was positively correlated with searching for alternatives (r = .16, p < .01) and likely offer (r = .16, p < .01), but negatively correlated with age (r = -.17, p < .01) and sex (r = -.12, p < .01). Image violation was positively correlated with satisfaction (r = .71, p < .01), but negatively correlated with searching for alternatives (r = -.08, p < .05) and withdrawal time (r = -.16, p < .01). Note that an increase in image violation indicates better fit with an organization; the lowest end of the image violation scale indicates poor fit or image violation. A very strong correlation was found between job satisfaction and image violation which indicates they are measuring some portion of the same construct. However, escape motives were not correlated with image violation or satisfaction, indicating they are not completely related. Some distinct differences exist among the escape motives, image violation, and satisfaction that warrants further investigation. The satisfaction component is negatively correlated with withdrawal time (r = -.17, p < .01) and positively correlated with age (r = .12, p < .01). Searching for alternatives was positively correlated with likely offer (r = .73, p < .01), but negatively correlated with withdrawal time (r = -.24, p < .01) and sex (r = -.09, p < .05). Likely offer was negatively correlated with withdrawal time (r = -.16, p < .01) and sex (r = -.09 p < .05). JOB HOPPING AND THE UNFOLDING MODEL 30 A regression was calculated to evaluate the intent to quit one’s current job (see Table 4). The regression included three steps, entered sequentially. The first step (Model A) included demographic variables such as race, sex, and age (R2 = .016, R Square Change = .016, F Change = 3.856, Sig. F Change = .009). The second step (Model B) expanded from the first model by combining demographic variables and the unfolding model components (R2 = .051, R Square Change = .035, F Change = 4.229, Sig. F Change = .000). The final step (Model C) included demographic variables, unfolding model components, and job hopping motives (R2 = .086, R Square Change = .035, F Change = 13.440, Sig. F Change = .000). Each step significantly improved the predictability of quit intentions, answering Hypothesis 7a. A second regression was conducted to evaluate the prediction of quitting propensity (see Table 5). For each participant, a new variable called quitting propensity was calculated by computing a ratio of number of jobs quit to number of jobs held. Examining this ratio controlled for the number of jobs a person has held in their life. The first step (Model A) included demographic variables such as race, sex, and age (R2 = .015, R Square Change = .015, F Change = 3.424, Sig. F Change = .017). The second step (Model B) expanded from the first model by combining demographic variables and the unfolding model components (R2 = .049, R Square Change = .034, F Change = 4.103, Sig. F Change = .000). The final step (Model C) included demographic variables, unfolding model components, and job hopping motives (R2 = .066, R Square Change = .017, F Change = 6.375, Sig. F Change = .002). Each model significantly improved the predictability of past quitting behaviors, answering Hypothesis 7b. Prediction of past quitting behavior and quitting intention was significantly enhanced when job hopping motives were included. The motives predict quitting behaviors and intentions above and beyond the traditional components of the unfolding model. JOB HOPPING AND THE UNFOLDING MODEL 31 Job hopper paths were derived a priori based on the specific and unique job hopper profile (Lake & Highhouse, 2014). Structural equation modeling was used to analyze the fit of the job hopper profile paths. The advancement job hopping path was evaluated by comparing employees with high and low advancement motives. Employees were divided into high advancement and low advancement using a median split technique (median = 3.75). Based on this median split, a dichotomous grouping variable (0 = low advancement motives; 1 = high advancement motives) was created. A set of SEM models was then run using this grouping variable as a moderator. Surprisingly, employees high on advancement motives fit Path 5a worse (χ2= 327.021, p < .001, TLI = 0.390, CFI = 0.512, RMSEA = 0.244), compared to employees’ low on advancement motives (χ2= 256.202, p < .001, TLI = 0.328, CFI = 0.462, RMSEA = 0.276) (see Table 6). The escape job hopping path was evaluated by comparing employees with high and low escape motives. Employees were divided into high escape and low escape using a median split technique (median = 2.75). Based on this median split, a dichotomous grouping variable (0 = low escape motives; 1 = high escape motives) was created. A set of SEM models was then run using this grouping variable as a