# Interaction in multiple regression

I am a psychology undergrad student and I'm not the best at stats. My thesis is due soon but I'm having trouble with my analysis despite extensively searching for an answer. It would be helpful to me if any answers are given in simple terms!

My dependent variable is victim blame and is continuous scale data. My independent variables are gender (male vs female) and scenario (scenarioA vs scenarioB vs scenarioC) as well as hostile sexism (continuous scale), benevolent sexism (continuous scale) and frequency of pornography use.

It has been suggested that I carry out a hierarchical regression but I am having a great deal of trouble with this. My hypothesis is that victim blame scores will differ depending on what scenario was read and that this difference will be moderated by benevolent sexism scores. Does this involve creating an interaction term of scenario x benevolent sexism? Is it necessary to first dummy code the scenario variable? Or is regression not the right approach at all? I'd reeeeeaaaally appreciate some help with this! Many thanks.

2. You don't say what program you are using, but all the statistics packages will do this for you. I know that SAS, R and SPSS will, and doubtless so will others. However, you may have to alter the reference level or the way the variable is parameterized. This will vary depending on software (they have different defaults).
• SPSS will create the dummy variable for you. If you add the interaction it will also create appropriate codes. However, I am not an SPSS expert, so I can't provide details. Perhaps someone else here can. Feb 10 '14 at 15:47