The Impact of the HOPE Scholarship on High School Graduation in Georgia Muhammad A. Yadudu Abstract Since Georgia enacted its widely acclaimed merit aid HOPE policies, many states have followed suit, ostensibly due to its large, positive, and significant impact on college enrollment. There has been muted interest in the impact of the policy on high school graduation of both students who will go on to college and those who will not. Using a differences-in-differences methodolog y, I contrast high school graduation rates in Georgia against similar states and found the enactment of the HOPE Scholarship has increased the probability of graduating high school by 12.5%. When viewed in light of a 7% point jump in college enrollment due to HOPE Scholarships (Cornwell et al., 2006), I conclude that the policy has increased the graduation rates—by about 5%—of students who will not be immediately enrolling in college after graduation from high school. 80 Spring 2017 Middle Tennessee State University 81 America may be facing a silent high school graduation problem. Although figures provided by government agencies seem to show a tremendous leap in graduation rates (see Figure 1), the raw numbers mask a hidden but fundamental problem. By 2011, around 12% of all high school graduates obtained this credential through the General Education Degree (GED) (Murnane, 2013).1 While statistically indistinguishable, labor market outcomes and long-term educational achievement is very different among GED holders than among true high school graduates (Cameron & Heckman, 1993). Figure 1. A plot of graduation rates over time. The dotted line shows the graduation rate, including the GED, while the dashed line shows the graduation rate without GED creden- tials, assuming the GED rate is 12% of the total every year. Source: https://eric.ed.gov /?id=ED524955. To develop and maintain a highly skilled workforce, many states have resorted to providing merit-based subsidies that incentivize college enrollment. However, Fenske et al. (1997) argue that policies that intervene earlier in students’ education may be just as important. Research on the efficacy of these merit-based policies tend to focus mainly on the impact of said policy on tertiary enrollment (Dynarski, 2002). In this study, I examine the impact of Georgia’s Helping Outstanding Pupils Educationally (HOPE) scholarship on high school graduation. Using a pooled 1. The General Educational Development (GED) certificate was supposed to be an equivalent cre- dential to the high School diploma for veterans of the Second World War. It has later morphed into the medium through which many Black, Hispanic and lower income individuals receive their diploma. As a result of this transformation, the value of the credential has weakened considerably (Cameron & Heckman, 1993). The Impact of the HOPE Scholarship on High School Graduation in Georgia cross-section of IPUMS CPS Annual Social and Economic supplement (ASEC) data, I analyze the impact of Georgia’s 1993 HOPE Scholarship/Grant programs on graduation rates. I use a differences-in-differences identification strategy to contrast the graduation rates of individuals in Georgia against those in other states after the HOPE Program was started. The main contribution of this paper is the better understanding it provides of the positive impacts of merit aid policies such as the HOPE Scholarship. I expect the rate of high school graduation to increase in Georgia relative to its peer states that lack a merit aid policy. High school graduation rates will increase because (1) the discounted cost of attending an institution of higher education is reduced substantially, and (2) the demand for labor in states with merit aid policies will gravitate towards students with higher educational attainment. I find that the enactment of the HOPE Scholarship increased the probability of students graduating from Georgia high schools by around 12.5%. In practical terms, this result echoes the conclusions of Henry and Rubinstein (2002) and Pallais (2009). In both instances, the performance of students on standardized tests increased in states where merit aid policies are instituted relative to peer states. It is not surprising that scores on standardized tests improved, since the eligibility conditions of merit