moderator. Employees high on escape motives did fit Path 5b better (χ2= 560.698, p < .001, TLI = --0.068, CFI = 0.003, RMSEA = 0.321), compared to employees who are low on escape motives (χ2= 575.063, p < .001, TLI = -0.061, CFI = 0.010, RMSEA = 0.348) (see Table 7). Overall, the proposed models did not fit high or low job hopping motives. As a part of the SEM-based moderator analysis, the computer program Mplus (Muthen & Muthen, 2007), was used to calculate the chi-square difference between high and low motive job hoppers in constrained (parameters jointly estimated for high and low motives) and unconstrained groups (parameters separately estimated for high and low motives). A significant JOB HOPPING AND THE UNFOLDING MODEL 32 chi-square value from this comparison indicates a significant moderating effect; parameters were different for high and low motive groups. Conversely, a non-significant chi-square value points to no meaningful difference between high and low motive groups. Results revealed that the model fit significantly better for employees with low advancement motives (df = 12, χ2 Difference= 68.818, p < .001), which fails to support Hypothesis 8a. A second chi-square difference test was calculated to see if there was a significant model improvement for employees with high and low escape motives. Results revealed that the model did not fit significantly better for employees with high escape motives (df = 14, χ2 Difference= -14.365, p = 0.423), which fails to support Hypothesis 8b. Although the model fit was poor for both paths, there was a significant improvement in model fit as employees demonstrated stronger advancement motives. Path 5a was a better fit for employees high on advancement compared to low on advancement. Path 5b was not a better for employees high on escape compared to low on escape. All things considered, neither path was a good fit to the job hopping profile, which indicates more research is needed to understand the step-by-step quitting process for job hopping employees. This type of employee may better fit in the model if the components were rearranged. Discussion The unfolding model (Lee & Mitchell, 1994) has been extensively studied and used for classifying and understanding how employees decide to quit. Examining the relationships (see Table 3) between job hopping motives and unfolding model components revealed similarities between constructs. Results suggest that the model is not capturing all of the aspects of the quitting process for employees with high advancement and escape job hopping motives; the model components are possibly missing some aspects of the quitting process that job hopping JOB HOPPING AND THE UNFOLDING MODEL 33 motives are measuring. As suggested, relationships between model components and job hopping motives may indicate ways the withdrawal process may differ. Further, prediction of past quitting behaviors and future quitting intentions is better explained when demographic variables are held constant and job hopping motives are added to the unfolding model components. Strengths and Limitations The primary strength of this research is that it was the first to incorporate job hopping motives into the very popular unfolding model (Lee & Mitchell, 1994). To date, the relationship between job hopping motives and model components has never been evaluated. Prediction of future quitting intention and past quitting behavior was better explained when job hopping motives were incorporated into the original model components. This is beneficial as it indicates job hopping motives are capturing a portion of the turnover variance that traditional turnover models are missing. A second strength of this research is that it used a very diverse pool of participants through Amazon’s MechanicalTurk. The sample had a wide variety of ages, races, and job sectors. The distribution of males and females was very even which will improve the generalizability of the findings to both sexes. This technique made it feasible to obtain a very diverse and large sample size to ensure most demographic groups were included in the study. The representation of diversity is a benefit, given that many studies attempt to draw conclusions from small and unrepresentative samples. For example, the unfolding model was created and used on a very simplistic sample of female nurses from a single hospital, which make the generalizability of the findings to job hopping employees very limited. However, common method bias is always a concern that may increase or inflate the observed relationships in a study. A limitation to the study is the data collection procedure was a JOB HOPPING AND THE UNFOLDING MODEL 34 cross-sectional survey design that collected data from participants at one point in time. There is little control over the