Howdy, Stranger!

It looks like you're new here. If you want to get involved, click one of these buttons!

Supported by

Multiple linear regression with a categorical IV.

I have a memory experiment with the following design:

Interval (IV, continuous): Memory delay in seconds.

Probe (IV, categorical): Whether the first or second memoranda is probed in the memory test.

The DV is the response error (continuous) for each trial.

According to Andy Field's textbook, this type of design should be analysed with either a multiple linear regression, or an ANCOVA. I would like to do this analysis in a Bayesian way, and I want to assess the interaction.

I don't think ANOCOVA is the right choice, because I don't want to "control" for the interval, and it doesn't appear to give the interaction. Likewise, if I dummy code the categorical variable and put it into a linear regression, it will give me BFs for the main effects, but not for the interaction. Both analyses give me a BF for "interval + probe", which I assume is the model containing both models, but not a model with the interaction.

Am I on the right track here?


  • Hi Rudale,

    When you say that your interval measure is continuous, do you mean that it can take on any value higher than 0 seconds? If you manipulated it, it seems it might be a factor with several levels (e.g., 2s, 4s, 8s). If it is the delay before someone answers then yes, that should be a predictor in a regression. When you also want to add a categorical variable to that regression, you are in fact doing an ANCOVA. So I think I agree with Andy F. here.



Sign In or Register to comment.