Stop us if you've heard this one. Two men are walking through the African grasslands when suddenly they see a hungry cheetah, teeth bared and headed straight for them. Their guns are broken, and they have no place to take cover. Both men drop to their knees – one to pray, the other to tighten the laces on his sneakers. The first guy says to the second, "What are you doing, you idiot? You'll never be able to outrun that cheetah!" And the second guy replies calmly, "I don't have to outrun the cheetah. I just have to outrun you."
(* rimshot! *) Thank you, thank you, we're here all week, and please don't forget to tip your waitress. But seriously folks, this ancient joke has a connection to American Idol, particularly in the wake of the stunning results from Andrew Lloyd Webber Night. Jason Castro and Brooke White, who delivered a pair of 1-star clunkers, were sent to safety early, thus allowing Syesha Mercado and Carly Smithson to set a new WNTS.com record they surely wanted no part of. The two ladies made up the Bottom Two/Three with the highest average performance rating, 74.0, in the show's history. (The previous record holders were a threesome from AI3 you might remember: Fantasia Barrino, LaToya London, and Jennifer Hudson, at 69.3.)
Afterwards, the Idolsphere got together and calmly asked each other, "WHAT THE #*%&^ JUST HAPPENED??!" To which we reply: Such language! What would your mother say? Here, take this bar of soap and go wash out your mouth, then have a seat in the corner with your blankie. You'll need to regress to your childhood days to understand the common Idol voting phenomenon that just transpired, one we've dubbed The Sesame Street Effect.
The Sesame Street Effect works like this. Imagine a Finals episode in which two popular contestants both happen to deliver train-wreck performances on the same night. Now it comes time to vote. Ernie's loyal fans reason, "OK, he was bad, but Bert wasn't any better, and Ernie so deserves not to go home before Bert!!" Thus, they dial and text for two solid hours. Bert's fans, of course, reason exactly the same way. One night later, Bert and Ernie are both safe, and America waves goodbye to Kermit.
The Effect is actually quite a reliable voting predictor. In one form or another, it explains many of the most shocking eliminations over the years. It's a by-product of American Idol's ladder-elimination format, its unlimited voting system, and basic human nature.
Let's start our study by seeing if you're smarter than a first grader. The subject is Math. On what percentage of American Idol episodes, do you suppose, is the contestant who delivered the lowest-rated performance of the night also the one who happens to be sent home by the voters? To keep things simple, we'll limit this to Finals (not Semifinals or Wildcard) episodes where each contestant performs just one song. Lock in your guess: is it A. 65%, B. 55%, C. 45%, or D. 35%?
The correct answer is...E. 24%. That wasn't one of the choices, you scream? Too bad – Fox's business model revolves around embarrassing average Americans on national TV. Take full credit if you guessed 35%, the closest option. But in fact, in the 41 episodes that meet the criteria (we won't count AI2's Final Nine or AI6's Idol Gives Back, when no one was eliminated), the contestant who delivered the lowest-rated performance was dismissed just 10 times.
Why is this percentage so low? We think the answer is obvious: fans of the lowest-rated contestant usually recognize that their boy or girl is in trouble, and they vote like crazy to save them. We've termed this the Mighty Mouse Effect, after the old cartoon superhero whose motto was "Here I come to save the day!" (Yes, we watched way too much TV growing up. Sue us.)
The chart at right shows the distribution of eliminations to reverse order of approval ratings. As you see, thanks to the Mighty Mouse Effect, it's statistically most dangerous to produce the second-lowest-rated performance of the night. If you do, you'll find yourself singing at the close of the Results Show roughly three times in ten. Note, though, that America does manage to send home one of the Lowest Two more than 50% of the time. Not great, but not bad.
But what happens if both of the Lowest Two happen to be well below the third-lowest singer of the night, making it clear to all that one or the other deserves to go? That's when Mighty Mouse and Super Grover join forces. Not only do their fans rally to save them, but they have a powerful added incentive: they sense weakness in another contestant. Ernie doesn't have to outrun the cheetah. He only has to outrun Bert.
Of the 41 episodes under discussion, we've identified 13 in which the Lowest Two either truly stunk out the joint compared to the other contestants, or when they both delivered performances abnormally below their usual standards. Let's look at how the voting transpired, and who played the role of Kermit on those fateful nights. (A green check mark means that neither of the Lowest Two went home; a red X means one or the other did.)
One final point before we tally the results. We didn't count the AI5 Final 10, in which three contestants famously turned in a 1-star disaster, because it was outside our target group. We were looking only for cases in which exactly two contestants' fanbases recognized that their favorites were in the most trouble. If Elmo happens to turn in a stinker along with Ernie and Bert, all bets are off. Fortunately for all of us, that very rarely happens.
