I read this at Disciples of Uecker, from back on July 4th:

Nearly everybody who reported on the Brewers clubhouse over the past few days has mentioned how McGehee didn't seem like himself, how his head just didn't seem right.

Link.

This prompts me to desire an experiment where beat reporters do not know the outcomes of games or how individual players are doing, but are still required to conduct interviews in the locker room before and after games and make judgments about the body language and psychology of players. Were the writer not to know that Casey McGehee was hitting .221, would s/he think there was something off about him?

This is a different side of the same type of bias that causes us to think that players aren't hitting the ball hard when they make outs. Over the course of a large sample, of course, players who hit poorly are likely the same as players who have poor statistics. But in the short run, it's very easy to think about a player's stats and selectively remember his weak grounders, popups, and strikeouts.

And all of that is, I think, related to one of the types of data issues that Colin Wyers and others have written about regarding batted balls -- are line-drives more likely to be scored hits, or are hits more likely to be marked down as line-drives?

None of this is to say that these effects don't exist. Old-timers like to claim that statheads think players are robots, so I guess I have to expressly say that this isn't true. Mood and energy level and psychology surely matter, players surely do develop mechanical flaws and go through periods where they don't see the ball as well, and, obviously, line-drives are more likely to be hits.

The point isn't what is and is not true. The point is what we can and cannot judge.