Cross Validated is a question and answer site for people interested in statistics, machine learning, data analysis, data mining, and data visualization. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I'm trying to run ANCOVA in SPSS, with a continuous IV as well as a continuous covariate.

Variable Information:

  • 1 DV: Continuous
  • 2 IVs: 1 Continuous and 1 Categorical
  • 1 Covariate: Continuous

I am most interested on the interaction of the 2 IVs on the DV.

My question: Do I need to enter the continuous variable into the "covariate" box in SPSS. In that case, do I then change from the "full factorial" model to a "custom" model where I select main effects of each IV and covariate, as well as the interaction of the 2 IVs? or something else?

share|improve this question
I've never found these terms (eg, ANOVA, ANCOVA, General Linear Model, etc) to be very helpful. Ultimately, these are all the same thing. Traditionally, an ANCOVA was when you were primarily interested in the effects of categorical IVs, but also wanted to adjust for some continuous covariates that weren't of substantive interest. In realiality, these are all forms of multiple regression. Part of your confusion is that SPSS makes you ask for analyses using this terminology. I would probably call your situation multiple regression. You should be able to do what you want via that in SPSS. – gung Nov 20 '12 at 22:36

General Points

There is no statistical difference between your two continuous predictors. It is all just a linear model. Your categorical predictor (unless it is binary) needs to be coded in an appropriate way (e.g., dummy coding) but after that they are all just predictors.

That said, I'm assuming that there is a conceptual difference between your variables. In particular, there is a difference in interpretation that results from independent variables that you have experimentally manipulated versus those that you have merely observed.


With regards to SPSS, if you are going to use analyze - GLM - univariate to perform your ANCOVA then you would probably put any numeric predictor into covariates. One exception might be an experimental ordinal factor (e.g., five ordered levels) which you might put into fixed factor and perhaps add a polynomial contrast.

By default this SPSS procedure does not report interactions involving covariates. If you want these you would need to specify them in the custom model. There's an example here.

share|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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