Losing Missouri Access: The effects of losing state needs-based financial aid due to academic criteria on student success

Presenter: 
Junpeng Yan, University of Missouri, jyy7f@mail.missouri.edu

The state of Missouri has three primary financial aid programs: Missouri Access (a need-based aid program), Bright Flight (a merit-based aid program), and the A+ Scholarship Program (a hybrid program with a community college focus). While each program has very different initial eligibility requirements, the year-to-year renewal requirement is the same for each program: to maintain a 2.5 cumulative grade point average (GPA) and maintain satisfactory academic progress as defined by a student’s school.
Different from the other two programs, the need-based Missouri Access grant, is initially assigned to applicants from low-income families, without considering their academic preparation. The initial eligibility is determined by the Expected Family Contribution (EFC) calculation through the Free Application for Federal Student Aid (FAFSA). Applicants whose EFC was below $12,000 were awarded grant aid. However, the subsequent renewal of the Missouri Access grant is dependent upon continuing to meet the initial eligibility (EFC < $12,000) and meeting an academic performance requirement (GPA > 2.5). This academic requirement for the renewal of a need-based financial aid program presents an interesting theoretical and policy design puzzle with little prior empirical evidence (Schudde & Scott-Clayton, 2016; Scott-Clayton & Schudde, 2019). On one hand, the performance requirement can motivate recipients to maintain good academic standing in their programs and increase aid efficiency. On the other hand, students from low-income families may be less-prepared to enter college, making it difficult to subsequently meet the renewal requirement. Losing their Missouri Access award may force them to leave higher education and jeopardize aid effectiveness.
In this study we explore the consequences of losing state-based financial aid due to not meeting the 2.5 GPA requirement for aid renewal on measures of academic success in the subsequent year: enrollment, credit hours attempted and completed, and grade point average. We apply a regression discontinuity (RD) design to state administrative microdata, which includes all first-time, full-time, and degree-seeking students from 13 Missouri 4-year public universities during the 2007-2012 academic years. Cumulative GPA serves as the running variable in our RD design with a GPA of 2.5 serving as the policy cutoff.
We utilize an intent-to-treat framework with estimates interpreted as the effect of being renewal eligible on subsequent academic outcomes. Preliminary findings suggest that being renewal eligible does not change the likelihood a student is enrolled within the Missouri public higher education system but does increase the likelihood that a student transfers to other public institutions within the system. Furthermore, there is no effect on subsequent year outcomes (credits hours and future GPA) conditional on being retained within the same institution. A possible explanation of these confounding results is, if the students who just pass the threshold but do not have the confidence to keep their future renewal eligibility, they may transfer to other institutions to receive Missouri Access continuously. And the rest students are more likely encouraged by the financial aid and then get a better grade in the following semester.
Our empirical data demonstrates grading policies vary among different institutions, with evidence of potential GPA manipulation around the cutoff for 4 of 13 institutions. After exclusion of these institutions, our results still confirm that being renewal eligible may discourage students to be retained in the same institution but also motivate the students to achieve a higher GPA if they choose to stay.
To address our contributions to the literature and practice, we notice that a few previous articles have explored the effects of non-renewal of financial aid on GPA cutoffs which coincide with satisfactory academic progress policy and/or institutional probation policy (Schudde & Scott-Clayton, 2016; Scott-Clayton & Schudde, 2019). Differently, this study of the effect of non-renewal of financial aid contributes to our understanding of the effects of losing financial aid by focusing on a local treatment effect that occurs at a GPA cutoff for which the student was otherwise making progress towards graduation. Thus, the findings of this study also have state policy implications related to whether incentive-based renewal cutoffs improve (through incentives to meet a 2.5 GPA) or hinder (by inducing dropout of successful students) state degree production.
Reference
Schudde, L., & Scott-Clayton, J. (2016). Pell grants as performance-based scholarships? An examination of satisfactory academic progress requirements in the nation’s largest need-based aid program. Research in Higher Education, 57(8), 943–967. https://doi.org/10.1007/s11162-016-9413-3
Scott-Clayton, J., & Schudde, L. (2019). The consequences of performance standards in need based aid: Evidence from community colleges. Journal of Human Resources, (January), 0717-8961r2. https://doi.org/10.3368/jhr.55.4.0717-8961r2

Comments

A really interesting study that spurred a couple of questions: (1) You say that you see evidence of manipulation at a few of the campus, that's the issue being 'addressed' on the campus side. Do you see any evidence of it being 'addressed' on the student side: i.e. through selective course taking patterns? (to ensure maintaining eligibility through easier classes) What about when students transfer? Any evidence that is happening? (You imply that are 'working harder' but maybe they are just savvier education consumers). (1a) Related to the above, do you see different patterns based on student year? i.e. are your effects about the same for the freshman -> sophomore year as for the sophomore -> junior or junior -> senior transition? Presumably students have more information about their abilities the further along they get. (2) Are there any other operative policies that have 2.5 GPA cutoffs? (i.e. are you sure that this is the only game in town when it comes to motivating GPA manipulation or transfer...NCAA eligibility is 2.3 (Div I) but at many schools you need a minimum GPA to apply to be, say, a business major...any evidence that is also a factor here? (One could make an argument that 'stacking' or aligning cut-points like this is a good thing rather than a bad thing)

This is a good start. I'd look next at the specific kinds of colleges are enrolling in. Since the effect seems to be on same-institution enrollment, we would expect students to move to cheaper and probably "lower quality" institutions. (I assume that the aid can be used at all MO institutions. If not, then institutional eligibility would be another thing to examine.)

I want to echo Ethan Hutt's point that the GPA manipulation itself could be interesting to study. In other words, it reveals something about the value students place on this aid. Don't necessarily view it as an RD limitation so much as a different phenomenon worthy of study. Heaping is interesting!

It's interesting that students transferred in-state at higher rates after losing the grant given the statewide criteria. Looking at where students enrolled next is really important.

Thanks for sharing this work! I think it would be useful to also think about exploring a bit about the "fuzziness on both sides" that is mentioned in the poster under Figure 1. It might be nice to descriptively explore where that fuzziness comes from (is it the EFC on the right hand side?). Beyond the points Ethan brought up about the manipulation which I think are quite useful, it could be useful to keep in mind is that the assignment mechanism for this regression discontinuity design actually has 2 forcing variables: GPA and EFC. Is there any way to model for both of those forcing variables? You could explore the options presented in Porter et al. (2017). Estimating causal effects of education interventions using a two-rating regression discontinuity design: Lessons from a simulation study and an application. Journal of Research on Educational Effectiveness, 10(1), 138-167. Or y'all could explore a fuzzy regression discontinuity. Thanks again for sharing this work and let me know if I can be helpful in the future!

Add new comment