setting in which participants complete the survey (e.g. location, distractions, participant mood). It would be ideal to evaluate job hopping behavior using a longitudinal research design to collect data at many different time points. Another possible way to overcome common method bias is to collect data from other sources (e.g. work history from resumes, manager description of the quitting process, an employment agency). Evaluating the quit from multiple sources, instead of self-report, could improve the accuracy of the data. Unfortunately, a cross-sectional design was more feasible. Another possible limitation to the current study is social desirability, which would prompt a participant to respond in a favorable manner to the survey questions. Voluntary quitting has a negative connotation and can be indicative of job hopping behavior. Social desirability describes the need to appear in a positive light to others in need of approval. The need to appear favorable may reduce or enhance the relationships between constructs. Their responses may not accurately describe the quitting experience or common quitting trends. Another possible limitation in the current study and turnover knowledge is the lack of clarity regarding the job hopping construct. To date, there is not a specific number of jobs an employee must hold in order to be determined a job hopper. The current study evaluated the propensity of job changes which is a ratio of jobs quit to jobs held. As voluntary turnover continues to be a pressing issue for academic professionals and organizations, a specific definition or qualification should be established in regard to the job hopper construct. Practical Implications The relationship between job hopping motives and the specific model components helped to better understand how different motivational forces may be related to turnover antecedents. JOB HOPPING AND THE UNFOLDING MODEL 35 There seems to be a portion of turnover variance that is explained by job hopping motives that traditional turnover antecedents do not capture. Organizations evaluate candidates by comparing applicable materials such as a resume, cover letter, and references. If an applicant has a resume with many jobs and short tenure, which is indicative of job hopping behavior, it may be alarming for an organization. However, better understanding of the motives that compelled the employee to quit many jobs by asking the candidate to explain their quitting process may be beneficial for organizations. Based on the results, image violation seems to be related to advancement motives. With this knowledge, organizations may be able to retain advancement driven employees by ensuring person-organization fit. Understanding this type of employees personal and career desires can help an organization prevent turnover. Internal promotion systems may be a way to retain advancement driven job hoppers. Also, succession planning could be beneficial to ensure talent is available. Information obtained from candidates may reveal they would not be a liability if hired. Based on the results, the script component seems to be related to escape motives. With this knowledge, organizations may be able to prevent or disrupt the script by understanding past quitting instances. Ensuring that the organization has managers that are willing to be supportive and help make the work environment more tolerable could help retain this type of employee. Understanding what the employee enjoys could help create a more flourishing and nurturing environment, which could reduce turnover. From a financial standpoint, organizations are concerned with selecting and retaining talented employees. Although given a negative connotation, job hopping employees are not necessarily poor investments. Understanding which motives the candidate demonstrated in JOB HOPPING AND THE UNFOLDING MODEL 36 previous quitting experiences may help retain talented candidates if selected. Measuring job hopping motives may allow organizations to retain talented employees and disrupt the script component for escape driven job hoppers. Highlighting an internal promotion system or skill development may retain advancement driven job hoppers. Future Research More research is needed to fully understand the impact of job hopping motives on traditional withdrawal processes. Factors such as age, family or marital status, and income may limit such job hopping behaviors. Following the same participants over an extended period of time would give a better understanding of how this job hopping behavior develops. It is possible that an employee may exhibit escape motives at one point in their life, but advancement motives at another point in time. It could be possible that certain model components from the unfolding model (Lee & Mitchell, 1994) become more