Quote:
Originally Posted by RonCo
I decided that I need to respond to this, just because it's an argument that comes up several times and is not correct. It is, in fact, quite normal to use very small sample sizes to adequately analyze entire populations.
|
I have never said that the issue is sample size. I said the exact opposite, post 136:
Quote:
Originally Posted by injury log
The issue isn't sample size; Tango's samples are huge. The issue is sample bias, and in the case of Tango's studies, the bias is extreme.
|
I don't know how I can say that more clearly.
If you want to apply Tango to the same population he did, the data is good. For big league players, OOTP should give results similar to Tango- I agree there. If you're extending it to a different population- minor leaguers, say- there's no reason to expect anything more than superficial similarity with Tango data. To use your polling example, it's akin to polling only those with >$10million annual incomes and hoping to get a read on how an election will play out. By looking only at Major Leaguers, you're only looking at guys who
a priori have developed well enough to play in the Majors. That's a pretty extreme sample bias if you're hoping for data that will apply to all minor leaguers.