A model of positive sequential dependencies in judgments of frequency

Positive sequential dependencies occur when the response on the current trial n is positively correlated with the response on trial n−1. They are observed in a Judgment of Frequency (JOF) recognition memory task (Malmberg & Annis, 2012), and we developed a process model of them in the REM framework (Malmberg, Holden et al., 2004, Shiffrin and Steyvers, 1997) by assuming that features that represent the current test item in a retrieval cue carry over from the previous retrieval cue. We tested the model with data that distinguish between the number of times two given items were studied (frequency similarity) and the similarity between stimuli (item similarity), which was varied by presenting either landscape photos (high similarity), or photos of everyday objects such as shoes, cars, etc. (low similarity). Two models of item similarity were tested by assuming that the item representations share a proportion of features and that the exemplars from different stimulus classes vary in the distinctiveness or diagnosticity. A comprehensive exploration of several variants of these models directly was conducted comparing BIC and SBICR model selection statistics. The analyses establish the plausibility of the basic model of positive sequential dependencies, which assumes that differences in the similarity of the stimuli and differences in vigilance to the JOF task account for the pattern of sequential dependencies that we observed. They also indicate that different decision criteria are used to classify different stimuli on the JOF scale.