Factor analysis in medical education research

Abstract

There is a large amount of survey research in medical education. Data is collected using newly created questionnaires with the desired content and the results are measured on a Likert scale. For example, The Dundee Ready Educational Environment Measure (DREEM) has 50 questions. Analyzing the results by displaying raw scores for each item would be excessive. Therefore, grouping content with a large number of questions from the beginning of the questionnaire creation with an expert panel to consider not only content validity but also construct validity. Interpreting the results by group scores makes it easier to understand. Grouping can also be done after the survey using statistical analysis to assist with construct validity. Popular techniques include factor analysis.