or less influential on the quitting process as an employee ages. For example, Allstate has conducted logistic regression analyses which revealed that predictive strength of turnover antecedent’s changes over time (Hom & Zedeck, 2011). For instance, newcomers’ satisfaction assessed on the 90th day of employment-but not work satisfaction a year after joining Allstate-predicted turnover (Hom & Zedeck, 2011). Therefore, future research should consider how voluntary turnover decisions unfold over time. Another avenue for future research could evaluate disengagement or withdrawal from an organization when voluntary turnover does not occur. Mobley (1977) highlighted that many other behaviors may be demonstrated if an employee has withdrawn from an organization. There may be additional negative consequences for an organization if employees’ withdrawal from the organization without quitting. Specifically, behaviors such as counterproductive work behaviors, absenteeism, decreased engagement, and lower performance could occur and serve as warning JOB HOPPING AND THE UNFOLDING MODEL 37 signs of voluntary turnover. Future research should investigate the relationship between job hopping motives and counterproductive work behaviors as a predictor of turnover. JOB HOPPING AND THE UNFOLDING MODEL 38 References Allen, D. G., Bryant, P. C., & Vardaman, J. M. (2010). Retaining talent: Replacing misconceptions with evidence- based strategies. The Academy of Management Perspectives, 24(2), 48-64. Anderson, J. R. (1983). A spreading activation theory of memory. Journal of verbal learning and verbal behavior, 22(3), 261-295. Beach, L. R. (1990). Image theory: Decision making in personal and organizational contexts (p. 254). Chichester: Wiley. Beach, L. R. (Ed.). (1998). Image theory: Theoretical and empirical foundations. Routledge. Bernardin, H.J. (1977). The relationship of personality variables to organizational withdrawal. Personnel Psychology, 30, 17-27. Branham, L. (2005). The 7 hidden reasons employees leave. American Management, New York, 172. Breaugh, J. A. (1981). Relationships between recruiting sources and employee performance, absenteeism, and work attitudes. Academy of Management Journal, 24(1), 142-147. Boswell, W. R., Boudreau, J. W., & Tichy, J. (2005). The relationship between employee job change and job satisfaction: The honeymoon-hangover effect. Journal of Applied Psychology, 90, 882-892. Davis, P. R., Trevor, C. O., & Feng, J. (2015, February 23). Creating a More Quit-Friendly National Workforce? Individual Layoff History and Voluntary Turnover. Journal of Applied Psychology. Advance online publication. http://dx.doi.org/10.1037/apl0000012 Dittrich, J. E., & Carrell, M. R. (1979). Organizational equity perceptions, employee job satisfaction, and departmental absence and turnover rates. Organizational behavior and human performance, 24(1), 29-40. Donnelly, D. P., & Quirin, J. J. (2006). An extension of Lee and Mitchell's unfolding model of voluntary turnover. Journal of Organizational Behavior, 27(1), 59-77. Retrieved from http://search.proquest.com.libpdb.d.umn.edu:2048/docview/621091528?accountid=8111 Ghiselli, E. E. (1974). Some perspectives for industrial psychology. American Psychologist, 29(2), 80. Gioia, D. A., & Poole, P. P. (1984). Scripts in organizational behavior.Academy of management review, 9(3), 449- 459. Harrison, D. A., & Hulin, C. L. (1989). Investigations of absenteeism: Using event history models to study the absence-taking process. Journal of Applied Psychology, 74(2), 300. JOB HOPPING AND THE UNFOLDING MODEL 39 Heckman, J. J., & Borjas, G. J. (1980). Does unemployment cause future unemployment? Definitions, questions and answers from a continuous time model of heterogeneity and state dependence. Economica, 247-283. Holtom, B. C., Mitchell, T. R., Lee, T. W., & Inderrieden, E. J. (2005). Shocks as causes of turnover: What they are and how organizations can manage them. Human Resource Management, 44(3), 337-352. doi:http://dx.doi.org/10.1002/hrm.20074 Hom, P. W., Mitchell, T. R., Lee, T. W., & Griffeth, R. W. (2012). Reviewing employee turnover: Focusing on proximal withdrawal states and an expanded criterion. Psychological Bulletin, 138, 831-858. Hom, P. W., & Griffeth, R. W. (1991). Structural equations modeling test of a turnover theory: Cross-sectional and longitudinal analyses. Journal of applied psychology, 76(3), 350. Hom, Peter W. Zedeck, Sheldon (Ed), (2011). Organizational Exit. APA handbook of industrial and organizational psychology, Vol 2: Selecting and developing members for the organization. APA Handbooks in Psychology., (pp. 325-375). Washington, DC, US: American Psychological Association, viii, 598 pp. http://dx.doi.org/10.1037/12170-011 Judge, T. A., & Watanabe, S. (1995). Is the past prologue? A test