As long as neither are categorical with greater than 2 groups, I would agree.
Hi!
I have a homework question that on the face looks rather simple to me, but I want to check with people who may have greater statistics knowledge than I!
With this regression:
var1 = β0 + β1var2 + β2var3 + β3(var2 × var3)
How many regression coefficients are used to represent var2 and how many are used to represent the interaction?
As far as I can tell, I would say one for both...or is there something about the interaction being representative of the individual variables in a regression?
Any help would be much appreciated!
Thanks,
smashing
As long as neither are categorical with greater than 2 groups, I would agree.
Stop cowardice, ban guns!
Thanks!
How does it change if some of the variables are indeed categorical with more than two groups?
Binary-categorical interaction: consider the following linear regression model: (Notice that
for simplicity, the variable with bold fonts represents a vector which can contain more than one
predictor.)
BMI = β0 + β1physact + β2nonwhite + β3(physact × nonwhite)
Binary-continuous interaction: consider the following linear regression model:
BMI = β0 + β1SBP + β2diabetes + β3(SBP × diabetes)
Categorical-continuous interaction: consider the following linear regression model: (Notice
that for simplicity, the variable with bold fonts represents a vector which can contain more than
one predictor.)
BMI = β0 + β1globrat + β2SBP + β3(globrat × SBP)
Continuous-continuous interaction: consider the following linear regression model:
BMI = β0 + β1glucose + β2SBP + β3(glucose × SBP)
I'm being asked how many coefficient terms represent the interaction, as well as globrat and physat.
Any help would be much appreciated!
Thanks,
smashing
Tweet |