Exploration and confirmation of the latent variable structure of the Jefferson scale of empathy
Mohammadreza Hojat1 and Marianna LaNoue2
1Jefferson Medical College of Thomas Jefferson University, Center for Research in Medical Education and Health Care, and Department of Psychiatry and Human Behavior, USA
2Jefferson Medical College of Thomas Jefferson University, Department of Family and Community Medicine, USA
Submitted: 23/01/2014; Accepted: 04/04/2014; Published: 20/04/2014
Int J Med Educ. 2014; 5:73-81; doi: 10.5116/ijme.533f.0c41
© 2014 Mohammadreza Hojat & Marianna LaNoue. This is an Open Access article distributed under the terms of the Creative Commons Attribution License which permits unrestricted use of work provided the original work is properly cited. http://creativecommons.org/licenses/by/3.0
Objectives: To reaffirm the underlying components of the JSE by using exploratory factor analysis (EFA), and to confirm its latent variable structure by using confirmatory factor analysis (CFA).
Methods: Research participants included 2,612 medical students who entered Jefferson Medical College between 2002 and 2012. This sample was divided into two groups: Matriculants between 2002 and 2007 (n=1,380) and be-tween 2008 and 2012 (n=1,232). Data for 2002-2007 ma-triculants were subjected to EFA (principal component factor extraction), and data for matriculants of 2008-2012 were used for CFA (structural equation modeling, and root mean square error for approximation).
Results: The EFA resulted in three factors: "perspective-taking," "compassionate care" and "walking in patient's shoes" replicating the 3-factor model reported in most of the previous studies. The CFA showed that the 3-factor model was an acceptable fit, thus confirming the latent variable structure emerged in the EFA. Corrected item-total score correlations for the total sample were all positive and statistically significant, ranging from 0.13 to 0.61 with a median of 0.44 (p<0.01). The item discrimination effect size indices (contrasting item mean scores for the top-third versus bottom-third JSE scorers) ranged from 0.50 to 1.4 indicating that the differences in item mean scores between top and bottom scorers on the JSE were of practical im-portance. Cronbach's alpha coefficient of the JSE for the total sample was 0.80, ranging from 0.75 to 0.84 for ma-triculatnts of different years.
Conclusions: Findings provided further support for under-lying constructs of the JSE, adding to its credibility.