of ghiselli's hobo syndrome. Journal of Management, 21(2), 211-229. Retrieved from http://search.proquest.com.libpdb.d.umn.edu:2048/docview/618680193?accountid=8111 Lake, C. J. & Highhouse, S. (2014, May). Identifying and assessing job hopping motives. Poster presented at the annual meeting of the Society for Industrial and Organizational Psychology, Honolulu, HI. Lee, T. W. (1996). Why employees quit. Decision making in the workplace: A unified perspective, 73-90. Lee, T. W., & Mitchell, T. R. (1994). An alternative approach: The unfolding model of voluntary employee turnover. Academy of Management Review,19(1), 51-89. Lee, T. W., Mitchell, T. R., Holtom, B. C., McDaniel, L. S., & Hill, J. W. (1999). The unfolding model of voluntary turnover: A replication and extension. Academy of Management Journal, 42(4), 450-462. Retrieved from http://search.proquest.com.libpdb.d.umn.edu:2048/docview/199821267?accountid=8111 Macdonald, S., & Maclntyre, P. (1997). The generic job satisfaction scale: Scale development and its correlates. Employee Assistance Quarterly, 13(2), 1-16. Maertz, C. P., & Campion, M. A. (1998). 25 years of voluntary turnover research: a review and critique. JOB HOPPING AND THE UNFOLDING MODEL 40 Maertz, C. P., Jr., & Griffeth, R. W. (2004). Eight motivational forces and voluntary turnover: A theoretical synthesis with implications for research. Journal of Management, 30(5), 667-683. doi:http://dx.doi.org/10.1016/j.jm.2004.04.001 March, J. G., & Simon, H. A. (1958). Organizations. Mobley, W. H. (1977). Intermediate linkages in the relationship between job satisfaction and employee turnover. Journal of applied psychology, 62(2), 237. Mobley, W. H., Horner, S. O., & Hollingsworth, A. T. (1978). An evaluation of precursors of hospital employee turnover. Journal of Applied psychology, 63(4), 408. Morgan, L. G., & Herman, J. B. (1976). Perceived consequences of absenteeism. Journal of Applied Psychology, 61(6), 738. Parker, D. F., & Dyer, L. (1976). Expectancy theory as a within-person behavioral choice model: An empirical test of some conceptual and methodological refinements. Organizational Behavior and Human Performance,17(1), 97-117. Paterson, D.G., Darley, J.D., & Elliot, R.M. (1936). Men, women, and jobs: A study in human engineering. Minneapolis, MN: University of Minnesota Press. Shaw, J. D., Gupta, N., & Delery, J. E. (2005). Alternative conceptualizations of the relationship between voluntary turnover and organizational performance.Academy of Management Journal, 48(1), 50-68. Ton, Z., & Huckman, R. S. (2008). Managing the impact of employee turnover on performance: The role of process conformance. Organization Science, 19 (1), 56-68. U.S. Bureau of Labor Statistics (U.S. Bureau of Labor Statistics) http://www.bls.gov/ Employee Tenure Summary (U.S. Bureau of Labor Statistics) http://www.bls.gov/news.release/tenure.nr0.htm Veiga, J. F. (1981). Plateaued versus nonplateaued managers: Career patterns, attitudes, and path potential. Academy of Management Journal, 24(3), 566-578. Retrieved from http://search.proquest.com.libpdb.d.umn.edu:2048/docview/616553595?accountid=8111 Verplanken, B., & Orbell, S. (2003). Reflections on Past Behavior: A Self‐Report Index of Habit Strength1. Journal of Applied Social Psychology, 33(6), 1313-1330. JOB HOPPING AND THE UNFOLDING MODEL 41 Wanous, J. P. (1976). Organizational entry: From naive expectations to realistic beliefs. Journal of Applied Psychology, 61(1), 22. Woo, S. E. (2011). A study of Ghiselli's hobo syndrome. Journal of Vocational Behavior, 79(2), 461-469. Youngblood, S. A., Mobley, W. H., & Meglino, B. M. (1983). A longitudinal analysis of the turnover process. Journal of Applied Psychology, 68(3), 507. Zikic, J., & Hall, D. T. (2009). Toward a more complex view of career exploration. The Career Development Quarterly, 58(2), 181-191. JOB HOPPING AND THE UNFOLDING MODEL 42 Appendices Appendix A: Original Interview Questions Appendix B: Job Shock Scale Appendix C: Scripted Quitting Behavior Scale Appendix D: Image Violation Scale Appendix E: Generic Job Satisfaction Scale Appendix F: Search for Alternatives Scale Appendix G: Likely Offer Scale Appendix H: Withdrawal Period Scale Appendix I: Job Hopping Tendency Scale Appendix J: Demographics JOB HOPPING AND THE UNFOLDING MODEL 43 Appendix A Original Interview Questions Lee & Mitchell (1994) JOB HOPPING AND THE UNFOLDING MODEL 44 1. Can you describe the circumstances surrounding the time you first began to feel or think that you should leave your job at the hospital? 2. Was there a particular event that caused you to think about leaving? Please describe that event and the circumstances surrounding that event. 3. Were you asked to leave? 4. How soon after you began thinking of quitting did you make up your mind to leave? 