Validation of the Universal Measure of Bias–Fat (UMB Fat) among Malaysian health practitioners using Rasch analysis

Nor Akma Yunus*, Elizabeth Sturgiss, Sze Ee Soh

*Corresponding author for this work

Research output: Contribution to journalArticleResearchpeer-review

Abstract

Objective: This study aimed to evaluate the structural validity of the Universal Measures of Bias – Fat (UMB Fat) among Malaysian healthcare practitioners using Rasch analysis. 

Methods: Data from a cross-sectional survey of 268 public and private doctors and allied health practitioners in Peninsular Malaysia were used for this analysis. Using Rasch analysis, overall model fit and item fit of the summary UMB Fat and domain scores were examined, together with unidimensionality, response threshold ordering, internal consistency, measurement invariance, and item targeting. 

Results: Data showed overall misfit to the Rasch model for both the summary UMB Fat score and domain scores. Whilst unidimensionality was observed for the domain scores, this was not evident for the summary score where multiple local dependencies were present. Disordered thresholds were observed for the response format, in which the majority improved with modification. Suboptimal targeting was also detected with an uneven distribution of items at the upper and lower end of the logit scale for the summary and domain scores. Despite this, excellent internal consistency reliability was observed (person separation index: 0.76–0.89), and no measurement invariance was detected. 

Conclusion: The Rasch model supports reporting of the UMB Fat domain scores but not the summary score. Several issues related to local dependencies and response format were identified that could benefit from refining the UMB Fat to improve measurement accuracy, particularly when used by healthcare practitioners in Asian countries.

Original languageEnglish
Pages (from-to)477-484
Number of pages8
JournalObesity Research and Clinical Practice
Volume17
Issue number6
DOIs
Publication statusPublished - 1 Nov 2023
Externally publishedYes

Cite this