How should PEC be graded? As the last few polls trickle in, let me give a suggestion for how to evaluate predictions after the election. Late tonight I’ll give actual predictions (and give you a chance to record your own predictions).
November 3rd, 2014, 7:00pm by Sam Wang
Coming into the home stretch, President Obama’s net approval/disapproval rating is at minus 8%. Not good…but 4% better than June. This is what candidates face as in-person voting starts tomorrow morning.
November 3rd, 2014, 10:05am by Sam Wang
I am thinking about how to get the most accurate last-minute snapshots of races, and how to turn that into a scorecard for you (and me) to use on Election Night. I’m also thinking about Brier scores as a means of evaluating the various prognosticators, including me.
In the meantime, here’s your morning reading: an excellent analysis by David Rothschild, Sharad Goel, and Houshmand Shirani-Mehr on the problems endemic to current polling practice. They analyze 2012 in great detail, and identify errors that go well beyond sampling error. Because these errors are unlikely to have been fully corrected this year, they think there’s a good chance that Democrats will outperform poll aggregates. In other words, all poll aggregators, including PEC, might carry a hidden bias. My own view is that based on historical data, errors have gone in either direction by several percentage points across the board. Go read their article!
November 2nd, 2014, 12:00pm by Sam Wang
I got into poll aggregation in 2004 to reduce endless chatter about outlier polls. Hmmm, how’d that work out…
— Sam Wang (@SamWangPhD) November 2, 2014
At least one journalist is chattering about whether there’s a late break in polls for Republicans…based on one data point, which is probably statistical noise. Some people are hopeless. Then again, several polls today have pushed the Meta-Margin almost as far as it’s been toward Republicans this campaign season.
He is missing a far more important point: Final election results can vary across-the-board from midterm polls – in the same direction. This last-minute polling bias is typically 2-3 percentage points – five times larger than the bias in presidential years. The direction of the bias is unpredictable. (Read this for a review of the subject.) This is why I care about the exact margins for front-runners McConnell (R-KY) and Shaheen (D-NH). Their states are early-reporting. From them we can make a rough estimate of nationwide polling bias, as follows. [Read more →]
November 1st, 2014, 10:15am by Sam Wang
New Yorker: Tweet
Republicans will probably win the national House popular vote, but even if they didn’t, it wouldn’t matter. Why not? In The New Yorker, I discuss gerrymandering (a big cause) and population patterns (a smaller cause).
November 1st, 2014, 8:45am by Sam Wang
This year, an unusually high number of incumbents are threatened, both governors (last elected in the wave of 2010*) and senators (last elected in the wave of 2008). Previously, I identified 14 races where party control will change or the incumbent is at serious risk. In the home stretch, many governors have recovered slightly, but are still at risk. At least two will be turned out of office, in Kansas and Pennsylvania. Here are the poll medians of surveys completed in the last two weeks (or the last three surveys, whichever is more data).
October 31st, 2014, 9:37pm by Sam Wang
I’ll be on CNN (Smerconish, sometime between 9 and 10) and MSNBC (Up with Steve Kornacki, live between 9 and 9:30). Key point on CNN (watch it here): is this Tuesday’s election an Obama referendum, or an echo of previous wave elections in 2008 and 2010?
October 31st, 2014, 12:05am by Sam Wang
Senate polls in individual states have moved around…but the Meta-Margin and the average seat count have stayed stable. Nonetheless, the crystal ball is cloudy. Why is that? The Midterm Polling Curse. Spoooooky!
As I wrote last week, everyone’s calculations are, to an extent, built on sand. Historically, in any given year midterm polls have been off in the same direction by a median of 2 or 3 percentage points. Depending on the year, either Democrats or Republicans end up outperforming polls. In current poll medians, six races are within less than 2 percentage points: Alaska, Colorado, Georgia, Iowa, Kansas, and North Carolina. Therefore all six of these races could be won by Republicans…or all six could be won by Democrats.
The other races total 48 Republicans and 46 Democrats/Independents. Republicans are slightly favored to take control, since an even split of the six close races would give them the 51 seats they need*. However, the likely possibilities range anywhere from a Republican majority of 54-46 to a Democratic majority of 52-48. As of today, cranking through the math and the uncertainties gives a probability of 55% for a Republican takeover.
I am thinking about how to guess on Election Night where in this range the race will land. I’ve decided that key races to watch are…Kentucky and New Hampshire. Wait, what?? Here is why. [Read more →]
October 29th, 2014, 9:43am by Sam Wang
Over at FiveThirtyEight is a new essay saying that there hasn’t been any concerted movement during this campaign season. Broadly, I agree with their objection that journalists and pundits overuse (and thoughtlessly use) the word “momentum.” However, in this case a close look at the data suggests there was strong movement from September through early October – which then stopped or reversed slightly.
The movement is easier to see in a polls-only snapshot, which is what is published here. In particular, Iowa and Colorado moved toward Republicans throughout September – and in the last few weeks, moved back toward Democrats. One possible cause is early voting – which should show up in surveys. [Read more →]
October 28th, 2014, 9:24am by Sam Wang
(click the above image for a high-resolution version)
The amount of Ebola coverage is amazing: 1,869 stories from October 20 to 24 alone. That coverage came on the heels of the death of one patient in Dallas, Texas. The level of coverage is amazing considering the far greater impact of other infectious diseases in the United States: rotavirus, which kills dozens of small children every year; West Nile virus, a similar number of adults; and of course influenza, which kills thousands even in years when there is no epidemic. One way to look at this is to calculate the ratio of stories to deaths. It’s about 6 million times higher for Ebola than for influenza.
Ebola appeals to our fears: the disease is grisly. It is a serious threat with tremendous public health implications – in western Africa. That is the reason for sending relief workers overseas – fighting it there so we don’t have to fight it over here. Unfortunately, popular intuitions about it are often wrong. Many people seem unaware that asymptomatic individuals are not contagious, and the disease is not transmitted by airborne means. It is unfortunate that more coverage does not focus on evidence-based information such as this New England Journal of Medicine editorial.
@SamWangPhD you are more likely to be a living member of The Beatles than to have died of Ebola here.
— P M (@OrdoSeclorum) October 27, 2014