Does political science matter? On Sunday, Ezra Klein, one of the rare journalists who seems genuinely interested in what political scientists have to say, wrote a column distilling some key lessons from political science, all of which revolve around the fact that most of what politicians do doesn’t actually make much of a difference, at least in the face of broad underlying forces: Presidential speeches don’t matter. Elections are determined by underlying economic conditions. Lobbyists aren’t as important as people think.

Over at the Monkey Cage, John Sides, who is also quoted in the article, argues that this lack of control should be good news to politicians: “Political science really does empower politicians. It tells them to ignore a lot of gossip and trivia. It tells them not to sweat every rhetorical turn of phrase.”

I must admit, I’ve been asking myself this question of whether political science matters ever since I started a Ph.D. program six years ago (and completed it a month ago). And I have a few quibbles with the conclusions that both Klein and Sides draw, and a few warnings for politicians and journalists lest they over-extrapolate from the received political science wisdom.

Those who consume political science research need to understand that political scientists are largely interested in finding patterns and proving that two variables are correlated (which is a necessary but not sufficient condition for proving causation, the holy grail of political science). But patterns tend to be rough patterns, and there are always other forces at work.

Figure 1 shows a typical XY scatterplot with a regression line, the most basic graphical tool of the social scientist.  While journalists tend to be interested in the dots (i.e., the real-life cases), political scientists are interested primarily in the regression line (i.e., the underlying relationship). Both are important, and ignoring one at the expense of the other inevitably leads to a limited perspective.

Figure 1: A Typical XY Scatterplot

Also notice: While there is clearly a relationship between X and Y variables very few dots (the actual real-life cases) fall directly on the line, meaning other factors are at work.  Of course, very little social science takes place at the bi-variate (two variable) level. Most published models explain outcomes with multiple factors, which account for more of the deviations from the predicted values (i.e., the regression line).

But even the best models still are often unable to explain half of the deviations.  What this means is that while large structural forces do drive political outcomes, there is also almost always room for idiosyncratic forces to operate as well, and they can often be decisive.

Sometimes candidates win victories they shouldn’t because of opponent gaffes. Sometimes lobbyists do get what they want, and make their clients very rich in the process. Sometimes presidential speeches do make a difference.  Political scientists are interested in the general case, and are fond of couching their findings in the cautionary language of ceteris paribus (all else being equal). But all else is rarely equal.

The thing is, it’s very hard for political actors to know when something they are going to do will make a difference. Often the impact hinges on unpredictable timing and unanticipated resonances. Anyone who has spent time in politics knows that you never know when something is going to “pop.”  So it’s almost always worth gambling because they pay-off could be big.

For a lobbyist, for example, even if on average you won’t win, when you do win, you might just win big enough to make it all worthwhile. A scoring system that treats all wins and losses equally ignores the fact that some lobbying wins are very big wins indeed, wins that far make up for a long strong of losses. For example, getting Medicare prescription drug coverage was a huge, huge win for pharmaceutical companies, surely worth many losses.

Additionally, one of the reasons why many political actions may seem to NOT matter is because political actors on both sides think that they DO matter. Consider the empirical conclusion that presidential speeches do not move public opinion. One reason that public opinion is unlikely to move is because there is always an opposition to respond, and so citizens who are skeptical of what the president has to say can easily latch on to messages that are critical. Thus they remain unmoved.

But say conservatives suddenly followed this wisdom, decided that speeches didn’t matter, and therefore didn’t bother to respond to anything Obama said. My guess is that Obama’s unrefuted speeches would start having more of an impact. Or alternately, say Obama decided speeches didn’t matter, and stopped giving them. He’d be leaving it to the conservative opposition to define him, without speaking up for himself and giving voice to his supporters. My guess is this would also have an impact on public opinion.

Likewise with lobbying or campaigning. These things exist in a kind of equilibrium. Given a rough balance of power, actions on each side are countered with actions on the other side, and hence drained of their perceived impact. Lobbyists on one side respond to lobbyists on the other side, thus neutralizing their efforts (this seemed to me one of the most important points of “Lobbying and Policy Change: Who Wins, Who Loses, and Why?”) Campaign contributions on one side are often matched by campaign contributions on the other side.

But were one side to decide outcomes were beyond their control and stop sweating, my guess is they’d very quickly prove the conventional political science wisdom wrong.

In short: there is a lot that political science can teach both politicians and journalists about underlying structural forces and general patterns that drive political outcomes. But politics takes place in specific cases, not general cases. Journalists and politicians ought to have a better appreciation for these general patterns, which will provide context for specific cases.

But politicians better not take political science wisdom too much to heart, lest they undermine it by upsetting the equilibrium that exists when they believe what they do on a daily basis actually DOES matter.