Possible that selecting multiple observations of the same case will have the effect of holding constant some off the true causes of the outcome of interest, even if the dependent variable spans a considerable range
Simultaneously: whatever potential causal factors do vary within the single case over time will seem to explain differences in the outcomes