# Thread: Comparing regression coefficients between groups

1. ## Comparing regression coefficients between groups

Hello!

I am a psychology student and I’m doing a research which I need help with.

The research is about parenting stress in parents of children with autism. The criterion variable is parenting stress and predictor is social support.

One hypothesis of the research states that “social support provided by partner, friends and family is a significant predictor of parenting stress and its predictive value will not differ for mothers and fathers”.

I assume that I need to compare regression coefficients between groups, but I don’t know how to do that in SPSS.

Mila

2. ## Re: Comparing regression coefficients between groups

parenting stress (continuous) = social support (continuous) + sex of parent (binary) + error

So for clarification, you don't have data for two different people within the same household?

3. ## Re: Comparing regression coefficients between groups

I have two regression models, one for mothers and one for fathers. The only predictor is social support and it's a continuous variable.

4. ## Re: Comparing regression coefficients between groups

Also, I don't have data for two different people within the same household.

5. ## Re: Comparing regression coefficients between groups

Is it feasible to merge all data together and add a term for sex into the model? Then you can see directly from the coefficient for the sex term in the model that there is no difference between the groups, while also controlling for the support variable and estimating its association.

6. ## Re: Comparing regression coefficients between groups

Is it feasible to merge all data together and add a term for sex into the model? Then you can see directly from the coefficient for the sex term in the model that there is no difference between the groups, while also controlling for the support variable and estimating its association.
Ack, if observations are independent (never father + mother from the same family are sampled), then one should collapse the data.

Since the original question was whether the effect of support differs between men and women, an interaction term sex*support should be added.

With kind regards

K.

 Tweet