4
$\begingroup$

I have to run a linear regression analysis with an interaction effect of two categorical variables:

  • Modality (audio, visual and audio-visual)
  • Repetition (1x, 2x and 4x)

I have already dummified the variables. Before I can run the analyses I have to compute the interaction term. But I don't know how I have to do this in the case of a variables with 3 categories (Because I have two dummy variables for each independent variable).

$\endgroup$
1
$\begingroup$

As a general principal, interactions are generally obtained by taking the product of the dummies included in the model

For example, if you include "visual" and "audio-visual" as levels of the first factor, and "2x" and "4x" as levels of the second (leaving "audio" and "1x" as baseline), then you'd have the four interaction dummies

"visual"x"2x"
"visual"x"4x"
"audio-visual"x"2x"
"audio-visual"x"4x"

calculated as an ordinary product of the main effect dummies. But I bet SPSS has a way of doing it more directly than you actually calculating them by multiplication (it's so many decades since I used SPSS I don't recall).

$\endgroup$
  • $\begingroup$ SPSS has a function where you can enter the interaction term, so by doing this you can include it in your regression analysis. The two independent variables that I use have 3 categeries (so I have two dummy variables for each independent variable): Modality_Dummy1 Modality_Dummy2 Repetition_Dummy1 Repetition_Dummy2 I don't know how to enter the interaction term in this case. When the independent variables have two categories you can just enter modality * repetition. But do you think that I have to include more interaction terms in this case? $\endgroup$ – NadineV Jul 23 '15 at 9:46
  • $\begingroup$ If you can just put in the categorical variables, it should calculate all the individual terms for you. $\endgroup$ – Glen_b Jul 23 '15 at 9:54

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.