aid rests on the students attaining a minimum cutoff on the tests. What is surprising is that distribution scores have shifted upwards; fewer students perform poorly on standardized tests when merit aid policies are enacted. Along similar lines, I also find the distribution of graduating grade point averages (GPAs) is shifting upwards towards the cutoff point. Background As part of legislation aiming to introduce a lottery into the state of Geor- gia, Governor Zell Miller in 1991 proposed a unique model where the bulk of the proceeds of the lottery scheme would be devoted to educational causes. In 1991,the Georgia legislature passed the lottery amendment while statewide residents nar- rowly passed a referendum supporting the amendment in 1992. This paved the way for the enactment of the HOPE Scholarship and the HOPE Grant programs. The HOPE Scholarship The HOPE Scholarship was created to encourage the academic achievement of Georgia’s high school students and Georgians seeking degrees from postsecondary institutions (Georgia Student Finance Commission, n.d.). Funding for the program is derived from Georgia Lottery for Education, and is administered by the 82 Spring 2017 Scientia et Humanitas: A Journal of Student Research The Impact of the HOPE Scholarship on High School Graduation in Georgia Middle Tennessee State University 83 Georgia Student Finance Commission. To qualify for a HOPE Scholarship, a student must be a U.S. citizen, a Georgia resident for at least 12 consecutive months prior to the award, and must graduate high school with a 3.0 GPA. To retain the scholarship, the student must maintain a 3.0 or B average throughout his/her academic program (limited to 6 years) with regular inspection by the state to ensure compliance. The scholarship is very generous—estimated around $4,700 in the 2003 academic year (Cornwell et al., 2006). When the HOPE Scholarship was first enacted, a student could get $1,000 towards tuition at private universities. This amount was later increased to $3,000 to reflect the growing trend of tuition increases. At the beginning of the program, eligibility was capped for students whose annual family income exceeds $66,000. The $66,000 hard cap was first increased to $100,000 in 1994 but then later abolished in 1995. The HOPE Grant This is the second plank of the lottery-funding program introduced under the HOPE Program. The HOPE Grant program provides assistance to residents of Georgia pursuing certificates or diplomas at Georgia’s public tertiary institutions. Its aim is to encourage residents to obtain practical training provided in 2-year colleges to broaden the talent pool of Georgia’s workforce. The eligibility criteria for the HOPE Grant program are the same as for the scholarship with one exception: students need not graduate with a 3.0 GPA. Merit Aid: Incentive, Effort, and High School Student Quality Bishop (1996) argues that pecuniary incentives that subsidize college costs are also expected to elicit more effort by high school students in order to meet or exceed the threshold of the merit aid policy. To demonstrate how this transpires, consider the following. Students cannot determine ex-ante if their academic effort will enable them scale the eligibility threshold. To obtain a grade point average sufficient for the merit aid scholarship, students who are either at or are perceived to be right below the threshold will likely exert more effort to scale that hurdle. That is, a student with a 2.9 GPA may exert moderately more effort to cross the 3.0 mark. At the same time, a student with a 2.5 GPA may substantially improve his/her effort in order to scale the threshold. Even if the student with the 2.5 GPA falls short of the cutoff (finishing with a 2.9 GPA, for example), that student’s probability of graduation nonetheless increases immensely. In essence, merit aid creates a positive externality that leads students to perform better, hence improving the chances of graduation (Betts, 1998; Kremer, Miguel, & Thornton, 2009) Indeed, the number of students graduating with a 2.90-2.99 GPA rose faster in Georgia than in surrounding states (Bugler, Henry, & Rubenstein, 1999). Another externality introduced by merit aid policies is that they improve the quality of high school students across the ability spectrum (Betts, 1998). Employers will realize that when standards are improved, the quality of students seeking to attain those standards is higher, and they therefore should be compensated differently. While the general consensus is that merit aid improves quality, Price suggests that merit aid simply rewards students who do not need the help (2001). He contends that since well-to-do students will attend college with or without merit aid, the policies do not lead to the attainment of their main objective: providing “equal opportunity.” Both higher achieving students through the HOPE Scholarship and lower achieving students through the HOPE Grant benefit from merit aid policy. Note that the main eligibility criteria of the HOPE Grant are high school graduation regardless of grade. Rational students who had no thought of obtaining post-secondary education are incentivized to graduate with the promise of a free technical education. The skills learned from these technical colleges would have taken years to obtain if the student had started working immediately upon graduation from high school ( Jepsen, Troske, & Coomes, 2014). Data and Empirical Specification Data I use data from three major sources in this study: the March Annual Social and Economic Supplement (ASEC) from the Current Population Survey (CPS);2 Common Core data from the National Center for Educational Statistics; and Local Area Unemployment Statistics from the Bureau of Labor Statistics. I pool ASEC data from 1993–2003 to create a large data set containing the constructed graduation status of all individuals. I create the graduation status variable as a binary variable coded as 1 for any individual in the data with an above 12th-grade education and coded as 0 for any individual who has a 12th-grade education and below. Due to a large sample size and compatibility issues, the analysis is limited to the 16 comparable states in the Southern Regional Educational Board (SREB)3 as done in Cornwell et al. (2006). 2. ASEC CPS provides the usual monthly labor force data, but in addition it also provides supplemen- tal data on work experience, income, noncash benefits, educational attainment, and migration. https:// www.census.gov/prod/techdoc/cps/cpsmar13.pdf 3.These states are Alabama, Arkansas, Delaware, Florida, Georgia, Kentucky, Louisiana, Maryland, Mississippi, North Carolina, Oklahoma, South Carolina, Tennessee, Texas, Virginia, and West Virginia. 84 Spring 2017 Scientia et Humanitas: A Journal of Student Research The Impact of the HOPE Scholarship on High School Graduation in Georgia Middle Tennessee State University 85 Table 1 shows the similarities between Georgia and the SREB states. I include as control variables parental income, school investments in facilities and teachers, and macroeconomic variables such as unemployment and minimum wage. Data on state expenditures were extracted from NCES’ Common Core data while unemployment figures were derived from the Bureau of Labor Statistics. All other variables were sourced from the CPS ASEC data. TABLE 1 Means of key variables in the SREB sample Georgia SREB Graduation Rates (%) 66 68 Hours Worked (Weekly) 16.8 17.54 Household Income ($) 70,015 66,764 Parents with HS Diploma (%) 85 81 State Per Capita Education Expenditures ($) 7,493 7034 Unemployment Rates 5.26 4.56 Youth Labor Force Participation (%) 53 57 No. Observations 329 6847 Notes. Income figures are expressed as 1995 dollars. I removed all entries with negative income figures and limited all the income variables to within three standard deviations of the mean. I limited hours worked weekly at 60. These adjustments did not alter the sample size significantly. To obtain an accurate representation of high school graduation, I construct a variable that looks at the current state of residence and tied it to the migratory history of the respondent. CPS data has a variable that captures the migration history of respondents over one-year intervals. For example, if a respondent indicates Alabama as his current state of residence and also indicates no history of migration over the past year, then this suggests that the state of residence and the state of high school graduation are the same. A secondary outcome variable that estimates whether or not the HOPE Program increases the probability of young