In my day job, I focus on two subjects: strategic behavioral science and analytics. Really smart people do most of the heavy analytics work. I deal with the human behavior (neuroeconomics, psychology, etc.). In our narrow field of loyalty and motivation, I think we’re the best in the business. Not me, but the people I work with. For example, one client asked us to develop a strategy for overhauling their loyalty program. The result was nearly a 20 percent increase in sales YOY against market growth of 8 percent.
Before I moved into strategy, I was a software architect. The two fields, strategy and coding, are more alike than you might imagine.
Both fields help you develop a sense for patterns and anomalies. There might be a name for this sense, but I don’t know what it is. But I know it when I feel it.
After Tuesday’s primary, I felt it, so I’ve been crunching numbers.
St. Louis County Primary Numbers Were Weird
As I said, I’m not the hard numbers guy, but I have taken a few stats classes. And I use some statistical analysis every day. When I think a number looks goofy, I do is a quick analysis (ANOVA, chi square, whatever it takes) to see if the differences are really freaky or if it’s just my imagination.
Last night, I took another look at voting in St. Louis County from 2006 to 2014. I looked only at primaries. Here’s what I found: they’re really freaky. Especially if you’re a County Republican.
Democrat ballots increased steadily from 2006 to 2012. But Democrat ballots in 2014 were insanely high. In fact, they were nearly two standard deviations over 2006 to 2014 period.
That should concern the County Republicans, but not as much as Republican trends
Republican ballots were more than one standard deviation high in 2010 and 2012. In 2014, they were not ridiculously low; they reverted to the mean.
If you notice, Republican ballots were more than one standard deviation high in both 2010 and 2012.
I’m going to speculate that Tea Party activism contributed some of that bump in 2010 and 2012. Face it, we haven’t been nearly as visible in 2014 as we were in those years.
The GOP’s real problem with the Tea Party (and will free market conservatives in general) goes deeper than fewer grassroots activist demonstrations. I’ll get to that shortly.
Lee Presser’s excellent guest post on Hennessy’s View asks similar, disturbing questions:
If Republican voters are not adrift, why did so many Republican voters who voted for Steve Stenger continue to vote for other Democrats? Asked more directly, why were Republican voters so unsure their County Executive candidate was going to win in November that they had to vote for Stenger to make sure Dooley would be gone.
So I’m not the only one asking these questions or noticing these trends in St. Louis County. So I kept digging.
Nate Silver Sees Something In the Number, Too
Even after I did the analysis, I was tempted to say, “nah, Hennessy; you’re seeing things,” and drop this until I could do more analysis. Then I read this article on Nate Silvers’ FiveThirtyEight blog. He begins:
One of the most universal lessons of sports prediction is that margins matter. An NFL team that wins a number of games by less than a touchdown might get banner headlines for its clutch performance. But a team’s record in close games is mostly just luck. A football team that thrives on winning close games is likely to see its luck revert to the mean and start losing its fair share of them. The same is true in baseball, basketball and most other sports.
Makes sense. And he’s right. Teams that rely on one freakish outlier statistic usually crash hard to the mean–in the Super Bowl or NFC Championship or Stanley Cup.
Politics is no different from sports. Statistics tell you something. They report on how humans perform in given situations. What I love about my day job is the combination of the two: numbers and people. If I could, I’d go back to school to study neuroeconomics.
Nate Silver tells us what Senatorial close calls mean for GOP incumbents.
Between 2004 and 2008, just four of 39 Republican senators running for renomination, or 10 percent of them, got less than 65 percent of the primary vote. This year, five of 10 have fallen below that threshold: not only Roberts, Cochran and McConnell, but alsoLindsey Graham of South Carolina and John Cornyn of Texas, who both benefited from running against divided fields.
