Objective:To develop a learning scale of scientific methods for medical students and to evaluate the reliability and validity of the scale.Methods:Based on the principles of planned behavior theory, the original scale was developed through literature analysis, pre-interview and experts’ consultation. A preliminary survey was conducted among 105 medical undergraduates with random cluster sampling, and the primary scale was formed by exploratory factor analysis. Then, 851 undergraduates were selected as the objects of formal survey, and the formal scale was finally determined by the reliability and validity evaluation and confirmatory factor analysis.Results:The formal scale was composed of 5 latent variables, 9 observed variables and 51 items. The overall Cronbach’s α coefficient and the overall retest reliability were 0.970 and 0.887, respectively. Composite reliability and KMO values of each latent variable were more than 0.8, and the values of average variance extraction were over 0.5. The fitness test ( P=0.155) showed acceptable fitting quality with the main fit indices of χ2/ df=1.405, RMR=0.039, RMSEA=0.022, etc. Conclusion:The learning scale of scientific methods shows satisfactory reliability, validity and fitness, which can be applied as a measuring tool to evaluate the learning behavior of scientific methods for medical students.

Objective:To develop a learning scale of scientific methods for medical students and to evaluate the reliability and validity of the scale.Methods:Based on the principles of planned behavior theory, the original scale was developed through literature analysis, pre-interview and experts' consultation. A preliminary survey was conducted among 105 medical undergraduates with random cluster sampling, and the primary scale was formed by exploratory factor analysis. Then, 851 undergraduates were selected as the objects of formal survey, and the formal scale was finally determined by the reliability and validity evaluation and confirmatory factor analysis. SPSS 22.0 and AMOS 22.0 softwares were used for statistical analysis of the measurement data.Results:The formal scale was composed of 5 latent variables, 9 observed variables and 51 items. The overall Cronbach’s α coefficient and the overall retest reliability were 0.97 and 0.89, respectively. Composite reliability and KMO values of each latent variable were more than 0.80, and the values of average variance extraction were over 0.50. The fitness test ( P=0.155) showed acceptable fitting quality with the main fit indices of χ2/ df=1.41, RMR=0.04, RMSEA=0.02, etc. Conclusion:The learning scale of scientific methods shows satisfactory reliability, validity and fitness, which can be applied as a measuring tool to evaluate the learning behavior of scientific methods for medical students.