adults joining the labor force is considered. This is a binary variable that indicates whether or not the respondent is in the labor force. CPS data also has demographic data about the respondent’s age, sex, gender, and family characteristics. I present some summary statistics in Table 2. TABLE 2 Breakdown of key variables by state STATES HOURS WORKED WEEKLY HOUSE- HOLD IN- COME ($) PARENTS WITH HS DIPLOMA (%) PER CAPITA EDUC EXP ($) UNP RATES (%) YOUTH LFP (%) #. OF OBS. ALABAMA 16.5 63,923 84 6,239 5.6 58 337 ARKANSAS 17.47 58,599 83 6,078 5.4 64 289 DELAWARE 20.1 77,302 86 9,599 4.2 58 294 FLORIDA 16.9 66,680 84 7,287 5.1 60 988 KENTUCKY 16.98 67,889 82 6,338 5.4 62 362 LOUISIANA 15.7 65,556 85 6,304 6.4 60 296 MARYLAND 19.57 93,635 93 9,012 6.4 56 286 MISSISSIPPI 14.12 56,132 76 5,351 6 50 316 NORTH CAROLINA 19.27 69,643 83 6,473 4.7 59 617 OKLAHOMA 17.66 65,581 88 5,980 4.5 59 368 SOUTH CAROLINA 16.97 70,885 82 7,280 5.6 55 305 TENNESSEE 18.33 71,118 84 6,041 4.9 60 292 TEXAS 18 59,325 70 7,215 5.7 55 1318 VIRGINIA 19.7 78,689 86 7,654 3.9 54 382 WEST VIRGINIA 15.12 57,287 84 7,790 6.9 50 397 GEORGIA 16.8 70,015 86 7,492 4.5 53 329 Notes. PER CAPITA EDUC EXP. Represents state expenditures per pupil. UNP RATES are mean unemployment rates. YOUTH LFP stands for youth labor force participation rates. NO. OF OBS. stands for number of observations Empirical Methodology The aim of this study is to examine whether Georgia’s merit aid (the HOPE Scholarship) has increased the rate of high school graduation. At first blush, the answer seems mechanically simple. When merit aid is available, more students will graduate from high school in pursuit of the program’s benefits. The eligibility criteria, which requires students to attain a 3.0 GPA, presents a twist on this presumption. That is, students with a 3.0 GPA or above typically graduate from high school with or without merit aid. Hence, the graduation rate of students with a GPA that exceeds the threshold will not be affected by the policy. Those at the margin and those within striking distance of the cutoff are the ones most affected by the policy. 86 Spring 2017 Scientia et Humanitas: A Journal of Student Research The Impact of the HOPE Scholarship on High School Graduation in Georgia Middle Tennessee State University 87 Therefore, the identifying assumption is that upon enacting the HOPE Program, high school students in Georgia will graduate at higher rates than their counterparts in surrounding states with similar characteristics. I use a differenc- es-in-differences method similar to its application in Singell, Waddell, and Curs (2006) to determine the impact of the HOPE Program on high school graduation: Grad it = ∂ (GA* HOPE)+yXi + αT+βs (ST)+ εit We compare each individual i at time t, to determine whether the policy as implemented causes the graduation status to differ between Georgia and the sur- rounding states. ST are state level macro variables that change over time, such as unemployment and education expenditures. GA is an indicator for the state of Georgia, and HOPE is an indicator for years after 1994. Even though the HOPE Scholarship was enacted in 1993, I am constrained to use 1995 as a cutoff due to the data setup. By design, the ASEC CPS data is collected in March of every year. Any respondent that indicates he or she has graduated must be referring to his status based on the previous year. That is, if a respondent in March 1995 indicates having completed 12th-grade, then it must mean that he or she graduated around June 1994. The coefficient of interest is HOPE.X, is a vector of individual level covariates, and is the error term. I include a linear time trend to control for secular growth in graduation rates. Graduation rates in the U.S. have been rising due to higher college premium, the sophistication of the U.S. economy, and laws aimed at keeping students in school. This time trend variable is expected to wash away most of the up-tick in graduation rates coming from other sources. NCES data show large disparities in graduation rates between male and female, white and black, and low-income and high-income groups (NCES, 2011). This makes it imperative to analyze the sample with respect to these dimensions. While it is hard to detach the impact of class from the impact of race