Premium
The Future of Medical Education Research: A Different Way to Analyze Data to Produce More Valid Results/Conclusions
Author(s) -
Lufler Rebecca,
Zumwalt Ann,
Hoagland Todd
Publication year - 2010
Publication title -
the faseb journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.709
H-Index - 277
eISSN - 1530-6860
pISSN - 0892-6638
DOI - 10.1096/fasebj.24.1_supplement.176.7
Subject(s) - confounding , analysis of variance , regression analysis , preference , statistics , variables , linear regression , psychology , statistical analysis , population , set (abstract data type) , statistical significance , mathematics , medicine , computer science , environmental health , programming language
Medical education research has been criticized for producing less than robust results. Perhaps in these studies significant differences exist, but they are being masked by the statistical analyses used. Typical analyses (t‐test, ANOVA) may not be appropriate in medical education research and may be disguising valid results or producing invalid results by not adjusting for confounding variables. Our purpose is to apply two different statistical analyses to the same set of medical education data and compare the outcomes. Using data from a previous study completed by the authors, analysis was completed in two ways: using ANOVA (assumes students are a homogeneous population) and using regression analyses (adjusting for variables that vary among students). ANOVA results showed no differences in practical exam averages across groups of students with different visual‐spatial abilities (VSA). Using regression analysis and adjusting for gender, learning preference, and MCAT scores, there was a significant difference between the performance of students with highest vs. lowest VSA, with adjusted practical exam average scores of 88.4 and 81.9, respectively (p=0.03). These data demonstrate that controlling for characteristic variables of the study population reveal confounding effects otherwise hidden by typical statistical analysis. Grant Funding Source : None