Regression

Regression

1. Simple Regression

2. Multiple Regression

   • Categorical predictions
   • Standard MR
     • Test relationship of all X variables in predicting Y
   • Statistical MR
     • Test which X variable subsets are lest redundant (most unique contributing) in predicting Y
   • Hierarchical MR
     • Test which X variables best predict Y by: 1. place variables into subsets by step order (either logically placed i.e. which one comes first or which one lest manipulable i.e. pre-existing factors) 2. use statistical regression to determine which X variable subsets most contributing (i.e. unique contributions)

3. Correlation

   • Relationship between variables
   • Effect size: r/d (Cohen's rule)