The authors developed the Group Activated Probability of Success (GAPS) model to replicate the relationship between the performances of minority group members that are impacted by stereotype threat and to examine the impact of stereotype threat from an item level perspective. To accomplish this, a traditional item response theory (IRT) model was modified to reflect the specific characteristics of the individual and the items that interact to influence the individual's estimated latent proficiency in the target domain. This resulted in non-uniform differential item functioning (DIF) in items of high difficulty for minority group members who are also of high proficiency in the target domain. The model was developed to simulate the effect of stereotype threat at the item level using item difficulty, individual latent proficiency, and group membership to alter the probability of success on a given item for individuals in the targeted group, controlling for proficiency in the target domain. Guided by stereotype threat research, the manipulation of these factors resulted in a model that successfully replicates the differences found in laboratory experiments and suggests a possible explanation for the lack of support for stereotype threat in applied research.