I am planning to run a set of analyses examining growth in variable A across 3 time points nested within individual children with a specific disability. I want to use mother variable B and father variable B to predict the intercept and rate of change in children's growth in variable A.
The families consist of: 1 child of interest, 1 mom, 1 dad (though there is missing data for some families - a few that don't have dads, 1 or so that doesn't have a mom). ~160 families.
I originally conceptualized this as a model including two levels - Level 1 (repeated measures within children) and Level2 (individual children). However, I am struggling to figure out conceptually how I can enter mother and father predictors in the same model, given mothers' and fathers' data is nested within the family unit and therefore influences each other, but there is not variance in the outcome variable A for children within families since there is only 1 child per family. Thus, a 3 level model (families, parents, children) doesn't seem to fit in my head since there will not be variance within parents in the children's variable A.
I don't know where to look for this specific kind of issue. Does anyone have some guidance or suggestions of places to look?
One easy way to handle this is to take a single-level, wide approach where you arrange your A variable as 3 variables (3 columns in the data) corresponding to the 3 time points. The mom and dad variables are just regular time-invariant covariates. So that's regular growth modeling in line with chapter 6 of the UG.