When “American Idol” wraps up its 15th and final season on Thursday night, this year’s winner will not be from North Carolina. But no matter who reaches the winner’s circle for 2016, it won’t change the Tar Heel state’s status as home to more “Idol” winners than any other state.
In one of the stranger coincidences of television game-show history, three “American Idol” winners have come from North Carolina: Asheville’s Caleb Johnson in 2014, Garner’s Scotty McCreery in 2011 and High Point’s Fantasia Barrino in 2004. Alabama comes next with two winners (2003’s Ruben Studdard and 2006’s Taylor Hicks). No other state has more than one.
“I think North Carolina fits perfectly with the audience that watches ‘Idol,’ and I’m sure a lot of votes come from the state to keep people in,” McCreery said in a recent telephone interview. “I’d also like to think the main reason is that North Carolina has a lot of talented folks coming out.”
Be that as it may, the statistical likelihood of this happening is rather small.
“North Carolina has about 3 percent of America’s total population,” said Jim Williams, a polling analyst at Public Policy Polling. “So you’d expect it to take 100 seasons of ‘American Idol’ to get three winners from here. Instead, three winners in 15 seasons is 20 percent. North Carolina is producing ‘American Idol’ winners at a rate seven times higher than you’d expect.”
What are the statistical odds of a state with 3 percent of America’s population producing 20 percent of all “Idol” winners? Run the numbers and it comes to less than a 1-in-100 chance – or 0.91 percent, to be precise, according to statistician Richard L. Smith.
But what does it mean?
“That meets the criterion for the result to be ‘highly significant’ in the language of statistical significance,” said Smith, who is director of the Statistical and Applied Mathematical Sciences Institute. “On the other hand, North Carolina is only one of 50 states, and one could have asked the same question about any of the other states. The probability that at least one state had at least three winners, assuming random outcomes, is about 56 percent, which is not at all a small probability. So from that point of view, the result doesn’t look significant.”
After consulting census tables, Smith ran calculations on the problem from a number of different angles, including the application of a “likelihood ratio test” to see if the results are “inconsistent with random.” That yields up a “p-value” (or true significance level, in layman’s terms) of about 0.045 – or one chance in 22 for the actual set of results.
“Usually, statisticians take a .05 level of significance, or one chance in 20, as the borderline for statistically significant,” Smith said. “So this just qualifies as statistically significant, but obviously it’s marginal. There is not strong evidence that the result is anything different from random.”
If one broadens the question beyond just North Carolina, however, the results are intriguingly anomalous. Through the first 14 “American Idol” seasons, 10 winners have come from the South.
Smith calculates the p-value of that at .0085, or a 1 in 118 chance. That qualifies as “highly significant” in the world of statistics.
Even more significant, the March 24 exit of Massachusetts finalist Sonika Vaid insures that this year’s final-season “American Idol” winner will again be a Southerner. The odds of 11 Southern winners in 15 years translates a p-value of .0042 – or 1 chance in 241.
But obviously, long shots do come in.