We Live in a (Multivariate, Probabilistic) Political World

I said something earlier today that I should talk a bit more about: when it comes to many things in politics, and particularly elections, it's important to remember that multiple factors are often in play, and that many things are best thought of in terms of probabilities, rather than clear, certain, cause-and-effect relationships. 

The first part of this means that we should never assume one result proves or doesn't prove that some explanatory factor did or didn't matter...it's possible that the losing candidate had the better campaign and was even in some sense the strongest candidate, but that other factors (partisanship, economic effects) overwhelmed very real, but smaller, campaign effects.  In other words, that Barack Obama beat John McCain doesn't mean that Obama (necessarily) ran the better campaign. 

The second part means that even when we know the magnitude and direction of some effect, there's often going to be some variation that can be treated as random, and not really explained by anything systematic. This one is a lot murkier than multiple causes...in some cases, such as explaining the votes of Senators on confirming a Supreme Court Justice, it's in my opinion not entirely easy to explain what it means to say that, say, something might change the chances of Susan Collins supporting a nominee from 52% to 58%.  Nevertheless, we can accurately say that there are plenty of things in politics that act as if they were consistent with the rules of probability.  Which also means that we can say that there are random unexplainable events -- what we might call luck (such as, for example, one party winning all of the close Senate races, as the Republicans famously did in 1980). 

A good example of what happens when a (good!) reporter gets this stuff wrong is in Dave Weigel's comments about birthers and truthers today in the wake of Nathan Deal's victory in yesterday's Republican primary in Georgia.  Weigel notes that Deal's win, compared to "truther" Cynthia McKinney's two primary losses, shows that "Dipping a toe into the 9/11 fever swamp cost McKinney her job. Dipping a toe into the birtherism fever swamp didn't stop Deal from winning a statewide primary."  Unfortunately, Weigel is pushing his evidence far beyond what it can support.  Did truther issues cost McKinney her job?  I don't know.  We know she was defeated in a high-profile 2002 primary, won her seat back in 2004, and then was defeated in another primary in 2006.  If I recall correctly, she had in 2002 won the opposition of national Jewish organizations for her views on Israel; I know that an incident with a Capitol security guard was at least reported to be an issue during her 2006 campaign.  I have no idea whether Cynthia McKinney had well-run constituency services, or the quality of her opponents (and what prompted them to run), or what other issues were prominent in those campaigns, or what her relationships were with important groups within the district's Democratic party network.  We would need to know all of those things to be able to conclude that it was her truther comments and actions that "cost KcKinney her job."

Now, we do have a bit of evidence for one, much weaker, claim: certainly Deal's (limited) birther comments have not disqualified him from Republican electoral politics.  Did it help?  Hurt?  We don't know.  For all we know, Sarah Palin (whose endorsement against Deal seemed to be important in the race, although again I don't know for a fact that part is true, either) may have specifically opposed Deal because of the birther stuff, and that might (might!) have almost sunk him.  Perhaps Deal only shows that one can go as far as he did (raising questions, nod-and-wink stuff), but no farther.  The runoff was very, very close -- would Weigel have said basically the same thing had a handful of voters gone the other way?  If not, he's almost certainly putting far more weight on the win/lose difference than it can really handle.

The lesson from all this: sometimes, it's just best to admit we don't know, or to scale down our claims to match what we actually do know.