Publications on Gender Disparities

This two-guide set for evaluators and others interested in evaluation grew out of a National Science Foundation funded effort to improve cross project evaluations. Guide 1, Designing a Cross-Project Evaluation, focuses on evaluation design including identification and operationalization of program goals, building of logic models, and selection of indicators and appropriate measures for these indicators. Guide 2, Collecting and Using Data in Cross-Project Evaluation, lays out multiple issues involved in data collection, strengths and weaknesses of different data collection formats, and methods for ensuring data quality, confidentiality, and the protection of human subjects.

Although many researchers have documented lower levels of upward mobility amongst black families, it is difficult to disentangle the effects of discrimination from differences in (sometimes unobservable) characteristics that also contribute to variation in employment, income, health, housing, and wealth outcomes across groups. As a consequence, findings regarding the presence or absence of discrimination tend to be controversial. This review pulls together several strands of research on the subject, including the statistical analysis of survey data, audit studies comparing market outcomes for similarly qualified individuals who differ along racial lines, and public opinion polling data on discrimination. (Review 1 of 11.)

Social Security spouse and survivor benefits advantage single-earner families relative to dual-earner families paying the same total taxes. Our paper considers earnings sharing—through which husbands' and wives' earnings records are combined and averaged throughout their marriage when computing benefits—as well as other changes to spouse/survivor benefits, including caregiver credits and minimum benefits. All the roughly cost-equivalent packages examined improve adequacy and horizontal equity compared to current law. The earnings-sharing proposal, however, only reduced poverty with significant adjustments to the treatment of surviving spouses. The packages reveal tradeoffs among beneficiary groups, with particular tensions around work and marital status.

Gender differences in academic achievement have long fascinated researchers and policy-makers alike. In this paper we analyze differences in math and reading test score growth rates by gender for four different race and ethnic groups -- white, black, Hispanic, and Asian students -- for six different time periods. Our data cover both the earliest years of education and the crucial years of adolescence. In addition, we have data bracketing one non-schooling period. Together these data enable us to get a very complete picture of how gender gaps evolve over the course of early elementary and high school years and how these trajectories differ by race and ethnicity. While the gender gaps are not always statistically significant, they are for 15 of 48 comparisons made, all during school. In addition, all of the statistically significant results suggest that males learn more math and females more reading during early elementary school and again during high school.

Although Social Security does not currently guarantee low-wage workers a minimum retirement benefit, several proposed reforms include benefit minimums. Our analysis suggests that such policies could help reduce poverty among older adults. However, the way the minimums are designed could affect the extent to which poverty rates decline. Using the Urban Institute's dynamic microsimulation model, we examine five alternative benefit options. Several lessons emerge from these simulations that could help guide designers of Social Security proposals. Understanding how a well-designed minimum benefit could reduce poverty is especially important now, when the program's long-term fiscal deficit threatens future benefit reductions.