Ready for the final tally? Of the 41 episodes in the study, the Sesame Street Effect was fully or partly in play on 13. On those, one of the Lowest Two went home four times; that's 31%. In the other 28 shows, one of the Lowest Two went home 18 times; that's 64%. Forty-one trials isn't enough to draw firm conclusions, only to observe trends. Still, it seems as though American Idol is one of the few places on earth where if two people snooze, neither one is likely to lose.
(Late breaking news: Since this article was originally published, we heard from readers who felt that we overlooked a few "hits," plus one obvious "miss". One that we agree deserves to be in the Hit column is the AI2 Final 8, where neither Caldwell nor Carmen Rasmusen got above 16 and Smith, third-lowest at 46, was sent home. That omission was a garden-variety brain cramp on our part. On the other hand, we consciously decided not to count as a Miss the AI6 Final 12, in which Malakar and Brandon Rogers both were in the 1-star range and Rogers was eliminated. But upon further review, that episode occurred one week before Sanjayamania truly took off with You Really Got Me, so we'll give it an X after all. The changes don't affect the Effect's tally very much: instead of 31% to 64%, it's now 33% to 65%. Carry on.)
So what can be done to fix the problem? The producers' response might be: "Problem? What problem??" Remember, to them Idol is neither a singing competition nor a sporting event. It's an extended, televised focus group study, nothing more and nothing less. If a contestant has a fanbase that will stay fiercely loyal through good performances and bad, so much the better. It suggests they'll probably stay loyal through his ups and downs as a recording artist, too. For further study, see the career of Clay Aiken.
It should also be noted that none of the Effect's victims were likely to win the title. That includes Hudson, Johns, and Smithson. The producers probably don't care very much in what order the contestants finish. They only seek to identify the ones who have staying power. Had Archuleta been caught in the Effect's backdraft last week (and we suspect he wasn't as far off as many would guess) it would certainly have caught the bosses' attention. As it is, it's just another bump in the road to the inevitable David-vs.-David Finale.
The Effect is much more of an issue to us viewers, who do see the show as more of a sporting event, and to whom the order in which the contestants finish does matter. Recognizing that the root of the problem lies in the voting system, many people across the Idolsphere have proposed sweeping changes over the years. We'll take a look at them more closely in next week's essay, and we'll even make a modest proposal of our own that might satisfy both the producers' agenda and the fans'. In the meantime, we gotta go; there's this cheetah we have to outrun....
Since this article first appeared in April of 2008, we've come to realize that the Sesame Street Effect is more prevalent than we originally thought. We'd concentrated only on those situations in which two contestants sang abysmally on the same night. Perhaps in the olden days of American Idol, when disastrous performances were more common, bad singing was the primary trigger of the Effect.
Today, however, we recognize multiple train wreckage as special cases of a more general rule: The Sesame Street Effect applies whenever the voters perceive that the "real" decision on the night is which contestant ought to go home, not which ones ought to stay. In other words, if a show's circumstances make it obvious that one of two contestants deserves a dismissal, the voters will cast their ballots so that their voices are heard on that case. Then, the Effect will do what it usually does: inflate Ernie and Bert's vote totals so that they're swept to safety past Kermit, the lowest vote getter among the Should Be Safe contingent.
One famous episode in which both Ernie and Bert sang well is the AI6 Final Three. Melinda Doolittle's record-setting approval rating suggested that the voters assumed, rightly or wrongly, that she was a lock for the Finale. The decision was widely perceived to be who should join her: Blake Lewis or Jordin Sparks, both of whom had sung perfectly decently that night and throughout the season. The Effect was nonetheless triggered, and Doolittle, who hadn't been in the Bottom Three all year, was sent home a night later.
Another trainwreck-free case occurred the night of the AI8 Final Five. All five contestants had sung very well in one of the highest-rated episodes ever. It was almost universally assumed that Adam Lambert and Danny Gokey were safe, particularly since the judges had anointed both as the likely Finale duo most of the season. Voters thus naturally believed they were being called upon to decide who should go among the other three finalists, particularly Allison Iraheta and Matt Giraud, both of whom were no strangers to the Bottom Three. The next night, Iraheta and Gokey were safe, while Lambert and Kris Allen made their only Bottom Three appearances. Giraud, however, was simply too far behind the Top 4 to catch a Kermit. (Two weeks later, still twitching over this close call and haunted by Doolittle's elimination, Simon Cowell practically got down on his hands and knees and begged viewers not to assume that Lambert was safe at the end of the AI8 Final Three.)
Whatever its cause, the Effect is a side-effect of a badly broken voting system that assumes that people will vote for one or two favorite contestants they want to keep, week in and week out. That naively ignores the fact that Idol fans do care in what order the contestants go home, even when they realize that the singers in question have no business winning the title. As long as the producers give America two completely different jobs to do, but only one tool to do them with, the Sesame Street Effect will always be with us.
- The WNTS.com Team