Attributions to discrimination are subjective judgments that involve attributing the cause for the unfair treatment of a person to that person’s social group membership. Here, we use signal detection theory (SDT) as a framework to differentiate between people’s sensitivity (i.e., their ability to differentiate between situations with and without cues to discrimination) and response bias (i.e., their relative tendency for attributing to discrimination) in such judgments. To enable the empirical application of SDT to attributions to discrimination, we introduce a new paradigm that is set in the context of pay raise decisions made by a committee in a fictional company. In the paradigm, participants are presented with a number of such decisions involving employees of different social groups (as indicated by a portrait picture and a name) and are asked to judge for every decision whether they suspect it to be biased. We present the results of an experiment using this paradigm in the context of gender discrimination where we examined the impact of the discrimination context on participants’ response bias and sensitivity. Our results suggest that people’s response bias in attributions to discrimination is influenced by decision outcomes and by their expectations about gender bias, whereas gender differences in sensitivity in our paradigm appeared independent of the discrimination context.