PSYC 202 Chapter Notes - Chapter 11: Multiple Comparisons Problem, Ratio Test, John Tukey
4 Dec 2017
School
Department
Course
Professor
Document Summary
How to test for which groups have diffs in a single factor analysis of variance. Categorical specifies the groups & if we have more than 2, cant use pairs of t -tests, instead test for overall diff among the groups using an f test. Happens only after you"ve looked at a anova table. Only if the analysis of variance is significant to you go this next level to figure out where diff is happening. Purpose to test for diff bw 2 groups while adjusting for the overall type i error rate in entire tail (entire set of data, called the family wise error rate. Ho= means of 4 groups are the same. Ha=means of groups are different somewhere so at least 2 means are diff. Reject null bc p value is less than 0. 05. Calculate effect size using summary statement from the linear model. If the intercept or 3 of the groups relative to the intercept.
