We adjudicate between these two sources of inconsistency by replicating analyses
across four different statistical strategies using the same population-based data set. If
our findings are consistent and robust across methodologies, this would suggest that
cohort differenresults replicate the range of estimates observed in past research, then we will argue
that past inconsistency is likely due to choice of statistical method.ces may underlie past observed inconsistencies. On the other hand, if