Differential weighting of common and distinctive components

Area of Science:

• Cognitive psychology
• Human perception
• Information processing

Background:

• Previous research indicates that common components hold greater weight in similarity judgments for verbal stimuli compared to pictorial stimuli.
• Understanding the factors influencing feature weighting in different representational formats is crucial for cognitive models of perception and memory.

Purpose of the Study:

• To investigate the underlying reasons for the higher relative weight of common components in similarity judgments for verbal versus pictorial stimuli.
• To examine how stimulus presentation and manipulation of component cohesiveness affect feature weighting in verbal and pictorial domains.

Main Methods:

• Experiments involved presenting stimulus components serially and manipulating the cohesiveness of composite stimuli (pictures and verbal descriptions).
• The relative weight (W) of common components in similarity judgments was measured under different experimental conditions.
• Stimuli included both verbal and pictorial representations of the same content to allow for direct comparison.

Main Results:

• Serial presentation of components did not alter W for verbal stimuli but increased the impact of both common and distinctive features for pictorial stimuli without changing their relative weight.
• Reducing picture cohesiveness (separating, scrambling, mixing components) increased W for pictorial stimuli.
• Enhancing verbal stimulus cohesiveness (imposing structure) decreased W for verbal stimuli.
• No systematic differences in W were observed when comparing verbal and pictorial representations of identical stimuli.

Conclusions:

• The cohesiveness of stimulus representation is a key factor influencing the relative weight of common components in similarity judgments.
• Pictorial stimuli's higher susceptibility to cohesiveness manipulations, compared to verbal stimuli, may explain the initial finding.
• Cognitive models should account for representational format and stimulus organization when predicting feature weighting in similarity judgments.