I’m guessing that something very odd is going on. Although, I confess that I was barely able to watch the show last night, this doesn’t match my impression of what went on:
|Contestant||WNTS Rating||Dialidol score||Bottom Group Previously?||Prev. Rating||Sex||Probability
of elimination (%)
According to the data, Hollie is in the most danger. She had the second lowest WNTS rating, did reasonably well in Dialidol (which is not very predictive at this stage), and is a woman. That moves her into the range for elimination at a probability of about 40%. This model does not take into account that she has not been in the bottom 3 before, because I have not found that to be statistically significant in the Top 9.
Heejun Han at only 7% is a weird call. I would think it’s wrong, but I’m way out of it this year. There were 5 standing ovations last night, and I didn’t understand a single one of them except for Elise Testone, predicted to be safe (though, again, the model does not take into account her status of having been in the bottom 3 before).
Skylar is again being singled out for being a woman, but does not get the boost of being country and having not been in the bottom 3. I don’t agree with the prediction. I think it will be either DeAndre or Heejun. But the model says Hollie.
If, dear reader, you want to see the results of a curious experiment, here are the predictions of a model with many different variables at modestly significant levels:
|Contestant||WNTS Rating||Dialidol score||Previous Rating||Sex||Probability of elimination|
This one holds Colton is the most likely, due to his WNTS rating going down by so much week-to-week and having a middling Dialidol score. This model also predicts Heejun is safe because his score improved by so much from last week. Take it for what it’s worth, which is probably very little.