Different predictors of memory scanning with unidimensional and digit stimuli

Abstract

The processing of digit and unidimensional brightness stimuli was examined in the traditional Sternberg fixed-set paradigm. From a total set size of 10, 18 different memory sets were studied, 6 each at memory set sizes of 1, 2, and 3. Three different variables, which are in general correlated, were used to predict performance: memory set size (M), a measure based on inferred distance relationships among the stimuli constituting the positive and negative sets (d), and a measure based on the number of inferred “decision boundaries” necessary to separate all memory set items from the items remaining in the set (DB). For the digit stimuli the results were consistent with previous research, and memory set size was the prepotent predictor. For the unidimensional brightnesses, DB was the prepotent predictor, and choice reaction times (CRTs) to both positive and negative probes were a function of the distance of the probe item from the closest member of the complementary set. However, for the digit stimuli this was found only for negative probes; CRTs to positive digit probes were independent of the distance from the closest negative set member.