5. Did you consider other job alternatives or options in making your decision? 6. Did you already have some job offers when you decided to quit? 7. Did you search for other jobs before or after you left the hospital? 8. Was the job search comprehensive? That is, how thoroughly did you gather information on other jobs? 9. Did you decide that you would better fit in one of these options? If yes, could you describe why you would fit better? 10. How would you rate the compatibility between your personal goals (which can include professional) values and those of your hospital? 11. Was your career progressing the way you expected it to? 12. Were your personal goals progressing the way you expected it to? 13. If you had stayed, would you have been able to achieve all of your career goals? Would you have been able to achieve all of your personal goals? 14. Have you been in a similar set of circumstances before (in terms of leaving a job)? If so, please describe what happened? 15. What’s your current job? If it’s not a nurse, is it related to nursing and/or the health care area? 16. How old are you? 17. How many years did you work for the hospital full-time or part-time? What service did you mostly work? What shift? 18. What year did you graduate from nursing school? 19. How many dependents do you have? 20. What proportion of family income does (did) your nursing job provide? JOB HOPPING AND THE UNFOLDING MODEL 45 Appendix B Job Shock Scale Adopted from Lee & Mitchell’s (1994) original interview & survey JOB HOPPING AND THE UNFOLDING MODEL 46 For some, but not all employees, an event occurs which makes them want to quit their job. Please refer to the quitting experience you previously described when answering the following questions. A particular event made me think about quitting A particular event made me quit my last job There was a clearly distinguishable jarring event that provoked my leaving Response Scale: 1 = Strongly Disagree, 2 = Somewhat Disagree, 3 = Neither Agree nor Disagree, 4 = Somewhat Agree, 5 = Strongly Agree JOB HOPPING AND THE UNFOLDING MODEL 47 Appendix C Scripted Quitting Behavior Scale Adopted from Lee & Mitchell’s (1994) original interview and survey JOB HOPPING AND THE UNFOLDING MODEL 48 Many factors may impact an employee’s decision to quit. Some employees base decisions off of another experience that was very similar. For others, quitting was very spontaneous. Please refer to the quitting experience you described when answering the following questions. *I based my decision to quit on my past work experiences I would likely quit another job again I had a pre-existing plan of action when I quit my last job *I made the decision to quit this job based on my experiences in similar situations *I have quit other jobs under similar circumstances as I have quit my most recent job Deciding to quit was an easy choice to make *I have a history of quitting jobs in this manner If the situation arises in the future, I would quit in the exact same way that I quit this job It is normal for people I know to quit in this way * indicates item was used after factor analysis * Response Scale: 1 = Strongly Disagree, 2 = Somewhat Disagree, 3 = Neither Agree nor Disagree, 4 = Somewhat Agree, 5 = Strongly Agree JOB HOPPING AND THE UNFOLDING MODEL 49 Appendix D Image Violation Scale Adopted from Lee & Mitchell’s (1994) original interview and survey JOB HOPPING AND THE UNFOLDING MODEL 50 Everybody holds personal images for different domains in their life. These images represent personal goals and desires that each person holds for themselves. For example, we hold images for work, family, friends, and recreation. Please refer to the quitting experience you described when answering the following questions. I felt like it “fit-in” with my previous employer I felt like I would experience a better fit with another organization My personal goals were very compatible with my previous employer’s goals My personal values were very similar to my previous employer’s values My career was progressing the way I had hoped at my previous job Quitting was my best option to progress my career If I had stayed, I would have been able to achieve my career and personal goals Response Scale: 1 = Strongly Disagree, 2 = Somewhat Disagree, 3 = Neither Agree nor Disagree, 4 = Somewhat Agree, 5 = Strongly Agree JOB HOPPING AND THE UNFOLDING MODEL 51 Appendix E Generic Job Satisfaction Scale JOB HOPPING AND THE UNFOLDING MODEL 52 Job satisfaction is important for employees and employers. Please refer to the quitting experience you described above when answering the following questions. I receive recognition for a job well-done I feel close to the people at work I feel good about working at this company I feel secure about my job I believe management is concerned about me On the