I’m going to let you in on a little secret. Walter Whetsell is a Republican operative who worked on Rick Perry’s 2016 presidential campaign. Before that, he created West Main Street Values PAC for Senator Lindsey Graham’s 2014 successful reelection bid. He is the founder of Starboard Communications, a public relations and marketing outfit based in South Carolina.
Now I want you to look at the recent polls out of South Carolina and see if you notice anything.
You can go back a little further and see that that there have been eight polls of the race between Graham and Democratic challenger Jaime Harrison done in October, and six of them have the race either tied or within two points. Excluding the Starboard Communications polls, four of the last six surveys have Harrison in the lead and the other has a tie.
To be sure, the exception is the highly-rated New York Times/Siena poll and they found Graham up by six in an October 15 release. My point isn’t that Harrison is definitely in the lead. My point is that there’s no information about Starboard Communications in the polling averages except that they have a B/C rating from FiveThirtyEight. There’s nothing to let you know that the pollster ran a super PAC for Graham the last time he was up for election. And Starboard isn’t just a little off from the consensus of polls. They’re a substantial outlier, and yet FiveThirtyEight lists their survey as carrying the most weight in their projection, which currently gives Graham a 79 percent chance of winning. That’s mainly because they’re the most recent poll, and their overall grade isn’t too bad. But doing mostly honest polling to get a decent grade is precisely what allows them to inject a last second poll for Graham that strongly moves the average of polls projection.
If Graham were actually anywhere near nine points ahead, and if he really thought he had a 79 percent chance of winning, he wouldn’t have just agreed to do one more debate with Harrison on the Friday night before the election.
Yeah there are a bunch of polls out there that are not explicitly partisan but are clearly disinformation campaigns that try to build attention to some “narrative”. Silver uses these because he doesn’t want to arbitrarily decide which groups fall into this, but it would be nice to have the model without such crap data inputs (the economist I believe excludes trafalgar for instance). Others include the AM greatness, Ras, and SasquehannaPR or whatever. These are not polling firms really, they are propaganda.
But coming to Harrison, it’s hard to believe he has more than a 1/4 shot winning. I’m sorry but SC is irredeemably racist and there’s not enough non-white growth to counter it this cycle.