Browsing by Subject "Propensity Score Matching"

Now showing 1 - 2 of 2
  • Thumbnail Image
    Item
    Cross-national impacts of public-private partnerships on equitable student achievement: a quasi-experimental assessment using PISA 2009
    (2013-04) Baum, Donald Rey
    Public-private partnerships are being increasingly supported and advocated for, ideologically and financially, as an approach to educational reform in many countries across the world. Proponents suggest that non-state involvement in the education sector has the potential to bolster international Education for All efforts, improve school governance, increase accountability to students and parents, and improve student cognitive outcomes at a lower cost than providing all basic education services through the state. Although the political support for these partnerships from various financial institutions and reform advocates is extensive, empirical evidence investigating achievement and equity impacts on students is greatly lacking. This dissertation adds a much-needed empirical voice to the debate. I perform a quasi-experimental examination of state education finance and provision practices across 17 countries using data from the Program for International Student Assessment (PISA) 2009. Using methods of propensity score matching to reduce the private sector selection bias and hierarchical linear modeling to estimate effects, accounting for a nested data structure, I investigate differences in student cognitive outcomes and equity between public and private education sectors. The work addresses three primary questions: (i) to what extent do students in publicly-funded private schools perform differently than students in public schools?, (ii) to what extent is student socioeconomic status associated with achievement differences?, and (iii) to what extent does student achievement depend upon the school-level characteristics of choice, competition, autonomy, and accountability? This third question deals with a prominent model for engaging the private sector in education, seeking to explain effective education provision through these four key criteria, which are theorized to improve student learning outcomes. The findings of this study somewhat contradict much of what is currently understood about public-private partnerships in education. Much of the current research shows that public-private partnerships have small performance advantages over traditional public schools. The results of this quasi-experimental empirical assessment provide evidence that, holding student and school socioeconomic indicators constant, students in public schools in a number of countries outperform students in publicly-funded private schools. Where these differences occurred, however, they were small, and in the majority of cases there was no difference in performance between public and private schools. In terms of student performance, I find no evidence of systemic inequity in either school sector. That is, low-income students appear to perform at equal levels in public and publicly-funded private schools. However, there is broad cross-country evidence of social discrimination in private sector school access. Enrollment in a private dependent school is associated with higher student socioeconomic status in 13 of 17 countries and publicly-funded private schools are more likely to discriminate in admissions by student academic ability. Keeping these equity findings in mind, I assert that government policy can be used to mitigate inequalities of opportunity through access to public and private school services. This notion entails a more modern view of the state, in which government is not recognized as sole education provider but assumes a larger supervisory capacity focused on regulating learning and guaranteeing equitable educational opportunity.
  • Thumbnail Image
    Item
    Long-term effects of Head Start enrollment on adulthood educational attainment and economic status: a Propensity Score Matching approach.
    (2009-06) Ju, Eunsu
    In 1965, the Head Start program was introduced as part of War on Poverty. There was a hope that intergenerational transmission of poverty could be broken down by a comprehensive preschool program for low income children and their families. Since then, a number of studies have explored the impact of the program. The majority of those studies have focused on cognitive development and educational outcomes, and they generally suggest that the program produces short-term effects, but that most of them disappear in the long-term. However, most previous studies suffered from comparability problems (i.e. different background characteristics) between program participants and non-participants. Further, although the main purpose of the program was to reduce poverty in the long run, confirmation of the idea has been almost ignored from the research for the last 40-plus years. The main purpose of this study is to conduct a more rigorous evaluation in regard to the long-term effects of the Head Start participation on adulthood educational attainment and economic status (as measured by personal earned income, the family income to poverty ratio, and welfare dependency). This study utilized data from the Panel Study of Income Dynamics (PSID) from 1970 to 2005. To examine the effects of Head Start participation on adulthood income, a total of 1,765 young adults (aged 19 to 35 in 2005) were selected from the original PSID sample: 161 with long-term Head Start participation, 171 with short-term Head Start participation, 611 with other preschool participation, and 822 with no preschool participation. Propensity Score Matching (PSM) analysis was employed to control preexisting differences that might have an influence on preschool experience. To ensure whether the findings are consistent regardless of how control groups are matched, four matching models were examined: one-to-one matching without replacement, one-to-one matching with replacement, two-nearest neighborhood matching, and radius matching. Using the matched samples, a series of multivariate analyses were conducted: ordered logistic regression for educational attainment, Tobit analysis for personal earned income, OLS regression for family income to poverty ratio, and logistic regression for welfare dependency. To control the influence of other factors, a variety of variables at the community, family, and child levels were also included in the analyses. Most of the pre-existing differences between groups were sufficiently controlled by the PSM. According to the results of the following multivariate analyses, Head Start provided sizeable gains to its participants in terms of educational attainment if they attended the program at least one year. Long-term Head Starters were more likely to have higher level of educational attainment by approximately 1.5 to 2 times than no preschool children, 2 to 3.2 times than other preschool children, and 1.9 to 2.1 times than short-term Head Starters. However, no significant effects were found among short-term Head Start participants. Regarding the effects on economic status, long-term Head Start participation had indirect effects on personal earnings and family income to poverty ratio through higher educational attainment as compared to short-term Head Start participation or no preschool experience. However, the effects on welfare dependency were not clear. Although some studies have questioned the long-term effects of Head Start participation, this study confirmed that Head Start had provided long lasting effects to its participants on their adulthood educational attainment and economic status. Consequently, more efforts should be made to minimize the barriers that interfere for low income children to attend the program. The programs’ coverage should be expanded and the quality of the program increased to reduce intergenerational poverty.