on high school graduation rates, Storer et al. (2012) still find the difference in drop-out rates on the basis of race statistically significant. Results and Discussion Full Sample I initially ran a naive model of equation (1) above in order to obtain a sense of the magnitude and direction of the coefficients. To do this, I ran a simple logit without any non-student covariates, hence obtaining the summary estimates of the model as seen in column 1 of Table 3. The reported coefficients are the marginal effects of the policy. Judging by the low F statistic of the specification, the model is not properly specified. The coefficient shows the positive correlation between the enactment of the HOPE Program and higher graduation rates. TABLE 3 Marginal Impact of merit aid policies on high school graduation (1) (2) (3) (4) (5) GA*HOPE 0.168** 0.121** 0.138** 0.125* 0.099* Parental Characteristics X Y Y Y Y Individual Characteristics X X Y Y Y State Level Covariates X X X Y Y No. of Observations 7176 7,176 7176 7,176 7176 Pseudo R2 0.05 0 0.19 0 0.26 Notes. *** p< 0.01, ** p< 0.05, *p < 0.1. I estimate a second model that measures parental income and present the results in column (2) of Table 3. This shows the enactment of the HOPE Schol- arship in Georgia has led to a 12.1% increase in graduation rates. I improve the model further by adding variables dealing with parental education as controls to determine the causal impact of the HOPE Scholarship. This is motivated by the work of Foley, Gallipoli, and Green (2014) who argued that there is a strong causal relationship between the educational attainment of parents and the graduation deci- sion of their children. After adding parental education measures, the coefficient for increases from 12.1% as seen in model (2) to 13.8% in this specification (column 3, Table 3). The fourth model is my preferred specification. It incorporates macroeco- nomic variables in addition to the covariates mentioned above. Column 4 of Table 5 shows the result of this analysis. It shows that the marginal effect of instituting the HOPE Scholarship program on graduation rates in Georgia is around 12.5% significant at a 90% confidence level. These results are comparable but smaller in magnitude to Keane and Wolpin (2000) who conclude that a proposed $25,000 subsidy to low-income households can decrease high school drop out rates by two-thirds. 88 Spring 2017 Scientia et Humanitas: A Journal of Student Research The Impact of the HOPE Scholarship on High School Graduation in Georgia Middle Tennessee State University 89 Other groups Table 4 presents the estimates of the impact of HOPE scholarships/grants on high school graduation for female students, black students, and low-income groups. The first column of Table 4 shows the regression results when the sample is restricted to female students. The marginal effect (0.143) is higher when compared to the full sample (0.125). The second column shows the result of the analysis when restricted to black students. The results show the HOPE Scholarship increasing graduation rates of black students by 36%. The analysis on black students may be spurious due to small sample size. It is expected that, as in the case of females, black students should respond more to the policy. Column 3 of Table 4 presents the result of the analysis on lower income student’s regression. TABLE 4 Marginal Impact of merit aid policies on different populations Female Students Black Students Lower Income Upper Income Low Educ. Highly Educ. GA*HOPE 0.143** 0.36+ 0.23** 0.07 0.042 0.158** (1.97) (4.57) (2.24) (0.91) (0.40) (2.44) Parental Characteristics Y Y Y Y Y Y Individual Characteristics Y Y Y Y Y Y State Level Covariates Y Y Y Y Y Y No. of Observations 3293 1,015 3561 3506 3,576 3600 Pseudo R2 0.22 0.18 0.14 0.24 0.16 0.23 Notes. *** p< 0.01, ** p< 0.05, *p < 0.10. The numbers is parentheses are clusters standard errors. +The results from the black student sample may suffer from low sample size. The marginal effect of 0.23 indicates an above average impact for this group. The implied subsidy provided by the HOPE Program is much higher for lower income families as a percentage of their income. It is not surprising that this group responds at a higher rate than their wealthier counterparts. These results are consistent