.png){kind=icon}
.png){kind=icon}
Return
Evaluating Avoidable Heterogeneity in Exploratory Factor Analysis Results
.png){kind=icon}
PRE
AI
DOI：10.1037/met0000589.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Meaningful interpretations of scores derived from psychological scales depend on the replicability of psychometric properties. Despite this, and unexpected inconsistencies in psychometric results across studies, psychometrics has often been overlooked in the replication literature. In this article, we begin to address replication issues in exploratory factor analysis (EFA). We use a Monte Carlo simulation to investigate methodological choices made throughout the EFA process that have the potential to add heterogeneity to results. Our findings show that critical decision points for EFA include the method for determining the number of factors as well as rotation. The results also demonstrate the relevancy of data characteristics, as some contexts are more susceptible to the effects of methodological choice on the heterogeneity of results. Translational Abstract When using psychological scales, such as a survey or questionnaire that is intended to measure depression, for example, meaningful interpretations of (depression) scores depend on the replicability of results from psychometric studies. Despite this, and despite unexpected inconsistencies in psychometric results across studies in the empirical literature, psychometrics has often been overlooked in the replication discussion. In this article, we begin to address replication issues in exploratory factor analysis (EFA)-a popular method for scale development and evaluation. We use a Monte Carlo simulation to investigate whether choices made throughout the data analysis process are potentially detrimental to the replicability of EFA results. Our find-ings show that critical decision points for EFA include the method for determining the number of factors as well as rotation. The results also show that the characteristics of the data are an important consideration as some contexts have a greater impact on efforts to replicate an EFA.
Keywords:
exploratory factor analysis
replication
methodological flexibility
researcher degrees of freedom
questionable research practices