Methodology and Statistics
Tilburg School of Social and Behavioral Sciences
Tilburg University

Phone: +31 (0) 13 466 3873
Academic webpage Andrea Stoevenbelt

Project

Psychometrics and statistics of stereotype threat

The gender gap in mathematics test performance is widely debated, and stereotype threat is often used to (partly) explain it. Stereotype threat theory states that women underperform on math tests after having been made aware of the negative stereotype stating that women are bad at mathematics. Numerous studies on the effect have been published, but concerns exist about the robustness of the effect as studies are often underpowered or make use of substandard or unregistered and hence potentially biased analysis techniques. The goals of this project are to shed light on the robustness of the effect, study implications of violations of popular analysis techniques used in stereotype threat research, and to study how performance decrements on math tests related to missingness. First, we will conduct a large-scale registered replication of a seminal stereotype threat study (Johns, Schmader, & Martens, 2005). Next, we will critically assess prevalence and implications of violations of assumptions underlying ANCOVA analyses that are often used in stereotype threat research. Third, we will build on earlier work that highlighted that slower working speed of threatened women might help explain the effect. We use missingness propensity models to study missing data patterns in the large-scale data obtained from the RRR.

Supervisors
Prof.dr. J.M. Wicherts, dr. P.C. Flore

Financed by
NWO, Tilburg University

Project period
1 September 2017 – 31 August 2121