We describe two SPSS macros that implement a matrix algebra method for comparing any two fitted values from a regression model. ![]() For many fitted value comparisons that are not captured by any of the regression coefficients, common statistical software packages do not provide the standard errors needed to compute confidence intervals or carry out statistical tests-particularly in more complex models that include interactions, polynomial terms, or regression splines. But the coefficients represent only a fraction of the possible fitted value comparisons that might be of interest to researchers. ![]() Therefore, each regression coefficient represents the difference between two fitted values of Y. ![]() ![]() In regression models with first-order terms only, the coefficient for a given variable is typically interpreted as the change in the fitted value of Y for a one-unit increase in that variable, with all other variables held constant.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |