Statistics in Educational Research Rebecca Henry, Ph.D. OMERAD College of Human Medicine Michigan State University
Working With Statisticians What will s/he want from you? The Research Question/Purpose ---Uhhh! Keep it Simple Description of Design 2 Group Case Control Medical School Dropouts Medical School Graduates Cross Sectional Survey of Preceptors Description of Variables IV & DV Level of Measurement Nominal/Dichotomous Ordinal Interval
Components of the Research Question/Hypothesis Variables Independent (predictor) The variable that is doing the causing or explaining. Examples: Gender, educational interventions, SES can all be powerful explainers or influences of outcome.
Dependent (criterion or outcome) A variable that is expected to be affected by an independent variable. Examples: Test scores, practice choice, satisfaction and performance abilities are often influenced by other variables. Caution - some studies don’t distinguish this way. For example, a relationship study that examined scores on NLME Part II with a department developed exam
Levels of Data Nominal: Categorical Male - Female Better - Worse Learning Styles Ordinal: Order Class Rank Tumor Classification Hi - Med - Low Interval: Equal Distance Between Point Temperature IQ Scores Standardized Measures Lickert Scales Cost Ratio: Meaningful Zero Height Weight
Bivariate Analysis INDEPENDENT Nominal (2 values) Nominal (> 2 values) Interval DEPENDENTNOMINAL INTERVAL(2 VALUES)(> 2 VALUES) t/z TestChi-square Chi-square (Mann Whitney U) Analysis of Chi-square Chi-square Variance (ANOVA) Correlation/LogisticDiscriminant regressionregressionanalysis