What is it about?
The research work reported in the paper served two purposes: (1) the investigation of the structural validity of the Online Self-Assessment (OSA) Figures scale and (2) the evaluation of the threshold-free approach of confirmatory factor analysis (CFA) for binary data in an empirical example. Over nine thousand participants did take part in the online self-assessment that required completing figural reasoning items. Unidimensionality was tested by two methods: (1) CFA with tetrachoric correlations as input and robust estimation and (2) CFA with probability-based covariances as input and an additional link transformation. Both methods provided support for structural validity. These results suggested that the link transformation of the threshold-free approach compensated for the deviation from normality that necessitated robust estimation otherwise.
Featured Image
Photo by Matt Reames on Unsplash
Why is it important?
The problem that CFA is not appropriate for binary data is addressed. Two approaches to overcome this problem are discussed and applied.
Perspectives
Writing this article was a great pleasure as it has co-authors with whom I have had good cooperation over the years. For me it was especially interesting to see how the two approaches of analyzing binary data would perform in a condition that could be assumed to be favorable for both approaches.
Karl Schweizer
Read the Original
This page is a summary of: Structural Validity of the OSA Figures Scale for the Online Self-Assessment of Fluid Reasoning, European Journal of Psychological Assessment, September 2018, Hogrefe Publishing Group,
DOI: 10.1027/1015-5759/a000345.
You can read the full text:
Contributors
The following have contributed to this page
