John Sides had an interesting piece in the New York Times’ FiveThirtyEight blog earlier this month. The full piece is well worth a read for anyone serious about understanding the latest political science research on campaign advertising. Here are links to part one and part two.
Sides is a widely published political scientist and the co-founder of one of my daily must-read blogs, The Monkey Cage, so I was excited to see his extended take on campaign advertising.
Covering a wide swath of the political science literature on the topic, I’m not going to try to cover everything Sides said in this piece. Instead I want to re-focus the discussion a bit on his original metaphor.
Sides carries the “Moneyball” metaphor only far enough to make fun of the “grizzled advisors” before launching into a fairly mundane re-capitulation of the state of the political science literature on campaign advertising.
There’s nothing wrong with a lit review, and having all of the information in one place is certainly valuable. But, it would have been more interesting to me as a practitioner of campaign research and analysis to see a direct assessment of the “Moneyball” nuggets hiding in the information he presented. I think there are a few such tidbits in his article.
First though it’s probably worth talking about “Moneyball” for those who aren’t familiar with the concept. “Moneyball” is the book, and later movie, about Billy Beane’s use of the findings of a branch of baseball research generally referred to as “sabermetrics” (SABR being the Society for American Baseball Research) as general manager of the Oakland Athletics.
Being a book, and movie, “Moneyball” almost has to tell the story as one about the conflict between Beane and the traditional baseball scouts and “grizzled advisors.” Sides chooses in his articles to mostly mirror that structure and talks a lot about how political science has proven typical political analysis wrong. That’s fair enough and he certainly has lots of opportunity to point out the failings of both pundits and campaign professionals.
But there’s another story in “Moneyball” that I think exists in the literature Sides reviews and in other science out there. What “Moneyball” is really about is the benefit of being an early-adopter of new information in your field. Billy Beane benefited from knowing that there was an undervalued resource out there—high on-base percentage power hitters who weren’t particularly athletic or good contact hitters—and taking advantage of it.
Interestingly, the story doesn’t stay the same forever. Players who combine high on-base percentage with the ability to hit for power are now hot commodities. Teams that are taking the same approach of using analysis to find undervalued resources are now focusing on defense.
The “Moneyball” approach isn’t just about baseball either, Daryl Morey of the Houston Rockets has used analytic tools to identify players who benefit their team by being on the floor despite not recording any real success in traditional statistical categories. And businesses of all types and sizes are increasingly using data and analytics to gain additional advantages and increase profits.
So what is the “Moneyball” of politics? Some of it is present in Sides’ article, but I also think he misses the point a little.
You see, “Moneyball” isn’t the macro effect that political scientists are usually, rightly, interested in. Billy Beane didn’t change the big picture of baseball. As a sage observer once said, “This is a very simple game. You throw the ball, you catch the ball, you hit the ball.” That’s as true for Billy Beane’s Oakland teams as it is for all of the others.
The “Moneyball” approach is about finding marginal advantages by using data, research, and analytics to understand the problem you’re facing better than the next guy. That’s how a Billy Beane uncovers the Giambi’s of the world and how Daryl Morey realizes the value of a Shane Battier.
So let’s look at some nuggets from the Sides article that might inform an actual “Moneyball” approach to politics.
Here are four findings that I thought were particularly relevant to someone, like me, whose job it is to inject a little “Moneyball” into the campaigns we work with at WPA.
- Campaign ads matter more when the candidates are unfamiliar.
While most political science research is about big races with well-known candidates (an overwhelming amount is about Presidential politics), this one finding is probably most important to those of us who work with candidates for Congress, down ballot statewide (e.g., Lieutenant Governor, etc.) and even state legislature.
This finding may suggest an increased value for early advertising in races where the candidates aren’t familiar because the marginal value of those dollars is higher during that initial introduction phase.
- Campaign ads matter more when a candidate can outspend the opponent.
While this is fairly obvious, the underlying information and analysis suggests some tantalizing things about optimizing campaign spending across different media markets and at different times.
- Campaign ads can matter, but not for long.
The “priming effect” of campaign advertising is something we’ve written about in the past, but it’s worth pointing out again. Thinking about the topic again does lead to some interesting questions about the value of hoarding resources for a late ad push. It makes one wonder whether the traditional aversion to “going dark” once advertising has started is really that important.
I also wonder how this finding would play out in the context of a campaign where the candidate was unfamiliar at the start—maybe it suggests two peaks in the campaign advertising cycle and less concern about what happens in the middle.
- Negative ads work, except when they don’t./There is no secret sauce. Really.
There’s a risk in taking these last two points together to arrive at some kind of “know-nothing” philosophy of actual campaign activity. Sides unfortunately edges toward this position to end his article.
The same information could also suggest that context matters a lot more than we typically think. There’s no “secret sauce” because you can’t always do the same thing in race after race and expect it to be successful. Negative messaging works in the right context and when you have the right message. Sometimes there’s nothing negative that will work.
Our own research using field experiments and random assignment of messages has demonstrated pretty clearly that, race-by-race, there are messages that work better than others (while this may seem basic, it’s amazing how often a message du jour will become the standard cookie cutter approach in races across the country.
But it’s not always the same message. And sometimes it’s a positive message while others it’s a negative one. And we often see different messages have very different impact by media market.
To us, not surprisingly, this all illustrates the importance of a good campaign researcher. A researcher who does more than just measure the race with antiquated more likely/less likely message testing. A researcher with the right message tools comes in: helping the candidate find the right message for each specific race, by market, by electorate, and be each specific moment in time.
And maybe that’s the “Moneyball” of politics.