My dataset is a stability dataset where I am looking at the degradation of a purity measurment (dependent variable) of multiple batches (categorical variable) over time (covariate). I have data from a total of 7 batches. Three of the batches have a purity measurement from four timepoints and 3 of the batches have a purity measurement from only two timepoints.
I want to assess the slope of all of the batches, but the batches with only two timepoints will have slopes that have no error. Should I exclude the batches with only two timepoints? Is in inappropriate to include them? I need a real statisticians point of view, because I just dabble and this is dataset is confusing to me.