A Hierarchical Unequal-Variance Signal Detection Model for Binary Data (original) (raw)

Gaussian signal detection models with equal variance are typically used in yes-no detection and discrimination whereas models with unequal variance require paradigms with multiple response categories or conditions. Here, a hierarchical signal detection model with unequal variance is proposed that is based on binary responses from a sample of participants. Introducing plausible constraints on sampling distributions makes it possible to estimate sensitivity, decision criterion and signal variance at the population level. The model is explored in simulation studies and applied to existing data from memory and reasoning tasks. The results suggest that this hierarchical unequal-variance model provides a promising alternative to equal-variance models.