We study a game theoretic model of standardized testing for college admissions. Students are of two types; High and Low. There is a college that would like to admit the High type students. Students take a potentially costly standardized exam which provides a noisy signal of their type.
The students come from two populations, which are identical in talent (i.e. the type distribution is the same), but differ in their access to resources: the higher resourced population can at their option take the exam multiple times, whereas the lower resourced population can only take the exam once. We study two models of score reporting, which capture existing policies used by colleges. The first policy (sometimes known as "super-scoring") allows students to report the max of the scores they achieve. The other policy requires that all scores be reported.
We find in our model that requiring that all scores be reported results in superior outcomes in equilibrium, both from the perspective of the college (the admissions rule is more accurate), and from the perspective of equity across populations: a student's probability of admission is independent of their population, conditional on their type. In particular, the false positive rates and false negative rates are identical in this setting, across the highly and poorly resourced student populations. This is the case despite the fact that the more highly resourced students can -- at their option -- either report a more accurate signal of their type, or pool with the lower resourced population under this policy.
more »
« less
Downstream Effects of Affirmative Action
We study a two-stage model, in which students are 1) admitted to
college on the basis of an entrance exam which is a noisy signal
about their qualifications (type), and then 2) those students who
were admitted to college can be hired by an employer as a function
of their college grades, which are an independently drawn noisy signal of their type. Students are drawn from one of two populations,
which might have different type distributions. We assume that the
employer at the end of the pipeline is rational, in the sense that
it computes a posterior distribution on student type conditional
on all information that it has available (college admissions, grades,
and group membership), and makes a decision based on posterior
expectation. We then study what kinds of fairness goals can be
achieved by the college by setting its admissions rule and grading
policy. For example, the college might have the goal of guaranteeing equal opportunity across populations: that the probability
of passing through the pipeline and being hired by the employer
should be independent of group membership, conditioned on type.
Alternately, the college might have the goal of incentivizing the
employer to have a group blind hiring rule. We show that both
goals can be achieved when the college does not report grades. On
the other hand, we show that under reasonable conditions, these
goals are impossible to achieve even in isolation when the college
uses an (even minimally) informative grading policy
more »
« less
- Award ID(s):
- 1763307
- NSF-PAR ID:
- 10100410
- Date Published:
- Journal Name:
- ACM FAT* 2019
- Page Range / eLocation ID:
- 240 to 248
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
We present a stylized model with feedback loops for the evolution of a population's wealth over generations. Individuals have both talent and wealth: talent is a random variable distributed identically for everyone, but wealth is a random variable that is dependent on the population one is born into. Individuals then apply to a downstream agent, which we treat as a university throughout the paper (but could also represent an employer) who makes a decision about whether to admit them or not. The university does not directly observe talent or wealth, but rather a signal (representing e.g. a standardized test) that is a convex combination of both. The university knows the distributions from which an individual's type and wealth are drawn, and makes its decisions based on the posterior distribution of the applicant's characteristics conditional on their population and signal. Each population's wealth distribution at the next round then depends on the fraction of that population that was admitted by the university at the previous round. We study wealth dynamics in this model, and give conditions under which the dynamics have a single attracting fixed point (which implies population wealth inequality is transitory), and conditions under which it can have multiple attracting fixed points (which implies that population wealth inequality can be persistent). In the case in which there are multiple attracting fixed points, we study interventions aimed at eliminating or mitigating inequality, including increasing the capacity of the university to admit more people, aligning the signal generated by individuals with the preferences of the university, and making direct monetary transfers to the less wealthy population.more » « less
-
Chaudhuri, Kamalika ; Jegelka, Stefanie ; Song, Le ; Szepesvari, Csaba ; Niu, Gang ; Sabato, Sivan (Ed.)Recent work highlights the role of causality in designing equitable decision-making algorithms. It is not immediately clear, however, how existing causal conceptions of fairness relate to one another, or what the consequences are of using these definitions as design principles. Here, we first assemble and categorize popular causal definitions of algorithmic fairness into two broad families: (1) those that constrain the effects of decisions on counterfactual disparities; and (2) those that constrain the effects of legally protected characteristics, like race and gender, on decisions. We then show, analytically and empirically, that both families of definitions almost always—in a measure theoretic sense—result in strongly Pareto dominated decision policies, meaning there is an alternative, unconstrained policy favored by every stakeholder with preferences drawn from a large, natural class. For example, in the case of college admissions decisions, policies constrained to satisfy causal fairness definitions would be disfavored by every stakeholder with neutral or positive preferences for both academic preparedness and diversity. Indeed, under a prominent definition of causal fairness, we prove the resulting policies require admitting all students with the same probability, regardless of academic qualifications or group membership. Our results highlight formal limitations and potential adverse consequences of common mathematical notions of causal fairness.more » « less
-
