3 ms·
> innovative bias-correction strategy. It might be innovative, but I don't think it's valid. What if having a longer name makes your 30% more likely to change
by bostonpete 13y ago
> innovative bias-correction strategy.
It might be innovative, but I don't think it's valid. What if having a longer name makes your 30% more likely to change your name and being richer makes you 30% more likely to change your name?
The Scrabble metric would still show that name length is a factor, but it doesn't necessarily serve as evidence that other factors weren't in play as well. Selection bias seems like an obvious suspect here and this Scrabble discussion just seems like hand-waving to me.
- azernik 13y agoMy assumption is that they ran a correlation test, which checks how much of the variation in name-changing is explained just by Scrabble score. If that's high enough, there just isn't much room for other factors that aren't themselves correlated with Scrabble score.