The extralist-feature effect: evidence against item matching in short-term recognition memory

Area of Science:

• Cognitive Psychology
• Memory Research
• Human Recognition Memory

Background:

• Current models of recognition memory often confound feature similarity with feature overlap.
• Understanding the precise factors influencing short-term memory performance is crucial for cognitive modeling.

Purpose of the Study:

• To investigate the role of feature overlap versus feature similarity in short-term item recognition.
• To test predictions of existing memory models against empirical data.

Main Methods:

• A short-term item-recognition task was employed, manipulating the similarity and number of features between negative probes and study items.
• Stimuli included both abstract stimuli (colored shapes) and words to ensure generalizability.
• Unconfounded comparisons were made to isolate the effect of feature overlap.

Main Results:

• Recognition performance was solely determined by the number of probe features present in the study set.
• Response time showed a linear relationship with the number of probe features in the study list.
• This effect was consistent across both manipulated features and words.

Conclusions:

• Existing global memory models, which average item matches, do not fully explain the observed data.
• A proposed alternative mechanism suggests direct comparison of probe features against a global representation of the study set.
• The number of presented features, a global list property, is a key determinant of recognition accuracy.