# Multiple regression

#### marigold

##### New Member
Hi,
I want to do a multiple regression , correlating offspring produced on mating one with offspring produced on mating two also with male age at the first mating as a variable. SO the 'dependent' variable would be offspring produced after second mating with offspring produced from first amting and male age at first mating as the other two (explanatory) variables.
There are over a hundres repeat measures of offspring counted for each mating. But only three different male ages 4days, thirteen days and one day old.
I was using SPSS for glm's then did stepwise in R, with many other variables, now im looking at multiple regression and nothing makes sense to me, i know what i want to do but not sure how?
Do i -> analyse ->regression -> linear ,then put both first mating offspring and male age in the independent box? Not sure what stats / plot to look at?

#### TheAnalysisFactor

##### New Member
I really hope these are insects.

I don't understand what you mean about hundreds of repeated measures of offspring at each mating. Does each male have more than one value for "mating one" or do you mean that there are hundreds of values across the data set?

If the latter, go ahead and use SPSS GLM Univariate.

Since there are only 3 male ages, you could think of it as either categorical or continuous. If you want it continuous, GLM will fit a line to model its relationship with the DV. If this is reasonable, go ahead and make it continuous. If you think it's not linear, make it categorical.

Continuous variables go into Covariates box. Categorical ones go into Fixed Factors. Don't use Random Factors.

You will look at the Tests of Between Subjects Effects, and most importantly the Parameter Estimates table. I recently did a free webinar about how to interpret the Parameter Estimates table. If you're interested, you can download the video recording at: http://www.analysisfactor.com/learning/teletraining4.html.

Karen