Deciding whether to use correlation, t-test, ANOVA, or chi-square

Last updated 3 years ago First published 15 April, 2017

Deciding which simple test to use for statistical testing for two variables

Research English Rating: ALL ages Shortest path: 3 nodes Longest path: 7 nodes Possible solutions: 4
Douglas Ferguson I work at the College of Charleston
  • Is your dependent variable a category? No. Yes.
  • Use chi-square.
  • Is your independent variable a category? Yes. No.
  • Use correlation (Pearson r).
  • How many levels? More than two (e.g., religions). Two (e.g., gender).
  • Use group mean difference (t-test).
  • Use group mean difference (ANOVA).

Related Topics

Related Trees

More from this iBosker