Question for the group: How might one go about modeling experimental data with twin respondents?
I ran a single-factor survey experiment with a small group of twin respondents (MZs and DZs, mainly same sex). There were two conditions. I randomly assigned twins to receive either:
Conditions A A
Conditions B B
Or
Conditions A B (or B A)
I'd of course like to decompose variance in responses to the treatment, but I also have to account for condition. How would you model these data? Any suggestions would be greatly appreciated. Thank you in advance. -Chance
Would it suffice to include, as a covariate, a per-individual dummy variable for condition, and just decompose the residual variance in the usual ACE or ADE way?
Possibly. Though would that covariate alone accurately capture a possible interaction between genetic relatedness and random assignment to unique conditions? (Let's say condition A is the treatment, condition B is the control).
Could you say more about the interaction you have in mind? For instance, do you hypothesize that experimental condition moderates the ACE variance components? Or is it more like, the between-twin phenotypic correlation is a function of zygosity and whether or not the twins were assigned to the same condition (something kind of like how the sibling correlation depends on number of alleles shared IBD at a given locus in linkage analysis)?
I suspect the latter: Correlation on the experiment’s DV as a function of zygosity and condition.