whole, I believe work is good for my physical health My wages are good All my talents and skills are used at work I get along with my supervisors I feel good about my job Response Scale: 1 = Strongly Disagree, 2 = Somewhat Disagree, 3 = Neither Agree nor Disagree, 4 = Somewhat Agree, 5 = Strongly Agree JOB HOPPING AND THE UNFOLDING MODEL 53 Appendix F Search for Job Alternatives Scale Adopted from Lee & Mitchell’s (1994) original interview and survey JOB HOPPING AND THE UNFOLDING MODEL 54 Prior to quitting an employee has the option of searching for other employment opportunities. Some quitting experiences are more sudden than others and leave less time to search for alternatives. However, some quitting experiences were prolonged and well thought out. Please refer to the quitting experience you described when answering the following questions. I considered other jobs when I decided to quit I searched for other jobs before I quit my last job My job search was very intense I was going to quit my previous job regardless of possible employment alternatives Response Scale: 1 = Strongly Disagree, 2 = Somewhat Disagree, 3 = Neither Agree nor Disagree, 4 = Somewhat Agree, 5 = Strongly Agree JOB HOPPING AND THE UNFOLDING MODEL 55 Appendix G Likely Job Offer Scale Adopted from Lee & Mitchell’s (1994) original interview and survey JOB HOPPING AND THE UNFOLDING MODEL 56 Please refer to the quitting experience you described when answering the following questions. I had a likely job offer when I quit my last job I felt confident that I would get a job offer elsewhere before I quit my last job I had a firm job offer when I quit my last job Response Scale: 1 = Strongly Disagree, 2 = Somewhat Disagree, 3 = Neither Agree nor Disagree, 4 = Somewhat Agree, 5 = Strongly Agree JOB HOPPING AND THE UNFOLDING MODEL 57 Appendix H Withdrawal Period Scale Adopted from Lee & Mitchell’s (1994) original interview and survey JOB HOPPING AND THE UNFOLDING MODEL 58 Please refer to the quitting experience you described when answering the following questions. My decision to quit was very quick My mental deliberations to quit were fast I didn’t take much time considering the quit It took me a great deal of time to decide to quit Response Scale: 1 = Strongly Disagree, 2 = Somewhat Disagree, 3 = Neither Agree nor Disagree, 4 = Somewhat Agree, 5 = Strongly Agree JOB HOPPING AND THE UNFOLDING MODEL 59 Appendix I Job Hopping Tendency Scale JOB HOPPING AND THE UNFOLDING MODEL 60 Please answer the following questions to the best of your ability. How many jobs have you had? __________ How many of these jobs did you voluntarily decide to quit? _______ How many times have you quit your previous job? ________ How long do you usually stay with one job? ________ How long do you usually have between jobs? _______ I plan on staying with my next employer for a very long time. Escape Motive Items 1. Becoming disinterested in a job is a good reason to move from job to job as often as desired. 2. When a person discovers they dislike their coworkers, they should move to another job, and keep switching jobs until they finally find a good place to work. 3. Because working for one company tends to create boredom, people should move from company to company often. 4. Repeatedly changing jobs is an ideal way to get a variety of job experiences. Advancement Motive Items 5. Even if someone has changed jobs several times, they should take a new job if it involves moving to a better position. 6. People should be willing to change jobs as many times as necessary to get the best job possible. 7. Frequently moving between jobs is perfectly justified when each job change leads to a more prestigious job. 8. It is desirable to periodically move from job to job, looking for the job that best improves one's lifestyle. Response Scale: 1 = Strongly Disagree, 2 = Somewhat Disagree, 3 = Neither Agree nor Disagree, 4 = Somewhat Agree, 5 = Strongly Agree JOB HOPPING AND THE UNFOLDING MODEL 61 Appendix J Demographic Information JOB HOPPING AND THE UNFOLDING MODEL 62 Age (in years) _________ Sex (M/F/No Response) Race (White, African American, Native American, Other, No Response) Type of Organization (Public, Private, No Response) How long did you work for your last employer prior to quitting? _______ Do you plan on quitting in the future? (Yes, No) How many jobs have you had? How many jobs have you quit? JOB HOPPING AND THE UNFOLDING MODEL 63 TABLE 1 Lee & Mitchell (1994) Interview and Survey Items Item Number Interview Survey 1 Was there a clearly distinguishable and jarring event? There was a particularly identifiable event that started me thinking about leaving. 2 Did you consider other job alternatives or options in making your decision? I considered other job options when deciding to leave. 