I've mentioned this before: one reason I love baseball is that it makes a great sandbox for statistical thinking. It features both a rich set of data and a steady supply of ignorance. And the lessons apply directly to the rest of the world. Here's an example. In the baseball world, sportswriter Chaz Scoggins explains his Hall of Fame ballot choices. He's gracious to talk to Baseball Prospectus, a primary source of the new wave of scientific analysis, and a forum in which his reasoning is not respected.

DL: Tim Raines received only 122 votes, one of which came from you. What impressed you about Raines?

CS: I look at Raines in a lot of the ways I look at Rickey Henderson. ... so he put the whole package together. But I did have to think about him; he didn't come as an obvious pick. When I got the ballot, I had to think about it for a couple of days, but I ultimately decided that Raines fit my criteria for a Hall of Famer.

DL: You didn't vote for Bert Blyleven. Why not?

CS: I just feel that Bert Blyleven was a little better than a .500 pitcher. I just never felt that he had the fortitude that it takes to win big games. People say that he had the misfortune of playing on a lot of mediocre and even bad teams, but to me, if you're a Hall of Fame pitcher you're able to lift your team up ...

DL: You also didn't vote for Alan Trammell. Did you give him serious consideration?

CS: I really gave him no thought. To me, Alan Trammell was just a very nice player. ...

It's fairly clear, especially if you read the entire interview and also read the stathead analysis in Baseball Prospectus, that Scoggins' selections are entirely subjective. His gut feeling tells him who is worthy, and then he cherry-picks statistics as needed to rationalize his choices. Now observe the same thinking in another context:

A task force created by 49 state attorneys general to look into the problem of sexual solicitation of children online ... concluded that the problem of bullying among children, both online and offline, poses a far more serious challenge than the sexual solicitation of minors by adults.

[...]

Richard Blumenthal, the Connecticut attorney general, who has forcefully pursued the issue and helped to create the task force, said he disagreed with the report. Mr. Blumenthal said it "downplayed the predator threat," relied on outdated research and failed to provide a specific plan for improving the safety of social networking.

"Children are solicited every day online," Mr. Blumenthal said. "Some fall prey, and the results are tragic. That harsh reality defies the statistical academic research underlying the report."

Notice the dichotomy. Statistics are not reality. Blumenthal knows the facts, and "statistical academic research" is not the facts. The actual reality, of course, is that "statistical academic research" is one of the most powerful tools available to understanding reality. It can be wrong in many ways, from bias to systemic error to total conceptual failure, but "it contradicts what I know in my heart and my vested interests to be true" is not a very good counter-argument.