Evaluating Factor Structure of Persian Version of Lifelong Learning Assessment Tool Using Ordinal Versus Quantitative Methods in Confirmatory Factor Analysis

Mahbube Abdollahi¹, Zeynab Avazzadeh¹, Maryam Salari², Farid Zaeri³, Fateme Pourhaji⁴, Elham Shaarbaf Eidgahi⁴, Zahra Khosravi Anbaran⁵, Fatemeh Zahra karimi⁶

¹Biostatistics Department, Tarbiat Modares University, Tehran, Iran. ²Biostatistics Department, Shahid Beheshti Univesity, Tehran, Iran. ³Health Promotion/Health Education Department, Tarbiat Modares university, Tehran, Iran. ⁴Department of Biostatistics and Epidemiology, Mashhad University of Medical Sciences, Mashhad, Iran. ⁵School of Nursing and Midwifery, Mashhad University of medical science, Mashhad, Iran. ⁶Department of Epidemiology and Biostatistics, School of Public Health, Zahedan University of Medical Sciences, Zahedan, Iran

DOI : http://dx.doi.org/10.13005/bbra/1937

ABSTRACT: In factor analysis, despite ordinal nature of the data, the assumption is that the data are normal quantitatively. The present study aimed to survey the factor structure of Persian Version of lifelong learning assessment tool JeffSPLL-MS using the most accurate method in confirmatory factor analysis. In this cross-sectional study, 430 students of Birjand University of Medical Sciences were selected randomly. Persian version of JeffSPLL-MS tool was used. First, assuming data scale being quantitative, three assessment methods were used, and by assuming data are ordinal, three assessment methods were used in confirmatory factor analysis, and eventually, the above mentioned six methods were compared. For comparing different methods, CFI, TLI and RMSEA indices were used. The three-factor structure of Persian version of JeffSPLL-MS had acceptable fit to the data. In all of the six methods, model indices confirmed the model fit, and ordinal models with correction had better fit compared to quantitative models. The result of the current study confirmed three-factor structure of Persian version of lifelong learning assessment tool. Although in confirmatory factor analysis, the common approach, taking into account the scale of quantitative data and using the maximum likelihood method, however, the results of this study showed that taking an ordinal scale data will result in improvement of fit model parameters.

KEYWORDS: Confirmatory factor analysis; ordinal scale; factor structure; JeffSPLL-MS

[ HTML Full Text]