I'm not going to get into a pissing contest with someone over the internet. As I mentioned before, it's nothing with you personally, I even apologized to you in the post, but I guess you're a tad sensitive-- very prideful. Again, sorry it was your post specifically. If you don't like my stats advice, you don't need to take it. Your opinion on what I know doesn't matter, so I'm not going to try to defend what I know or don't know to you.
You don't need thousands of data points with representative sampling. There are also many other ways to determine under or over prediction than the suggestion I offered, but eyeballing and using single cases is one of the least scientific and least methodologically sound options available. You can use it if you want, but it's not going to serve you well.
Congratulations, you understand sampling variation and the difference between an average or group's parameter value and individual values...
Right-- this is actually why people who know statistics argue that medicine needs to start incorporating more Bayesian statistical methods because the traditional Frequentist methods aren't typically geared towards answering questions for the individuals or single cases as they focus more on long run result without statements for individuals (which seems to be what you want). Your method of eyeballing a score is neither and can't be used with any measure of reliability that you would find in an actual methodology. Again, use it if you want.
Thank you so much for pointing that out! It's an incredibly useful observation (although I don't recall making the assertion that UWSA over or under predicts, especially in an absolute manner, which is the only thing your "piece of evidence" would be useful for refuting). It's just as useful as the observation you'd make for
every single person who took a UWSA before there test, and it would change quite frequently, not to mention it's always going to be an after the fact statement (which doesn't help people much before the exam...). It seems like a pretty meaningful observation. I don't know how I didn't see that before
What a powerful conclusion, you really know your stats! You forgot to tell us that the question you were trying to answer on an individual level, not on average, as you noted, could be answered with some Bayesian methods, maybe even a couple Frequentist ones, too, although you have different perspectives with either school of thought. You probably knew that, though, and were waiting to enlighten us all. Since I spoiled the surprise for you, maybe you could share with us the specific Bayesian or Frequentist method you'd prefer to use to answer the question? Sharing your vast knowledge would be far more beneficial than picking a fight and flexing your internet muscles...