Journal article
A Descriptivist Approach to Trait Conceptualization and Inference
Psychological review, Vol.123(1), pp.90-96
01/01/2016
DOI: 10.1037/a0039542
PMID: 26214685
Abstract
In their recent article, How Functionalist and Process Approaches to Behavior Can Explain Trait Covariation, Wood, Gardner, and Harms (2015) underscore the need for more process-based understandings of individual differences. At the same time, the article illustrates a common error in the use and interpretation of latent variable models: namely, the misuse of models to arbitrate issues of causation and the nature of latent variables. Here, we explain how latent variables can be understood simply as parsimonious summaries of data, and how statistical inference can be based on choosing those summaries that minimize information required to represent the data using the model. Although Wood, Gardner, and Harms acknowledge this perspective, they underestimate its significance, including its importance to modeling and the conceptualization of psychological measurement. We believe this perspective has important implications for understanding individual differences in a number of domains, including current debates surrounding the role of formative versus reflective latent variables.
Details
- Title: Subtitle
- A Descriptivist Approach to Trait Conceptualization and Inference
- Creators
- Katherine G. Jonas - University of IowaKristian E. Markon - University of Iowa
- Resource Type
- Journal article
- Publication Details
- Psychological review, Vol.123(1), pp.90-96
- Publisher
- Amer Psychological Assoc
- DOI
- 10.1037/a0039542
- PMID
- 26214685
- ISSN
- 0033-295X
- eISSN
- 1939-1471
- Number of pages
- 7
- Language
- English
- Date published
- 01/01/2016
- Academic Unit
- Psychological and Brain Sciences
- Record Identifier
- 9984627349702771
Metrics
16 Record Views