Comparing the classic factor analysis with Bayesian factor analysis on the data of Family Dermatology Life Quality Index

Abstract

Background & Aim: Chronic diseases impact not only on patients but also on their family members’ lives. This study aims to determine dimensions of Family Dermatology Life Quality Index (FDLQI) questionnaire by the use of classic and Bayesian factor analysis (BFA) factor. Methods & Materials: In this study, FDLQI questionnaire distributed among 100 family members of dermatological patients. BFA is exploited to determining dimension and contribution of items of the questionnaire in different sample size. BFA is examined using the Monte Carlo-Markov chain algorithm. All the above analysis is done in sample size 100, 130, 150. Results: In this study, 100 family members of dermatological patient attended to evaluate Persian version of FDLQI questionnaire. A mean age of participants was 37.1 years (± 12.3). A mean score of FDLQI was 15.4 (± 5.5) with maximum and minimum scores of 30 and 6, respectively.Exploratory FA revealed a one-factor solution that accounted for 45.87% of the total variance. The unidimensional model was concordance by confirmatory FA. For more exploration, BFA was performed. Two factors extracted when iteration is done. Conclusion: It appears that when sample size diminished, Cronbach’s alpha and Kaiser–Meyer–Olkin increased. Among 10 items of the questionnaire, item 9 mostly appears differently in results.

Keywords

• Bayesian factor analysis
• Monte Carlo-Markov chain algorithm
• Family Dermatology Life Quality Index