Most of our research is based on statistical analysis of the expansive WhatNotToSing.com database...but beware. No matter how many years we spend building the DB, and how many train wrecks we have to endure along the way, we'll never really have all the data we need. That's because American Idol is a living, breathing, ever-evolving entity. Not unlike a flu virus.
Thus, rather than include the same boring set of explanations, warnings, disclaimers, drug interaction advisories, and other menial boilerplate in every last one of our Idolmetrics articles, we decided to be merciful. The whole shebang is collected on this page, where we can simply reference it for all eternity.
Most of our studies are based on statistics compiled from semifinals and finals episodes, a.k.a. the "performance rounds". Through just the first seven years of the show, that covers 112 episodes, 193 contestants, 695 unique songs, 983 performances, and roughly two trillion insipid remarks by Ryan and the judges. This is a lot of data. It's enough to identify emerging trends with a fairly high degree of confidence. It's not, however, enough to draw conclusions so firm that we'd advise betting your house that all of them will hold up forever. So, don't bet your house. Got that?
The ever-shifting format of the competition (Top 24 vs. Top 32, wild cards vs. no wild cards, multiple elimination weeks, multiple song weeks, etc., etc.) makes it a challenge to set up consistent experiments. For example, the two occasions when Idol staged a no-elimination week – once in Season Two, once in Season Six – are the absolute %^&*# banes of our Statistics Department's existence. Whenever we fear these "one-off" anomalies might skew an analysis, we take the time to weed them out as best as we can. Other times, we ignore them. Cope.
Many of our studies rely on the WhatNotToSing.com web approval ratings for each performance. If you're new to our site and you're wondering what all those numbers and stars mean, read this. And this. And if you're really masochistic, this. If after slogging through all of that, you still do not feel that our rating system is an accurate enough measurement of the Idolsphere's collective opinion, that's your prerogative and we'll just respectfully disagree (but then, most of our Idolmetrics articles won't have much meaning to you.)
We have no agenda. We promise. If you believe some of our conclusions to be uncomfortable or unfashionable or even flat-out wrong...well, okay. Write us about it and we'll address your concerns, or maybe even re-run the experiment another way to see if the conclusions hold up. But always keep in mind, our goal is simply to advance the collective wisdom of the Idolsphere. Sometimes the truth hurts when you corner it.
We think the producers (and their lieutenants, the judges) are sometimes biased, usually manipulative, and always shameless. But you already knew that. However, at the end of the day, we believe they do try to put on something vaguely resembling an honest competition. Obviously, if every outcome was rigged from start to finish, then all statistical studies we run are meaningless. If you believe that to be the case, you're welcome to your opinion, and who's to say you're wrong? We still hope you enjoy reading our essays, though, if only for the occasional snark.
There's more...something about reading the prospectus carefully before investing, we think. But that's enough for now.
-- The staff of WNTS.com