Memory effects of semantic attributes: A method of controlling attribute contamination

Area of Science:

• Cognitive Psychology
• Psycholinguistics
• Semantics

Background:

• Semantic attribute ratings are crucial for memory research but often suffer from attribute intercorrelations.
• Existing rating norms lack a clear understanding of the underlying distinct semantic dimensions.
• Contaminated manipulations hinder accurate investigation of memory encoding processes.

Purpose of the Study:

• To identify the fundamental latent semantic attributes underlying word intensity and ambiguity ratings.
• To develop a method for controlling attribute contamination in memory research.
• To provide a normed database for future uncontaminated memory studies.

Main Methods:

• Compiled a database of word ratings across 20 semantic attributes from multiple projects.
• Utilized principal component analysis (PCA) to identify latent factors for intensity and ambiguity.
• Interpreted the psychological meaning of the extracted latent factors.

Main Results:

• PCA revealed three latent factors for semantic intensity: emotional valence, size, and age.
• PCA identified five latent factors for semantic ambiguity: discrete emotion, emotional valence, age, meaningfulness, and verbatim memory.
• A database of factor scores for intensity and ambiguity was created.

Conclusions:

• Semantic intensity and ambiguity can be explained by a smaller set of underlying latent factors.
• The identified factors provide a basis for designing memory experiments free from attribute contamination.
• The provided database facilitates more precise research into the memory effects of semantic attributes.