The 2016 US Presidential election is still 15 months away. The first debate amongst the Republican candidates will take place on August 6. To keep the size of the debate manageable, FOX News decided to limit it to the top ten candidates. In a previous article, Fox and Republicans, we made a quick prediction on who would be there.
How did we do?
Must place in the top 10 of an average of the five most recent national polls, as recognized by FOX News leading up to August 4th at 5 PM/ET. Such polling must be conducted by major, nationally recognized organizations that use standard methodological techniques.
As we discussed, these guidelines were very flexible. As Jon Stewart said, "Ah, so basically they're going to look at the polls and [Fox News boss] Roger Ailes is going to pick whoever he wants."
However, that vagueness was attractive. It allowed us to estimate probabilities and make predictions. Our final prediction for the debate was Donald Trump, Scott Walker, Jeb Bush, Marco Rubio, Mike Huckabee, Rand Paul, Ben Carson, Ted Cruz, Chris Christie, and John Kasich. We felt the first nine were safe, and that Kasich had a 3-to-1 lead over Rick Perry.
At approximately 6:00 pm ET on August 4, FOX News made public its decision. It matches our prediction in terms of the list.
What is helpful is that FOX News also listed the polls it used to make the list: "The five polls included in the average that determined the line-up were conducted by Bloomberg, CBS News, Fox News, Monmouth University and Quinnipiac University." This list is interesting as it excludes the NBC/Wall Street Journal poll of July 30, which was more recent than the Quinnipiac University poll of July 28. One wonders why FOX News decided that poll was not appropriate.
For the record, here is the FOX News ordering, with their estimated support percentages according to the video:
From these provided percentages and from the polls used, we can now tell how the "statisticians" at FOX News averaged the polls. They completely ignored the sample sizes and just found the arithmetic mean of the estimates. In our elementary statistics courses, we warn against this; polls using larger sample sizes should be weighted more than polls with smaller sample sizes. The larger polls are more precise.
That is rather interesting.
For more information, see our article at The Conversation.