Academic Commons

Theses Doctoral

Essays on Improving STEM Academic Outcomes and Reducing Gender and Race Graduation Gaps: The Effects of College Grades and Grading Policies

Minaya Lazarte, Veronica Milagros

A college degree is not a homogenous investment across fields of study (Arcidiacono, 2004; Zhang & Thomas, 2005). Even after accounting for selection, STEM degrees pay substantially more than other fields (Altonji et al., 2012) and earnings disparities across majors have increased substantially over time (Altonji et al., 2014).
Even though STEM degrees yield greater labor market returns, the number of STEM graduates and professionals remains low and the disparities in STEM attrition are alarming. As a result, STEM education has been elevated as a national priority in the U.S. and considered to be in high demand in the global economy. Yet, there is a lack of consensus on how to boost STEM graduation.
My dissertation is motivated by the need to improve the number and composition of STEM graduates and to evaluate policies that can mitigate STEM attrition. In my dissertation I focus on the effect of college grades and grading policies on STEM graduation. College grades are important determinants of course and major choices and research suggests that grades have differing effects for STEM minorities and non-minorities. Moreover, disparities in grades between STEM (low-grading departments) and non-STEM (high-grading departments) due to grade inflation and compression of grades near the top affect sorting into majors, making grades less informative and distorting major choices (Bar et al. 2012).
In my first essay, I examine the possible differential effect of college grades on STEM attrition gap by gender and race. Non-grade explanations such as pre-college factors, instructor gender and race and peer effects are also examined as potential determinants of STEM attrition gaps. However, I focus on grades because there is evidence that grades affect sorting into majors, and grades may have differing effects for minorities and non-minorities. This review uncovers evidence supporting the importance of institutional grading policies to shape student’s major and course choices. Despite the fact that institutional grading policies have been studied at some extent, none of these studies have addressed the differential effect of these policies on those who might be more sensitive to grades (i.e., women and racial minorities).
In the second essay, I explore what factors explain the gender and race disparities in STEM attrition. This study utilize Florida’s Education Data Warehouse to conduct a reweighted Oaxaca decomposition of racial and gender differences in STEM attrition, with a particular focus on how STEM- intending students respond to college grades in introductory courses. The decomposition results show that women mainly leave STEM by switching into non-STEM fields, particularly due to non-STEM college factors such as grades and credits attempted in lower-division courses. In contrast, racial minorities mainly leave STEM by dropping out of college towards graduation, and they differentially leave STEM due to their lower high school preparation in STEM and consequently lower grades in lower-division STEM courses during their first two years of enrollment.
In the third essay (which is also my job market paper) I examine the effect of changing the grading scale from whole-letter grades to plus/minus grades on STEM graduation/major choice. In this study, I examine the effect of changing the grading scale from whole-letter grades to plus/minus grades on STEM graduation/major choice. I use administrative data from the Florida Department of Education that combines students’ pre-college characteristics with students’ enrollment and transcript records. I rely on a difference in differences framework that compares STEM graduation/major choice rates during the early 2000s versus the late 1990s for students whose grading differentials between STEM and non-STEM courses were reduced versus students whose grades were not differentially affected. I find significant effects of changing the grading scale on reducing grading differentials and improving STEM graduation/major choice. These results represent the first direct, quasi-experimental evidence regarding the effect of changing the grading scale.


  • thumnail for MinayaLazarte_columbia_0054D_13333.pdf MinayaLazarte_columbia_0054D_13333.pdf binary/octet-stream 2.85 MB Download File

More About This Work

Academic Units
Economics and Education
Thesis Advisors
Scott-Clayton, Judith E.
Ph.D., Columbia University
Published Here
May 4, 2016