Sampling 3rd dose after two weeks would be statistically insignificant and not enough time to record stats from sickness.
After the 3rd vaccine, afaik, they compared the rates of hospitalizations of people do did vs. didn't yet get the 3rd vaccine, as compared to what happened the week before. They saw a drop in 3rd-vaccinated, and only in 3rd-vaccinated, implying that it did have an effect. And this wasn't a small population (I don't know exact numbers but I assume several 10s/100s of thousands).