I think Nate Silver is brilliant and very, very good at predicting elections, but he’s only as good as the data he has to work with, and he is capable of making small mistakes or developing models that are too clunky to capture some of the finer grains in campaigns. One example is his reliance on a “fundamentals” score that really serves as a stand-in for good data in the early stages before he has a robust sample of polls to work with.
If you look at the fundamentals of a state like Kansas, it’s going to look like the Republicans have a huge advantage over the Democrats. The state hasn’t elected a Democratic senator since the 1930’s and it vastly prefers Republican presidential candidates. The voters are very conservative and the president is quite unpopular there.
But this all masks a rather serious split within the Republican Party in Kansas. Under the leadership of ultra-conservative Governor Sam Brownback, moderate Republican lawmakers were purged from the legislature and replaced with wingnuts. Dave Weigel termed it The Great Kansas Republican Purge of 2012. And, yes, Koch Brothers money made it possible.
Much of the dispute focused on moderate Republican opposition to Governor Brownback’s proposed income tax cuts, which were seen as fiscally suicidal. And that has turned out to be the case, with both Moody’s and Standard and Poor’s recently downgrading the state’s credit rating.
The most prominent of those purged Republican lawmakers is former state Senate president Steve Morris:
Brownback’s critics say the degree of the tax cuts is threatening basic services, not to mention the state’s bond rating and requirement to balance the budget.
“People are concerned about education,” said Steve Morris, a member of Republicans for Kansas Values, a group of more than 100 GOP members who are backing [Democratic candidate Paul] Davis. “They’re concerned about transportation.”
Morris is a former Kansas state Senate president who lost a 2012 Republican primary to a Brownback-favored Republican. He said he is backing Davis not because of “sour apples,” but “concern about the direction the state’s heading in.”
The result is that Brownback trailed Davis on all three polls taken in August. Rasmussen had Brownback down by 10 points. Needless to say, when the incumbent governor is doing this badly, the “fundamentals” of the state don’t count for much. Brownback will be on the top of the ticket, not President Obama. And that’s a problem for Sen. Pat Roberts. With a very large contingent of angry moderate Republicans headed to the polls to defeat Brownback, might they also cross-over to beat Roberts?
That possibility is what led me to write about this race last week, and I noted that a Public Policy Polling survey had found Roberts narrowly beating Democrat Chad Taylor in a two-way race but losing to independent Greg Orman in a two-way race. Might Chad Taylor be enticed somehow to drop out of the race? Would Greg Orman agree to caucus with the Democrats?
Well, today, Chad Taylor dropped out of the race, leaving Orman a one-on-one contest with Sen. Roberts. We can only speculate about whatever behind the scene machinations went on to make this happen because Chad Taylor is keeping mum.
Taylor, reached by phone in his car, tells Kansas First News that he turned in the papers to withdraw at 4:15 p.m. Wednesday.
When asked why, Taylor declined to comment and says he “would do some press later in the week.”
As I pointed out last week, Prof. Sam Wang saw this possibility as a complete game changer:
If the [independent senators] and the Democrats win exactly forty-nine seats, Orman would have it in his power to provide—or deny—the critical fiftieth vote to control the chamber. In all the outcomes simulated in my model, this event has an almost thirty per cent probability of happening. Added to the Democrats’ chances of gaining control without Orman, the total probability of combined Democratic and independent control would be eighty-five per cent—a total game-changer.
Nate Silver still gives Roberts a 56% chance of winning against Orman, but that is because his “fundamentals” number gives the Republicans a 25% advantage. That’s an advantage that takes no account of the civil war that has gone on inside the Kansas GOP over the last two years. It takes no account of this, which happened yesterday:
More than 70 former Republican lawmakers announced their support for the independent candidate for U.S. Senate over incumbent Republican U.S. Sen. Pat Roberts on Wednesday.
Traditional Republicans for Common Sense, which is made up of moderate Republicans who have served in the Kansas Legislature, chose to endorse Greg Orman over Roberts and Democrat Chad Taylor.
“Our members know leadership because they’ve been leadership,” said Jim Yonally, the group’s chairman and a former state representative from Overland Park. “Our members have been there. They’ve been on the frontline. They’ve had to make the tough decisions.”
The group sees Orman as a pragmatist who can broker compromises between the two parties, said Rochelle Chronister, who has previously represented Neodesha in the House and served as chair of the Kansas Republican Party.
“This has been the most do-nothing Congress ever,” Chronister said. “They can’t even pass an appropriations bill to keep the government running.”
Like I said, I think Nate Silver is brilliant. But Harry Reid ain’t no dummy either. And I think Reid just cooked Pat Roberts’s goose.
