Date of Graduation


Document Type


Degree Name

Doctor of Philosophy in Public Policy (PhD)

Degree Level



Political Science


William Schreckhise

Committee Member

Todd Shields

Second Committee Member

Garu Ritter


Arkansas, Community development, Evaluation, Human capital, Logistic regression, Persistence, Poverty abatement, Retention, Workforce investment act


This study used a quasi-experimental cross-sectional quantitative model to evaluate the Youth Development Program, a component of the federal Workforce Investment Program. This evaluation study determines whether participation in the Youth Development Program reduced dropout rates among youth in secondary schools in seven school districts in the southeast Arkansas Delta. The impact of the Youth Development Program was examined in the following school districts: Dollarway, Pine Bluff, Watson Chapel, White Hall, Stuttgart, Dewitt, and Star City. The cross-sectional analysis covered the 2006 through 2008 program years. The evaluation model consisted of a comparison group that included 437 youth randomly drawn from the total population of youth and a treatment group of 147 youth which comprised the total population of youth receiving services in the seven districts. In addition to participation in the retention program, other key independent variables such as gender, age, race, grade level and test scores were analyzed through the use of both descriptive and inferential statistics. An ordinary least squares regression model was tested to determine whether the Workforce Investment Act's Youth Development Program achieved its goal of increasing high school retention rates. The analyses reveal initially that the program does seem to have an impact. Students who participate in the program are less likely to drop out of school. However, when student abilities are taken into consideration (and included in the OLS model in the form of test scores), the initial effect that program participation has on the likelihood of dropping out disappears. In other words, no statistically significant relationship exists between program participation and dropout rates. The findings of this evaluation shed light on the methods used for program selection and allude to possible defects within the structure of the agency. This study recommends future research in those areas.