PSYC 305 Lecture Notes - Lecture 20: Linear Regression, Analysis Of Variance, Royal Institute Of Technology

• School: McGill University
• Department: Psychology
• Course: PSYC 305
• Professor: Heungsun Hwang

Document Summary

We have so far learned the following parametric tests t/z test. Yes, they are all connected to each other t-test is a special case of one-way anova with only two groups to compare. Linear regression has anova tests as special cases. Anova may be viewed as a special case of multiple regression with nominal (categorical) ivs with multiple levels/groups. We need to transform nominal variables into dummy-coded variables to use in multiple linear regression. When one nominal iv is used in the regression model, this is equivalent to one-way anova. When two nominal ivs are used, this is equivalent to two-way anova. And so on because there is such a thing as a three- or four-way anova test. Dummy coding involves the assignment of binary values (0 or 1) to represent membership in each level of a nominal variable. 0 means not a member, and 1 means member. Expresses group memberships of observations using only zeros and ones.