Suppose a researcher runs a “forward entry” statistical multiple regression with a group of 20 candidate predictor variables; the final model includes five predictors. Why are the p values shown on the SPSS printout not a good indication of the true risk of Type I error in this situation? What correction can be used when you test the significance of the overall multiple R for a forward regression?
Answer to relevant QuestionsHow can we obtain significance tests for sets or blocks of predictor variables? What is centering? How do you center scores on a quantitative predictor variable in SPSS? When adjusted group means are compared with unadjusted group means, which of the following can occur? (Answer yes or no to each.) • The adjusted group means can be closer together than the unadjusted group means. • The ...In general, if p is the number of predictor variables and k is the number of groups for the outcome variable in a DA, how many different discriminant functions can be obtained to differentiate among groups? (This question ...Suppose that your first discriminant function has an eigenvalue of 1 = .5. What is the value of rc, the canonical correlation, for this function? What is the value of Roy’s largest root for this function?
Post your question