[pete fabrizio]

[ QUOTE ]
[ QUOTE ]
My point is that the "bottom 3 helps rally the troops" belief is probably overdone (and yes, largely bolstered by the regression fallacy). The fact that people drift in and out of the bottom 3 is perfectly consistent with normal variance. What the bottom 3 indicates to me is that the contestant is not very popular and didn't have a very good performance. Normally that's bad news.

[/ QUOTE ]
This all seems right.

Being in the bottom 3 is obviously a bad sign because it means you aren't very popular with the voters. But that's not the question I had in mind when I started this thread. Let me restate the issue with a current example.

An unknown performer ("Fred") finished in the bottom three this week. Did American Idol help or hurt Fred by not announcing that he was in the bottom 3?

1. The standard assumption seems to be that this hurts Fred. He's going to miss out on the rallying effect.

2. My intuition is that this actually helps Fred. Most people constantly seek validation of their personal opinions, especially in subjective matters like clothes and music. Seeing Fred in the bottom 3 would have caused many of his fans to start doubting their own ears. Next week instead of enjoying his performance they would subconsciously be picking it apart.

[/ QUOTE ]

I'm sure both of these effects (rallying and tarnishing) are real and do happen, I'm just not sure to what extent, and looking at the history of the thing doesn't help much.

On balance, I think there may be a minor rally right after someone reasonably popular appears in the b3 for the first time, but in the long run the tarnish probably hurts them more.