Don’t confuse the map with the territory

In a recent post on the Monkey Cage, under the heading, “How personality explains behavior in Congress,” Adam Ramey wrote:

In a new paper, my colleagues (Jonathan Klingler and Gary Hollibaugh) and I argue that personality and ideology work together to shape how legislators make decisions, voting and otherwise. Using every floor speech by every member of the U.S. Congress since 1996, we use some recent methods in computer science to generate the first estimates of legislator personality over time. For each member, we estimate their positions on the Big Five personality dimensions — Openness to new experiences, Conscientiousness, Extraversion, Agreeableness, and Neuroticism (or OCEAN, for short). We find that each of these dimensions helps us to explain lots of different behaviors that legislators engage in, even after accounting for ideology.

So far, so good. I clicked through to the paper and found the actual words that they coded:

The eight mini-markers for Openness to Experience are creative, imaginative, philosophical, intellectual, complex, deep, uncreative, and unintellectual.

The eight mini-markers for Conscientiousness are organized, efficient, systematic, practical, disorganized, sloppy, inefficient, and careless.

The eight mini-markers for Extraversion are talkative, extraverted, bold, energetic, shy, quiet, bashful, and withdrawn.

The eight mini-markers for Agreeableness are sympathetic, warm, kind, cooperative, cold, unsympathetic, rude, and harsh.

Finally, the eight mini-markers for Neuroticism are moody, jealous, temperamental, envious, touchy, fretful, unenvious, and relaxed.

I think they’re counting the number of each sort of word in each Congress member’s speeches.

Okay, I like it so far. The method is transparent, and if you think you can do better, you can start from the choices Ramey, Klingler and Hollibaugh made and go from there.

But I do object to how Ramey continues:

Most of the dysfunction we observe in Congress today is a product of both ideology and changing personality demographics…. [P]ersonality dimensions explain the kinds of bills legislators propose, how often they buck the party line, how they use press releases and Twitter to disseminate information, and more…. [P]ersonality is more than a feeling. It’s a driving force behind partisan, polarizing tactics and it’s reshaping how Washington works.

They don’t have direct information on personality; they have counts of words in speeches. That’s fine — I think this is fascinating research — I’d just prefer to keep the measurement front and center.

Instead of “personality is … a driving force,” I’d prefer they write something like, “Counts of characteristic words are consistently correlated with….” It’s not as snappy, but it’s more accurate. And it’s just fine: what they found is interesting enough on its own without needing any descriptive shortcuts.

Again, I don’t have a problem with the research; I’d just prefer more precise descriptions of what’s been found.

Andrew Gelman is a professor of statistics and political science at Columbia University. His books include Bayesian Data Analysis; Teaching Statistics: A Bag of Tricks; and Red State, Blue State, Rich State, Poor State: Why Americans Vote the Way They Do.
Comments
Show Comments
Most Read Politics
Next Story
Akos Lada and Maria Snegovaya · April 15