First-generation (FG) and/or low-income (LI) engineering student populations are of particular interest in engineering education. However, these populations are not defined in a consistent manner across the literature or amongst stakeholders. The intersectional identities of these groups have also not been fully explored in most quantitative-based engineering education research. This research paper aims to answer the following three research questions: (RQ1) How do students’ demographic characteristics and college experiences differ depending on levels of parent educational attainment (which forms the basis of first-generation definitions) and family income? (RQ2) How do ‘first-generation’ and ‘low-income’ definitions impact results comparing to their continuing-generation and higher-income peers? (RQ3) How does considering first-generation and low-income identities through an intersectional lens deepen insight into the experiences of first-generation and low-income groups? Data were drawn from a nationally representative survey of engineering juniors and seniors (n = 6197 from 27 U.S. institutions). Statistical analyses were conducted to evaluate respondent differences in demographics (underrepresented racial/ethnic minority (URM), women, URM women), college experiences (internships/co-ops, having a job, conducting research, and study abroad), and engineering task self-efficacy (ETSE), based on various definitions of ‘first generation’ and ‘low income’ depending on levels of parental educational attainment and self-reported family income. Our results indicate that categorizing a first-generation student as someone whose parents have less than an associate’s degree versus less than a bachelor’s degree may lead to different understandings of their experiences (RQ1). For example, the proportion of URM students is higher among those whose parents have less than an associate’s degree than among their “associate’s degree or more” peers (26% vs 11.9%). However, differences in college experiences are most pronounced among students whose parents have less than a bachelor’s degree compared with their “bachelor’s degree or more” peers: having a job to help pay for college (55.4% vs 47.3%), research with faculty (22.7% vs 35.0%), and study abroad (9.0% vs 17.3%). With respect to differences by income levels, respondents are statistically different across income groups, with fewer URM students as family income level increases. As family income level increases, there are more women in aggregate, but fewer URM women. College experiences are different for the middle income or higher group (internship 48.4% low and lower-middle income vs 59.0% middle income or higher; study abroad 11.2% vs 16.4%; job 58.6% vs 46.8%). Despite these differences in demographic characteristics and college experiences depending on parental educational attainment and family income, our dataset indicates that the definition does not change the statistical significance when comparing between first-generation students and students who were continuing-generation by any definition (RQ2). First-generation and low-income statuses are often used as proxies for one another, and in this dataset, are highly correlated. However, there are unique patterns at the intersection of these two identities. For the purpose of our RQ3 analysis, we define ‘first-generation’ as students whose parents earned less than a bachelor’s degree and ‘low-income’ as low or lower-middle income. In this sample, 68 percent of students were neither FG nor LI while 11 percent were both (FG&LI). On no measure of demographics or college experience is the FG&LI group statistically similar to the advantaged group. Low-income students had the highest participation in working to pay for college, regardless of parental education, while first-generation students had the lower internship participation than low-income students. Furthermore, being FG&LI is associated with lower ETSE compared with all other groups. These results suggest that care is required when applying the labels “first-generation” and/or “low-income” when considering these groups in developing institutional support programs, in engineering education research, and in educational policy. Moreover, by considering first-generation and low-income students with an intersectional lens, we gain deeper insight into engineering student populations that may reveal potential opportunities and barriers to educational resources and experiences that are an important part of preparation for an engineering career.more » « less
-
The Greater Equity, Access, and Readiness for Engineering and Technology (GEARSET) Program, an NSF funded S-STEM program was developed GEARSET to address several institutional needs at the university. The original target population for the GEARSET program was identified as a subset of the students who applied to the College of Engineering and do not meet all the admissions requirements and are admitted to an Exploratory Studies major in the university’s University College. Historical data indicates that approximately 170 students per year with a high school GPA of 3.00 or higher are admitted to Exploratory Studies because they do not meet the College of Engineering admissions criteria. Of these, roughly 78 students remain at the University after one year. Of those 78, only about 45 students per year transition to college of Engineering majors by the end of their first year. These numbers do not accurately reflect the ability of these students, but rather are due in part to curricular bottlenecks, lack of institutional support, and lack of significant relevant exposure of students to material meant to engage their engineering future selves. This data motivated the creation of the GEARSET program. Specifically, the program was designed to 1. Increase recruitment, retention, student success, and transfer rates into engineering of students who are not admitted directly to engineering but who are instead admitted to the university’s University College. 2. Increase meaningfulness and engineering relevance of pre-engineering curriculum. 3. Increase diversity within the student population of various engineering departments in the College of Engineering. 4. Remove bottlenecks in curriculum and improve access to engineering and decrease length to degree. A key aspect of the program is a curated curriculum. All students in the GEARSET program are enrolled in multiple courses historically proven to promote better understanding of the key areas of Math, Chemistry and Physics needed to be successful engineers. All students have access to advisors within the COE to help them better understand the programs, curriculum and professional outcomes of each discipline of Engineering. Another key component of the program is that low income students in the GEARSET cohort who successfully transfer to a major within the COE after one year receive scholarship support. Here we describe the Program, the results to date, and the impact of the recent global pandemic and the subsequent transition to test optional admissions criteria on the definition of the GEARSET cohort, program implementation, and student participation.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- https://doi.org/10.1145/3287560.3287578
