Update on state size and governors’ popularity

Nick Obradovich saw our graphs and regressions showing that the most popular governors tended to come from small states and suggested looking at unemployment rates. (I’d used change in per-capita income as my economic predictor, following the usual practice in political science.)

Here’s the graph that got things started:

Governors Approval Rating 2009.png

And here’s what Obradovich wrote:

It seems that average unemployment rate is more strongly negatively correlated with positive governor approval ratings than is population. The unemployment rate and state size is positively correlated.

Anyway, when I include state unemployment rate in the regressions, it pulls the significance away from state population. I do economic data work much of the day, so when I read your post this morning and looked at your charts, state unemployment rates jumped right out at me as a potential confound.

I passed this suggestion on to Hanfei, who ran some regressions:

lm (popularity ~ c.log.statepop + c.unemployment)
coef.est coef.se
(Intercept) 48.57 2.07
c.log.statepop -3.98 2.42
c.unemployment -2.95 1.20

n = 49, k = 3
residual sd = 14.46, R-Squared = 0.27

lm (popularity ~ c.log.statepop * c.unemployment)
coef.est coef.se
(Intercept) 47.50 2.31
c.log.statepop -3.90 2.42
c.unemployment -2.79 1.20
c.log.statepop:c.unemployment 1.09 1.06

n = 49, k = 4
residual sd = 14.45, R-Squared = 0.29

lm(formula = popularity ~ c.log.statepop + c.income.change + c.unemployment)
coef.est coef.se
(Intercept) 48.57 2.05
c.log.statepop -3.47 2.44
c.income.change 1.97 1.62
c.unemployment -2.76 1.20

n = 49, k = 4
residual sd = 14.38, R-Squared = 0.29

Also some other versions including lots of non-statistically-significant interactions. The “c” in the variables “c.log.statepop”, “c.unemployment”, etc., represent centering. We centered each predictor to allow easy interpretation of each main effect in the presence of interactions.

The punch line is that, in these models, state-level unemployment is highly predictive of lower popularity for the governor. Even after controlling for these economic predictors, governors of smaller states remain more popular, but this trend is no longer statistically significant.

The next step is to look at other years and other statewide offices. I think there are some articles on this in the political science literature that Shigeo found when we last looked at the topic.

In the meantime, I’ll take the position that the most popular governors tend to be from smaller states, and some of this pattern appears to be explained by economic factors.

5 thoughts on “Update on state size and governors’ popularity

  1. In the U.S. smaller states have relatively more representation in the Senate, and so tend to get disproportionate shares of federal funds, no?

  2. May anyone post code in their favourite language without saying which it is? Or is this a privilege reserved for R users?

  3. Carl: Yes.

    Nick: Any language is ok. But if you post regression results in R, you have to use coefplot() or display(), never the default print(), plot(), or summary().

Comments are closed.