No matter how experienced you are, you will still make the most obvious errors, and sometimes, (if you are me), the more obvious it is the harder it is to spot.
I mentioned wasting time trying to figure why my quintiles weren’t exactly 20% each when the answer was pretty obvious.
I am getting better, though. I re-ran an analysis yesterday and spent a few minutes trying to puzzle out why it was very close but not exactly the same as the prior results. Then, it dawned on me that this analysis takes a random sample from the population and when I re-ran it, of course, it took a DIFFERENT random sample. Given the good old central limit theorem, these results were similar but not identical. (Although, theoretically, they could have been very different, that would have been rare.)
Did you spot it? Well, I didn’t. Of course it is supposed to end with html.
So, today’s lesson, for the 1,467th day in a row is, check the obvious things first.