Reading focus
Syllabus: How Do We Combine Evidence Across Studies?
Follow the search, coding and callback measure before examining the time-trend analyses.
Why you’re reading it. Connect eligibility and comparability to a substantive claim from a literature.
A question to consider. How could changes in sampled jobs or audit methods affect a time trend even if each individual study provides a credible comparison?
Why this reading
This synthesis measures unequal treatment through employers' responses to matched applicants. It offers a way to study discrimination that does not depend on people reporting their attitudes, while showing the additional assumptions needed to compare studies across years.
Before you read
Correspondence studies vary signals on applications; in-person audits use matched testers. A callback ratio compares response rates, so a ratio above one is a relative difference, not a percentage-point gap. Meta-regression examines differences between studies.
Read for
- Check what is held comparable within each kind of audit and what the racial signal may communicate.
- Separate the evidence for a callback difference within studies from the evidence for a trend across study years.
- Compare the estimates and uncertainty for African American and Latino applicants; the title compresses different results.
Consider afterwards
- An employer's attitude survey improves while an application audit still finds unequal callbacks. What do the two measures establish, and what would you investigate next?
Editorial context
The main time comparison covers U.S. fieldwork from 1989–2015; it is not a statement about the present or every country. Callback behavior is an entry-stage outcome, not the whole employment relationship. Audits identify unequal treatment under specified application conditions without directly revealing a recruiter's motive. Study settings, available comparisons, and between-study uncertainty constrain the time-trend claim.