But consider this: what if I published a paper based on trends in worship attendance? What if I followed a large group of people over 20 years and tracked their attendance with phone data or surveys? Would this data be better? If so, in what specific way? (Here is a hint: there is a big problem that following people through time solves in these situations.)