The Problem of Infra-marginality in Outcome Tests for Discrimination⇤
نویسندگان
چکیده
Outcome tests are a popular method for detecting bias in lending, hiring, and policing decisions. These tests operate by comparing the success rate of decisions across groups. For example, if loans made to minority applicants are observed to be repaid more often than loans made to whites, it suggests that only exceptionally qualified minorities are granted loans, indicating discrimination. Outcome tests, however, are known to su↵er from the problem of infra-marginality : even absent discrimination, the repayment rates for minority and white loan recipients might di↵er if the two groups have di↵erent risk distributions. Thus, at least in theory, outcome tests can fail to accurately detect discrimination. We develop a new statistical test of discrimination—the threshold test—that mitigates the problem of infra-marginality by jointly estimating decision thresholds and risk distributions. Applying our test to a dataset of 4.5 million police stops in North Carolina, we find that the problem of infra-marginality is more than a theoretical possibility, and can cause the outcome test to yield misleading results in practice. ⇤We thank Cheryl Phillips and Vignesh Ramachandran of the Stanford Computational Journalism Lab for helping to compile the North Carolina tra c stop data, and the John S. and James L. Knight Foundation for partial support of this research. We also thank Stefano Ermon, Avi Feller, Seth Flaxman, Andrew Gelman, Lester Mackey, Jan Overgoor, and Emma Pierson for helpful comments. Our dataset of North Carolina tra c stops is available at https://purl.stanford.edu/nv728wy0570, and our analysis code is available at https://github.com/5harad/ threshold-test. This paper is forthcoming in the Annals of Applied Statistics.
منابع مشابه
Testing for Racial Discrimination in Police Searches of Motor Vehicles ⇤
In the course of conducting tra c stops, o cers have discretion to search motorists for drugs, weapons, and other contraband. There is concern that these search decisions are prone to racial bias, but it has proven di cult to rigorously assess claims of discrimination. Here we develop a new statistical method—the threshold test—to test for racial discrimination in motor vehicle searches. We use...
متن کاملTwo-Point Discrimination Test in the Treatment of Right-handed Females with Lumbosacral Radiculopathy
Non-somatic causes of pain may aggravate painful complaints and complicate the conservative management of diseases such as lumbo-sacral disk root disease. The two-point discrimination test has been used for assessment of diseases, which change patients' skin sensation. This study aims to find out how applicable is two-point discrimination test in the conservative treatment of lumbo-sacral disk ...
متن کاملمقایسه نتایج ارزیابی آزمون ارتقاء دستیاران در گروههای آموزش بالینی دانشگاه علوم پزشکی اراک
Introduction: Designing a tool for measuring of residents’competency with attention to their main role in education and practice of university. This study aims to assess the residents’ promotion tests of clinical departments at Arak University of Medicals Sciences. Methods: This cross- sectional study that was undertook in 2010 at Arak University of Medical Sciences. Seven hundred and fifty mul...
متن کاملسایکوآکوستیک و درک گفتار در افراد مبتلا به نوروپاتی شنوایی و افراد طبیعی
Background: The main result of hearing impairment is reduction of speech perception. Patient with auditory neuropathy can hear but they can not understand. Their difficulties have been traced to timing related deficits, revealing the importance of the neural encoding of timing cues for understanding speech. Objective: In the present study psychoacoustic perception (minimal noticeable differen...
متن کاملOutcome prediction of different groups of patients using a modified scoring system
Abstract Background: In this study we aimed to examine the discrimination and calibration of a severity characterization of trauma (ASCOT) in our setting to determine whether its usage is appropriate to predict outcome of our trauma patients. Methods: This study was conducted in three hospitals. All patients admitted in studied hospitals were divided randomly into two equal subgroup...
متن کامل