In fact, the average share of the primary vote received by Republican incumbent senators so far this year is 73 percent. Not only is that lower than 2004 through 2008, when incumbents averaged 89 percent of the vote — it’s also lower than 2010 and 2012, the years when the tea party was supposedly in ascendancy, when GOP incumbents got an average of 78 percent.
Ascendancy and maturity are two different things.
A lot of Tea Partiers suffer from Republican Fatigue Syndrome. RFS is a disease spread through casual contact with the establishment. Republican Fatigue Sydrome explains both the Senate incumbent challenges and the St. Louis County primary numbers. I’ll give you that the County Executive race lifted the Democrat ballots above their eight-year trend line. But that doesn’t account for the sharp fall in GOP ballots this year.
(Before I forget, here’s the link to my Excel spreadsheet.)
For example, Lee Presser noticed that Ann Wagner get fewer votes in St. Louis County in 2014 than in 2012 And the Democrats got more:
Compare the numbers above  to the vote totals from the 2012 St. Louis County primary election. Then, four Republican primary candidates for Missouri’s 2nd Congressional district received 63,978 votes. Of that total, Mrs. Wagner received 42,573 votes (2106 more votes than she received in 2014). Four 2012 Democrat primary candidates received a total of 22,446. 2014 candidate Lieber received 48,742, which was 26,296 more votes than four 2012 Democrat candidates.
Remember, Republican ballots in 2012 were a full standard deviation above the eight-year trend in St. Louis County.
The Science of Persuasion Could Affect November Turnout
Psychologically, Tuesday’s turnout was really bad news for Republicans in St. Louis County. While voting is dynamic and events can alter trajectories, two psychological factors will help Democrats in November: consistency and social proof.
The easiest way to get people to take a big step, like voting in a general election, is to get them to take a small step, like putting a yard sign up or donating (even $1), or voting in a primary. As the charts above indicate, Democrat primary votes in St. Louis County have been steadily increasing for a decade. Tuesday was a step change. People who pulled a Democrat ballot on Tuesday are very likely to vote Democrat again in November. That’s the consistency part.
Studies show that Democrats tend to hang around with other Democrats and Republicans hang around with Republicans. Plus, the number one predictor of whether someone actually votes is if someone in their network votes. And that effect extends to people two-degrees of separation from the voter, as I mentioned in my pre-election blog:
Even when we control for alternative sources of similar behavior, such as having the same income, education, ideology, or level of political interest, the typical subject is about 15 percent more likely to vote if one of his discussion partners votes. But does this influence spread beyond that to the rest of the network? As it turns out, we see a correlation between people who are directly connected and also between people who are indirectly connected via a common friend. In other words, if you vote, then it increases the likelihood that your friends’ friends vote as well.
Christakis, Nicholas A.; Fowler, James H. (2009-09-09). Connected: The Surprising Power of Our Social Networks and How They Shape Our Lives (p. 185). Hachette Book Group. Kindle Edition.
Those 130,000 people who voted Democrat on Tuesday likely have 5 close friends. Each of those friends has 5 friends. Three from each circle is likely to vote, in part because of the person who voted in August. You do the math.
Moreover, the number of registered voters has declined over that period while the number of Democrat voters has risen.
I haven’t done all the math, but it appears that the loss of registrations in St. Louis County comes disproportionately from Republican voters. That aligns to my post-2012 analysis that concluded Republicans are dying faster than Democrats, and the GOP isn’t replacing those deceased voters. I also discussed the problem here.
So the GOP has a lot of work to do before November. It needs to reignite traditional Republican voters, overcome a growing Democrat advantage in the County, and find a cure for Republican Fatigue Syndrome among free market conservatives.
Or maybe things are right where the Establishment want them—the Tea Partiers returning silently to their pre-2009 lives or moving over to the Libertarian Party. If you wonder, “why would Tea Partiers abandon Republicans,” look no further than the race-baiting fiasco in Mississippi that has now embroiled Erick Erickson of RedState.
If that’s the way the Establishment wanted things, all I can say is “Mission accomplished.”