Psychometric Properties of the Learning Approach Inventory: A Confirmatory Factor Analysis

Objective: To determine the internal consistency and construct validity of the Learning Approach Inventory (LA-i) among first year medical students in a Malaysian medical school. Methods: Cross sectional study was done on 196 first year medical students in Universiti Sains Malaysia (USM). The items of the LA-i were framed based on characteristics of three learning approaches. The Cronbach’s alpha reliability analysis, exploratory factor analysis and confirmatory factor analysis were applied to measure internal consistency and construct validity. These analyses were done using Predictive Analytics SoftWare (PASW) version 18 and Analysis of Moment Structure (AMOS) version 19. The Composite Reliability (CR) and Average Variance Extracted (AVE) were calculated manually to measure construct reliability and convergent validity. Result: A total of 196 medical students responded to this study. Exploratory factor analysis showed that three potential constructs were extracted from the inventory. The confirmatory factor analysis showed the three factor model with nine items had a good fit with the latent constructs (X2 (df) = 26.07 (20), p = 0.163, RMR = 0.04, GFI = 0.969, AGFI = 0.93, NFI = 0.967, RFI = 0.941, IFI = 0.992, TLI = 0.985, CFI = 0.992, RMSEA = 0.04). Each domain of the final model of the LA-i has three items. The Cronbach’s alpha value of the AL-i was 0.86. The Cronbach’s alpha values of surface, strategic and deep approach domains were 0.62, 0.73 and 0.88 respectively. Most of learning approach domains had Composite Reliability and Average Variance Extracted values were more than 0.6 and 0.5 respectively indicating good construct reliability and adequate convergent validity. Conclusion: This study suggested that the three factor model with 9-items of the LA-i has a good fit and shown good psychometric values. It is a valid and reliability measurement to determine learning approaches among first year medical students.