Multivariate analyses depend on multivariate normality assumption.

Introduction:USMaP-i is an English, 66-item self-administered inventory, consisting of personality (60 items, 5 factors) and faking (one factor) components, which was mainly developed to measure personality traits among Malaysian students based on local cultures and values. The personality component was based on the Big Five dimensions as suggested by numerous personality researchers. Previous exploratory studies showed promising validity, reliability and stability of USMaP-i. Objective: To provide further validity evidence of USMaP-i for use among medical degree program applicants by confirmatory factor analysis (CFA). Methods: Data were collected as a part of screening of medical degree program applicants for year 2010-2013 intakes in Universiti Sains Malaysia (USM), of which 657 cases were suitable for analyses following a data screening measures. CFA was performed by bootstrap maximum likelihood estimation due to non-normality of items at multivariate level. Results: Although the revised five-factor model of personality showed good model fit (X2(df) = 144.36(55), P-value < 0.001; CFI = .944, TLI = .921; RMSEA = .050; SRMR = .032, Bollen-Stine bootstrap P-value = 0.004), the reliability of the factors is very poor (composite reliabilities (CR) = .483 to .650). In contrast, the unidimensional faking component exhibited good model fit (X2(df) = 14.15(5), P-value = 0.015; CFI = .984, TLI = .968; RMSEA = .053; SRMR = .011, Bollen-Stine bootstrap P-value = 0.068) and factor reliability (CR = 0.731). Conclusion: The personality component should be revised and revalidated due to poor reliability, despite showing good model fit. In contrast, the faking component showed good model fit and reliability. Further validation studies are recommended before its use among medical degree program applicants.

Introduction: For pre-post and cross-over design analysis of numerical data, paired t-test is the simplest analysis to perform. In planning such studies, it is imperative to calculate appropriate sample size required for the test to detect hypothesized difference. However, the sample size formula requires determination of standard deviation of difference, which is not commonly reported. In this article, the author guides the reader to calculation of standard deviation of difference from standard deviation of each separate occasion.