3 Was the job search comprehensive? That is, how thoroughly did you gather information on other job options? I gathered lots of information about other job options. 4 Did you already have some job offers when you decided to quit? I had actual job offers in hand before I decided to leave. 5 How would you rate the compatibility between your personal goals and values (which can include professional) and those of your hospital? My values match up well with the organizational values of the hospital I left. 6 Was your career progressing the way you expected it to? My professional goals were being met in the hospital I left. 7 Have you ever been in a similar set of circumstances before (in terms of leaving a job)? I’ve left hospitals before for similar reasons. 8 Were these deliberations quick? I deliberated a long time before I decided to leave. 9 Is there anything that your peers, supervisor, or the hospital could have done before you actually quit which might have caused you to stay? If the hospital had done a couple of things for me, I would have stayed. 10 What’s your current job? I am currently working as a nurse. JOB HOPPING AND THE UNFOLDING MODEL 64 TABLE 2 Coefficient Alpha Scale Name Coefficient Alpha Shock Scale .91 Script Scale .73 Image Violation Scale .85 Satisfaction Scale .88 Search for Alternatives Scale .78 Likely Offer Scale .86 Withdrawal Time Scale .94 Advance Motive Scale .78 Escape Motive Scale .77 Cronbach’s alpha >.70 represents internal consistency 65 TABLE 3 Correlations Variable 1 2 3 4 5 6 7 8 9 10 11 1. Advance 2. Escape .39** 3. Shock -.03 -.04 4. Script -.03 .18** .02 5. Image Violation -.13** .04 .02 -.01 6. Satisfaction -.02 .02 .01 -.06 .71** 7. Search for Alt .04 -.02 -.17** .16** -.08* -.01 8. Likely Offer .06 .06 -.17** .16** -.05 .05 .73** 9. Withdrawal Time .02 .08* .32** .03 -.16** -.17** -.24** -.16** 10. Age .00 -.04 .04 -.17** .06 .12** -.06 -.03 -.01 11. Sex┼ -.08* -.15** .04 -.12** .02 .04 -.09* -.09* -.05 .05 ┼ P<.01 ** P<.05 * Male = 1 Female = 0 66 TABLE 4 Regression for Quitting Intentions Model & Components R2 R Square Change F Change Sig F Change Standardized Beta t p A .016 .016 3.856 .009 Age .004 2.85 .005 Sex -.041 -1.35 .178 Race -.011 -0.98 .325 A & B .051 .035 4.229 .001 Age .004 2.80 .005 Sex -.051 -1.66 .097 Race -.010 -0.85 .394 Shock .051 3.02 .003 Script -.022 -1.13 .259 Image .030 1.24 .215 Satisfaction -.056 -2.47 .014 Alternatives .002 0.09 .922 Likely Offer -.024 -1.45 .146 A, B & C .086 .035 13.440 .001 Age .004 2.85 .005 Sex -.070 -2.32 .020 Race -.005 -.482 .630 Shock .047 2.87 .004 Script -.013 -.661 .509 Image .022 0.90 .371 JOB HOPPING AND THE UNFOLDING MODEL 67 Satisfaction -.050 -2.22 .027 Alternatives -.007 -.336 .737 Likely Offer -.018 -1.07 .286 Advancement -.061 -2.51 .012 Escape -.064 -3.16 .002 Dependent Variable: future quitting intentions A. Demographic Variables: race, sex, age B. Unfolding Model Components: shock, script, satisfaction, image violation, search for alternatives, likely offer components C. Job Hopping Motives: advancement and escape motives JOB HOPPING AND THE UNFOLDING MODEL 68 TABLE 5 Regression for Past Quitting Behaviors Model & Components R2 R Square Change F Change Sig F Change Unstandardized Beta t p A .015 .015 3.424 .017 Age -.001 -1.085 .278 Sex .064 2.992 .003 Race -.004 -0.539 .590 A & B .049 .034 4.103 .001 Age -.001 -0.955 .340 Sex .070 3.266 .001 Race -.006 -0.659 .487 Shock -.035 -2.960 .003 Script .037 2.693 .007 Image -.003 -0.162 .871 Satisfaction .038 2.365 .018 Alternatives .001 0.053 .958 Likely Offer -.006 -0.491 .624 A, B & C .066 .017 6.375 .002 Age -.001 -0.988 .323 Sex .079 3.722 .000 Race -.008 -0.995 .320 Shock -.033 -2.832 .005 Script .033 2.391 .017 Image .001 0.076 .939 JOB HOPPING AND THE UNFOLDING MODEL 69 Satisfaction .034 2.166 .031 Alternatives .005 0.369 .713 Likely Offer -.009 -0.783 .434 Advancement .030 1.740 .082 Escape .031 2.173 .030 Dependent Variable: number of quit / number of jobs held A. Demographic Variables: race, sex, age B. Unfolding Model Components: shock, script, satisfaction, image violation, search for alternatives, likely offer components C. Job Hopping Motives: advancement and escape motives JOB HOPPING AND THE UNFOLDING MODEL 70 TABLE 6 Path 5a Advancement Profile Motive X2 RMSEA TLI CFI High 327.021 0.244 0.390 0.512 Low 256.202 0.276 0.328 0.462 Note: Median split was used to create high and low groups on the job hopping motives. JOB HOPPING AND THE UNFOLDING MODEL 71 TABLE 7 Path 5b Escape Profile Motive X2 RMSEA TLI CFI High 560.698 0.321 -0.068 0.003 Low 575.063 0.348 -0.061 0.010 Note: Median split was used to create high and low groups on the job hopping motives. JOB HOPPING AND THE UNFOLDING MODEL 72 FIGURE 1 Current Model Paths JOB HOPPING AND THE UNFOLDING MODEL 73 FIGURE 2 Job Hopper Profile Paths