2
$\begingroup$

I wanted to carry out a two-way ANCOVA for my data. However, SPSS isn't liking that I have only one IV.

I have

  • One IV: Group (3 levels, i.e. 2 experimental groups and a control group),
  • One DV: outcome measure (2 levels, i.e. time 1 and time 2),
  • One covariate (I have checked and this demographic continuous variable moderately correlates with the DV).

Please could someone help as to whether I am missing something in terms of putting this into SPSS?

Improve this question
$\endgroup$
3
  • 2
    $\begingroup$ you sound strange. "Two way" ANOVA or ANCOVA needs 2 independent categorical vars ("factors") by definition. Your data is one-way ANCOVA. Next, because your predictant is only 2 levels (fixed, I assume) you'd better go to Binary Logistic regression procedure instead of GLM. $\endgroup$ – ttnphns Mar 14 '11 at 16:08
  • $\begingroup$ @ttnphns Assuming "you sound strange" means "let me correct you on some points" and "predictant" means "dependent variable," I agree with what you're saying. $\endgroup$ – rolando2 Mar 14 '11 at 21:50
  • $\begingroup$ I think the confusion here is a matter of language. "Time" is your second IV. The DV does not have two levels. "Time" has two levels. Your DV, I am assuming, is metric. $\endgroup$ – Jeromy Anglim Apr 15 '11 at 3:48
2
$\begingroup$

I am guessing that you have your data in wide format. This seems to be causing you confusion because in wide format it is less obvious that "time" is your second independent variable.

One option in SPSS for doing what it sounds like you want to do:

  1. Analyze -- GLM -- Repeated Measures
  2. Enter time as repeated measures factor with two levels
  3. Enter Group as a between subjects factor
  4. Enter the dependent variable (presumably represented as two variables in SPSS as your two times points)
  5. Enter your covariate in the covariate box

For more information, see Andy Field's SPSS Repeated Measures tutorial.

However, given that repeated measures ANOVA already controls for individual differences in some respect, you may want to think about the meaning of including a time-invariant covariate in such a model. See here for a discussion of the issues related to covariates in repeated measures ANOVA.

Improve this answer
$\endgroup$
1
$\begingroup$

Set it up as a regular regression, they are mathematically equivalent, just make sure your non-continuous variables are defined appropriately (as categorical or whatever) to SPSS.

Improve this answer
$\endgroup$
1
  • $\begingroup$ this question was dramatically altered after this answer was made $\endgroup$ – Patrick McCann Apr 21 '15 at 16:43

Your Answer

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