Bloody mice, chewing up everything. And just because our cat thinks she needs to bring in live prey.
ActiVote? What The Hell?
I’ve been excluding ActiVote results because of a perceived erratic rightward tilt, and because it’s not rated nor known to FiveThirtyEight. But today, I glanced at their first three new results, and, much to my shock, two of them were reasonable. Since I’ve been writing this series not only to keep readers up to date on the Senate campaign, but also as a learning experience, I thought I’d put them up as a lesson in … the power of money.
In our hypotheses, we’ll take a potential high rating on FiveThirtyEight’s Pollster Rating Page as a proxy for future earnings potential; the better the rating, the higher the earnings.
We can put together these hypotheses, and expected behaviors.
- ActiVote is a prolific, honest pollster. They are non-proficient, so results are erratic compared to other pollsters, and often diverge from top pollsters. Codename: Honest.
- ActiVote is a prolific, honest pollster, and is excellent. Their divergence from top pollsters is simply the result of other pollster’s poorer results. Codename: Excellent.
- ActiVote is a prostitute pollster, meaning it issues results favoring its sponsors at their request. The goal is to influence voters’ behavior at the polling booth or moral equivalent; accuracy is not a priority. Thus, their results may be much like item 1 results, only more so. While their goal is both dubious and probably impractical, being caught is of low probability, and punishment may be negligible. Codename: Prostitute.
- ActiVote is a grifter of sponsors, meaning it issues results that pleases sponsors without a request to do so, nor without revealing their activities, and with no regard to accuracy. Their goal is to increase repeat sales through pleasing results, often via confirmation bias. Codename: Grifter.
Naturally, for the Prostitute and Grifter there is a tension between immediate profit and high ratings from 3rd parties such as FiveThirtyEight. Thus, for those who qualify in either category and are good pollsters, their last polls should converge with other known excellent pollsters. We may have seen that in today’s batch of polls, but I don’t know if this was the last for our prolific example.
As a practical matter, hypothesis Honest will be considered the most likely if ActiVote’s results prove to be a pseudo-random scatter. Prostitute and Grifter will be indistinguishable, since there is currently no access to ActiVote’s internal communications, and considered most likely if ActiVote’s results change from previous results, converging with Election results in most cases. Finally, Excellence will be the most likely if ActiVote does well in its final results for each region it polls without a shocking change compared to previous results.
I’ll try to remember to return to this topic once Election results are released.
How Is The Abortion Issue Playing Into The Election?
Thanks for asking, I just wrote a post on that.
Examine And Classify The Nitty-Grit … Grit … Grit …
- In Arizona, unknown ActiVote has thrown in the towel and claimed Rep Gallego (D-AZ) has a substantial lead over election-denier Kari Lake (R-AZ) of 54%-46%, or nine points after rounding and maybe a few other dance routines. Not only is it large, but it has Gallego well over 50%.
Data for Progress (2.6) gives Rep Gallego a 50%-45% edge, with a margin of error of ±3 points.
Finally, YouGov (3.0) has the smallest lead for Rep Gallego at 49%-45%, which seems a bit of a puzzle.
The only question in my mind for this contest is why are so many Arizonans voting for Lake?
- Nevada has generated a flood of polls from higher quality pollsters for the contest between Senator Rosen (D-NV) and challenger Sam Brown (R-NV), most of which give the lead to Senator Rosen: Emerson College (2.9) 49%-45%, YouGov (3.0) 51%-44%, Noble Predictive Insights (2.4) a close 48%-46%, and Data for Progress (2.6) with a more substantial 49%-42% lead. Only Susquehanna Polling & Research (2.3) gives Brown the advantage at 47%-46% on a smallish sample size of 400, the smallest of all the Nevada polls this time around, although not by a great deal, and a large margin of error of ±4.9 points.
I’ve been expecting a high single digit to a low double digit victory margin for the Senator; perhaps I’m a bit optimistic?
- There’s an even larger flood of polls in Pennsylvania, all showing Senator Casey (D-PA) ahead of challenger David McCormick (R-PA?): YouGov (3.0) has a very reasonable 50%-44%, Data for Progress (2.6) gives the Senator a reasonable 49%-45%, Suffolk University (2.9) takes 49%-46% and rounds it to a four point lead, Marist College (2.9) a very small lead of 50%-48%, WaPo (unrated) suggests a 49%-46% lead, and finally the problematic ActiVote (unrated) has the Senator leading 51%-50%, presumably through a surfeit of fingers used for counting. Or rounding, or rounding, yes, yes. Stop yelling. ActiVote’s summary? “McCormick Closing in on Casey in Pennsylvania.” Maybe. I kinda doubt it, though.
- Michigan is only getting a rain shower of polls, all finding Rep Slotkin (D-MI) leading hard-right extremist former Rep Rogers (R-MI) for the open Michigan Senate seat: YouGov (3.0) gives Rep Slotkin a big lead of 51%-42%, Marist College (2.9) has the Rep ahead by a smaller gap of 52%-46%, EPIC-MRA (1.9, which is a bit low but I have nothing else against them) is at 47%-42%, and finally the ever doubtful ActiVote, now functioning as a cautionary example, giving Rep Slotkin a minuscule lead of 51%-49%. I think an eight point victory is not out of reach, but we shall see.
- Wisconsin’s up here, where we finally received snow, and next to Lake Michigan, so they’re getting a lake-effect snow squall of polls, all showing Senator Baldwin (D-WI) leading Republican Eric Hovde (R-WI?): YouGov (3.0) shows a lead of 50%-45%, Marist College (2.9) has a measurement of 51%-48%, and ActiVote, oh ActiVote, measures a respectable lead of 54%-46%, which sounds right even if their sample size is a mere 400 voters, with a margin of error of ±4.9 points.
It appears Senator Baldwin is pulling away, if the polling is accurate, which many progressives will dispute. And this Daily Kos article is interesting: The pollsters say Wisconsin is a toss-up. But the polls aren’t matching 2024’s reality.
- But do debates matter for Senate races? I’ve never seen that question asked, much less answered. But Steve Benen claims that the debate between Missouri Senator Hawley (R-MO) and his challenger, Lucas Kunce (D-MO), was a disaster for the Senator, as does some dude named Brian Tyler Cohen. The latter link is to a short clip from the debate, and I hear Kunce saying, Josh Hawley is a selfish swamp creature… I kid you not.
It’s noteworthy that a debate has occurred at all. It may mean that the Senator’s private polls showed that he is in trouble, and perhaps he thought that he needed to win a debate to show he’s the dominant political personality in Missouri. It seems that this idea swirled down the toilet, if Cohen and Benen are to be believed. I haven’t the time to hunt down a recording of the event and watch.
But will it matter? Do Missourians watch debates, or is the state too polarized? There is very little public polling, and I’ve just assumed the Senator will win. But Missourians certainly have good reasons to kick the
bumSenator out, from being just a showboating mouth to encouraging Jan 6th insurrectionists to supporting Christian Nationalists. I guess we’ll have to wait for results from Missouri on Tuesday.
Oh, Stop That!
Polls from the following pollsters, entirely or in part, have been shoveled into the bin, taken to the Pit, and placed on the Spike, who has been tied down and smothered in mushrooms: Torchlight Strategies (no rating, appears to be a lapdog for Senator Fischer) and OnMessage (1.1, which is awful).
Enjoy that visual. And, yes, I meant Spike from Buffy, The Vampire Slayer. Thanks for asking.