Psychometric properties of DREEM in a sample of Malaysian medical students

Dear Sir

Over the past 15 years, medical and allied health educators across the globe widely used the Dundee Ready Educational Environment Measure (DREEM) to appraise their institutions’ educational climate (Roff & McAleer 2001). Its validity and reliability was well reported across institutions, but none of the article reported about it in Malaysia settings. This study aimed to evaluate its psychometric properties in a sample of Malaysian medical students. It was designed to answer a few questions: (1) Is DREEM a valid and reliable tool to measure educational climate in the studied population?, (2) What is the best fit model of DREEM in the studied population? (3) Does internal consistency vary across years of study?

The author carried out a cross-sectional study on 656 medical students. A total of 511 (77.9%) medical students completely responded to the DREEM: 156 (30.5%) first year, 196 (38.4%) third year and 159 (31.1%) fifth year. Majority of them were female (61.1%). Confirmatory factor analysis (CFA) and reliability analysis were analysed by SPSS version 18 and AMOS 19 to assess the construct validity and internal consistency.

The CFA showed that the proposed five-factor structure of the original DREEM was not fit as all the goodness of fit indices did not signify for model fit. This finding consistent with previous studies reported that its construct validity was not well supported by the empirical data (Jakobsson et al. 2011). On further CFA, the author found that the five-factor structure of DREEM with 17 items (i.e. the shortened DREEM) was fit as all the indices signify for model fit. Cronbach's alpha values across years of study for the subscales of both DREEM versions ranged between 0.53 and 0.82.

Our study did not support the proposed five-factor structure of DREEM. The shortened version demonstrated a better fit with the proposed model and was found to be as reliable as the original version across years of study. Continued research is required to verify and maximize psychometric credentials of the DREEM across institutions and nationalities.