EFFECTS OF RATING SCALE AND FORCED-CHOICE FORMAT MEASUREMENT METHODS ON THE QUALITY OF RELIABILITY AND CONSTRUCT VALIDITY IN BIG FIVE PERSONALITY TRAITS MEASUREMENT
Abstract:
The purposes of this research were 1) to compare the reliability of the big five personality traits measurement on the rating scale and forced-choice format, 2) to validate the construct validity of the big five personality traits measurement on the rating scale and forced-choice format, 3) to analyze the effects of measurement methods (the rating scale and forced-choice format) of the big five personality traits measured by controlling for the effects of an unmeasured latent methods factor technique (CEUL) and controlling for the effects with correlated trait factor and correlated method factor technique (CTCM) and 4) to analyze the effects of measurement methods (positive and negative items) of the big five personality traits measured by using CEUL and CTCM techniques. The sample consisted of 1,200 students in grade 12 from the school under the basic education commission, Bangkok. The research instruments consisted of the big five personality traits which measured by the rating scale and forced-choice format. Data analysis were conducted by using z-test and confirmatory factor analysis. The results of this research were as follow: 1) the internal consistency reliability by using Cronbachs alpha coefficient of the big five personality traits measurement on the rating scale higher than the reliability of the forced-choice format at the .01 level of significance, the stability reliability by using the Pearson product moment correlation coefficient of the big five personality traits measurement on the forced-choice format higher than the reliability of the rating scale at the .01 level of significance, 2) the results of confirmatory factor analysis of the big five personality traits measured by the rating scale model (CFA-R) and forced-choice format model (CFA-F) were not fit to the empirical data, however, after the models were adjusted, the CFA-F model were fit to the empirical data than the CFA-R model at the .01 level of significance (CFA-F; Chi-square = 2,442.08, df = 1,650, Chi-square/df = 1.48, RMSEA = .020, GFI = .94), 3) the analysis results showed that the effects of the rating scale and forced-choice format measurement methods of the big five personality traits measured by using CEUL technique were fit to the empirical data and appropriated than by using CTCM technique (R-MR-CEUL; Chi-square = 11,207.09, df = 1,640, Chi-square/df = 6.83, RMSEA = .070, GFI = .76, low level effects of measurement methods), 4) the analysis results showed that the effects of the positive and negative items measurement methods of the big five personality traits measured by using CEUL technique were fit to the empirical data and appropriated than by using CTCM technique (R-MRN-CEUL; Chi-square = 12,069.24, df = 1,676, Chi-square/df = 7.20, RMSEA = .072, GFI = .75, medium level effects of measurement method) and 5) The results revealed that the most effective technique to analyze the effects of measurement methods of the big five personality traits measurement on the rating scale and forced-choice format were used CEUL technique.
Chulalongkorn University. Office of Academic Resources