I am having troubles figuring out if my assumptions for the appropriate SPPS tests are correct.
I am doing a secondary analysis from an existing SPSS file.
I have two independent variables (with Yes/No Answers), and two dependent variables.
I aim to do multiple regression analysis and also investigate if two other variables moderate the effect of my independent variable.
First I will do reliability analysis, reporting Cronbach's alpha. I have to do this for my independent variables and my moderator variables only, right? (the ones that are part of the Questionnaire).
Initially I was thinking of doing SPSS -> Linear -> Multiple Regression. Where in Block (1) I will input independent variables and dependent variables, and in Block(2) input the two moderator variables. Is that how I should approach my research analysis?
Best Answer
Some of the details here are unclear to me. Because you say "I have one two independent variables" it isn't 100% clear to me how many IVs you have. You also mention doing a reliability check on the IV(s), but you said they are binary, so you wouldn't test the reliability of a single yes/no question.
Going back to the overall modeling question, there's no need for the blocks. Put the interaction terms (IV * moderator) in the model and if the estimate is significant, then it is significant. If you are adding the "moderators" without an interaction term, you aren't testing moderation.