with Angrist and Lavy (2002), who also find that the impact of high school cash incentives leads to stronger graduation rates for low-achieving students. Discussion The social and economic gains of having more graduates in a state are both substantial and long lasting (Catterall, 2011; Levin et al., 2007). However, those who drop out and those who graduate differ in some important characteristics (Eckstein & Wolpin, 1999). Factors such as a healthy labor market may incentivize students to drop out (McNeal, 2011). As such, providing a financial incentive may encourage students to stay in school and graduate. The impact of a comprehensive policy like the HOPE Program, which incen- tivizes college enrollment, should also encourage students to graduate from high school. The estimates presented in Tables 3 and 4 show graduation rates increasing by around 10% to 12.5% when the HOPE Program was instituted. The results also confirm the heterogeneity of the policy’s impact on different groups. Disaggregating the results for different subgroups is very useful, especially as it relates to formulating policy objectives because different groups have different elasticities to the policy. For instance, we can see from Table 4 that lower income groups have a higher response rate to the policy when compared to the full group. The results suggest that around 5,793 more students are graduating from high school in Georgia as a result of the HOPE Program. This large increase in the number of high school graduates has significant impacts on the economy of the state. Using D’andrea’s (2010) baseline estimates, every high school dropout is expected to cost the state $1,100 in Medicare costs and $950 in incarceration costs annually. At this rate, Georgia could save around $6.3 million in Medicaid costs and about $5.5 million in incarceration expenses per year. Conclusion I analyze the impact of merit aid policy in Georgia (HOPE) on high school graduation rates. Unlike many studies on this policy, my analysis seeks to find the pre-college benefits of the policy. Results from differences-in-differences shows a 12.5% increase in high school graduation due to the policy. The responsiveness to the policy is more pronounced on the traditionally lower graduating groups—women, minorities, and students from poorer backgrounds. While the results presented here are robust, care must be taken in extrapolating the results to other states. Because Georgia was the first to implement the policy, it is uncomplicated to use Georgia as a treatment. States that have not implemented a similar policy were similarly uncomplicated to use as controls. Many states unveiled several policies to improve graduation rates in the 1990s, so extending the same model 90 Spring 2017 Scientia et Humanitas: A Journal of Student Research The Impact of the HOPE Scholarship on High School Graduation in Georgia Middle Tennessee State University 91 to other states will be susceptible to confounding effects. What can be concluded from this analysis is that when incentives are provided encouraging students to enroll in college, these same policies will have a positive effect on all students—not just those who will eventually enroll in college. The Impact of the HOPE Scholarship on High School Graduation in Georgia 92 Spring 2017 References Angrist, J. D., & Lavy, V. (2002). The effect of high school matriculation awards: Evidence from randomized trials. (NBER Working Paper No. 9389). Cambridge, MA: National Bureau of Economic Research. Retrieved from http://dx.doi. org/10.3386/w9389 Betts, J. R. (1998). The impact of educational standards on the level and distribution of earnings. The American Economic Review, 88(1), 266-275. Bishop, J. (1996). Signaling, incentives, and school organization in France, the Netherlands, Britain, and the United States. In E. Hanushek and D. Jorgenson (Eds.), Improving America’s Schools: The Role of Incentives (pp. 111- 146). Washington, DC: National Academy Press. Bugler, D. T., Henry, G. T., & Rubenstein, R. (1999). An evaluation of Georgia’s HOPE Scholarship Program: Effects of HOPE on grade inflation, academic performance, and college enrollment. Retrieved from Council for School Performance: http://www.issuelab.org/resources/4112/4112.pdf Catterall, J. S. (2011). The societal benefits and costs of school dropout recovery Education Research International, 2011, 1-8. Retrieved from http://dx.doi. org/10.1155/2011/957303 Chapman, C., Laird, J., Ifill, N., & KewalRamani, A. (2011). Trends in high school dropout and completion rates in the United States: 1972-2009. Compendium report (NCES 2012-006). Washington, DC: National Center for Education Statistics. Cornwell, C., Mustard, D. B., & Sridhar, D. J. (2006). The enrollment effects of merit-based financial aid: Evidence from Georgia’s HOPE Program. Journal of Labor Economics, 24(4), 761-786. D’andrea, C. (2010). Tennessee’s high school dropouts: Examining the fiscal consequences. The Foundation for Educational Choice. Retrieved from http:// www.edchoice.org/wp-content/uploads/2015/09/Tennessees-High- School-Dropouts-Examin ing-the-Fiscal-Consequences.pdf Dynarski, S. (2002). The behavioral and distributional implications of aid for college. American Economic Review, 92(2), 279-285. Retrieved from http://users.nber. org/~dynarski/2002%20Behavioral.pdf Eckstein, Z., & Wolpin, K. (1999). Why youths drop out of high school: The impact of preferences, opportunities, and abilities. Econometrica, 67(6), 1295-1339. Middle Tennessee State University 93 Scientia et Humanitas: A Journal of Student Research Fenske, R. H., Geranios, C. A., Keller, J. E., & Moore, D. E. (1997). Early intervention problems: Opening the doors to higher education (ASHE-ERIC Higher Education Report, vol. 25, no. 6). Washington, DC: Office of Educational Research and Improvement. Foley, K., Gallipoli, G., & Green, D.A. (2014). Ability, parental valuation of education, and the high school dropout decision. Journal of Human Resources, 49(4), 906-944. Retrieved from http://muse.jhu.edu.ezproxy.mtsu.edu/ article/558437 Georgia Student Finance Commission. (n.d). HOPE Programs. Retrieved from https://gsfc.georgia.gov/HOPE Jepsen, C., Troske, K., & Coomes, P. (2014). The labor-market returns to community college degrees, diplomas, and certificates. Journal of Labor Economics, 32(1), 95-121. doi:10.1086/671809 Henry, G. T., & Rubenstein, R. (2002). Paying for grades: Impact of merit-based financial aid on educational quality. Journal of Policy Analysis and Management, 21(1), 93-109. Keane, M. P., & Wolpin, K. I. (2000). Eliminating race differences in school attainment and labor market success. Journal of Labor Economics, 18(4), 614- 652. Kremer, M., Miguel, E., & Thornton, R. (2009). Incentives to learn. The Review of Economics and Statistics, 91(3), 437-456. Retrieved from http://www.mitpressjournals.org/doi/pdf/10.1162/rest.91.3.437 Levin, H., Belfield, C., Muennig, P., & Rouse, C. (2007). The costs and benefits of an excellent education for all of America’s children. Retrieved from http://www3. nd.edu/~jwarlick/documents/Levin_Belfield_Muennig_Rouse.pdf McNeal Jr., R. B. (2011). Labor market effects on dropping out of high school: Variation by gender, race, and employment status. Youth & Society, 43(1), 305-332. Retrieved from https://eric.ed.gov/?id=EJ915799 Murnane, R. J. (2013). U.S. high school graduation rates: Patterns and explanations. Journal of Economic Literature, 51(2), 370-422. Pallais, A. (2009). Taking a chance on college: Is the Tennessee Education Lottery Scholarship Program a winner? Journal of Human Resources, 44(1), 199-222. Retrieved from http://muse.jhu.edu.ezproxy.mtsu.edu/article/467071/pdf Price, D. V. (2001). Merit aid and inequality: Evidence from baccalaureate & beyond. Journal of Student Financial Aid, 31(2), 5-18. Retrieved from https://eric. ed.gov/?id=EJ643306 Singell Jr., L. D., Waddell, G. R., & Curs, B. R. (2004). HOPE for the Pelll? The impact of merit-aid on needy students. Southern Economic Journal, 73, 79-99. Retrieved from https://scholarsbank.uoregon.edu/xmlui/ bitstream/handle/1794/237/UO-2004-5_Singell_HOPE _Pell. pdf ?sequence=1&isAllowed=y Storer, H. L., Mienko, J. A., Chang, Y., Kang, J. Y., Miyawaki, C., & Schultz, K. (2012). Moving beyond dichotomies: How the intersection of race, class and place impacts high school graduation rates for African American students. Journal of Sociology & Social Welfare, 39(1), 17-44. Retrieved from http://scholarworks.wmich.edu/cgi/viewcontent. cgi?article=3648&context=jssw 94 Spring 2017 Scientia et Humanitas: A Journal of Student Research