Sorry to be OT, but I just couldn’t resist:
http://digbysblog.blogspot.com/2014/09/rip-mtp.html
Bring on the legacy kids. Can’t imagine that Luke’s following is any larger than Chelsea’s (who recently resigned from her $600,000/year gig (in preparation for Mom’s run for the WH?).
Meanwhile those who want to do the real work of journalists in the field have to do so as freelancers with lousy pay and no institutional training and protection. So, they take high risks to try to get the story and some end up getting their heads chopped off.
As you said, he’s [Silver] only as good as the data he has to work with,
Analysts that use all the data/information and not just the available quantitative data always do better. As the poll numbers were of good enough quality and electorate stable enough since 2002, using statistics to call the elections wasn’t a difficult task. (Not that partisan Republicans ever accepted the numbers when not in their favor. It was truly painful for me to read all the dKos happy talk before the 2002 and 2004 elections when the numbers just weren’t there. Then all the fretting in 2006 and 2008 when the numbers were there.) Silver may have been lucky to enter the election prognostication game when he did. And may not look as proficient as various electorates become unsettled.
Maybe, but it seems that humans are predictable to a very high degree by statistics. I guess we’ll see.
Who’s been better than Silver who uses non-polling data? The only one to beat him has been Wang, and even there mostly only in predicting lower uncertainty, and Wang also uses overwhelmingly polling data.
Silver’s big problem this cycle, as you’ll see if you look at details on his site, is that the polling data is crap. Alaska didn’t have any recent nonpartisan polls last time I looked. Many other critical states like Arkansas and NC have wild swings between polls that show there are some serious methodological issues going on. A lot of his near-50% probabilities basically reflect that he has no idea what’s going on.
Honestly, given how close it is, we’re all flying blind, and we’re really not going to know what’s going on until late on Election Night. More reason to organize and donate, as I see it.
Probably nobody.
But I guess you missed my point that the quantitative stuff alone is often good enough — but the best analyses incorporate the numbers as one of the relevant factors. And should there not be good enough polling data, other not ignored factors are more heavily weighted.
He reminds me a bit of the Wall St. quants and there ever more complex formulas. They worked really well — until they didn’t. Basic statistical analyses of the polls since 2002 with a bit of information on the specifics of the race seemed to be as good as what Silver has done.
Yep. I know my “models” (look at polling in my spreadsheets and throw in some “gut”) got at least as good as Silver, and in one of the GE’s (can’t remember if it was 2008 or 2012) got the vote percentages better than he did.
and I know you came close, too.
I’m not organized or obsessive enough to bother with calculating projective win/lose margins. Win/lose is good enough for such an exercise. However, the actual vote counts and margins are data for future reference.
Not sure I use a gut test factor. It’s usually several more tangible factors. A particularly powerful one is a comparative visual charisma test. While I’m more persuaded by words than presentation styles the opposite seems operative among a majority of the electorate. As the test instrument, it’s easier for me if I have never seen either candidate before which reduces pre-test bias. Watch a debate without sound and score it. If too close to call, add the sound. Will have to try this more often.
Read about the future today at Booman Tribune.
So another Republican that might be enticed to be a LieberDem. My heart beats faster.
you were on this before anyone else, except for Sam Wang.
Good job, BooMan
As I’ve said before, Nate Silver is good in math but has no understanding of politics whatsoever.
Departures from historical patterns can undo statistical expectations of elections by hitting in the improbable area. Whatever Kansas elections are this year, they do not conform to historical patterns of behavior. Whether that plays out in the statistical results remains to be seen in November.
I’m being lazy, but how have Silver and Wang compared when it comes to midterm elections? If one of them is substantially better at midterms, then that is who I’d go with.
If they’ve both been pretty accurate with neither being clearly better, than it goes to show how tenuous midterms are, since liberals and democrats are assholes about voting for anyone but POTUS.
Oops —
Kobach says Taylor stays on ballot for U.S. Senate in Kansas.
How could we all have overlooked that Kobach is KS SOS? He’s one scary and very determined rightwinger/GOP.
So Kobach is ratf-cking again?
Who could have predicted that? I suppose we should be thankful that he’s a RWNJ because otherwise he would be extremely dangerous.
I live in Kansas now… who should I vote for? That Orman dude is going to screw us just like Ben Nelson did. I don’t see the point, really, of preserving a Democratic majority in the Senate. It shouldn’t even be close, but it is, so we might as well let the radical GOP have their way for a while and how much they can screw things up. Nothing is going to happen for 3 1/2 years… at the soonest. And if the country wants to empower the radical GOP some more… why should I care?