1
$\begingroup$

I have a design involving 1 between-subjects categorical factor, 1 between-subjects continuous factor and 2 within-subjects categorical factors.

This is theoretically a case of multiple regression, but as in this question here, I was able to use lm() in R to specify my model (actually, using the aov.car() function in the afex package).

This means that my output was in the form of an ANOVA table, with F and p-values. Can I report my analysis as though it was an ANOVA, or should I find a way to force the output of regression coefficients and report the analysis as though it were a multiple regression? I know that I shouldn't use an ANCOVA in this case, because my continuous predictor variable is certainly not a nuisance variable!

Improve this question
$\endgroup$
1
  • 1
    $\begingroup$ (i) There's no actual difference between ANCOVA and regression (aside from point of view); (ii) reporting an ANOVA table for a regression is fine, if that serves your purpose; (iii) reporting coefficients is also fine, if that suits your purpose. The choice is dictated more by area conventions (what do people expect to see) and what you're trying to show than any major statistical considerations. $\endgroup$
    – Glen_b
    Commented Jun 25, 2015 at 6:32

1 Answer 1

Reset to default
1
$\begingroup$

You can get an ANOVA table even if your model is a simple linear regression with a continuous variable. There is nothing invalid about that. You can consult the journal to see if there are conventions that they prefer stylistically, but you can report the $F$ and $p$ values from the ANOVA table.

On the other hand, if you want to know the point estimates for your effects (e.g., the slope of your continuous variable), you will need different output to see them.

Improve this answer
$\endgroup$
1
  • $\begingroup$ I should be able to specify "lm" output and then use summary() if I wanted to see the slope of my continuous variable. Untested as yet. $\endgroup$
    – luser
    Commented Nov 6, 2014 at 20:32

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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