Recall the counter intuitive result that you can explain the vast majority of all voting behavior with only two metrics, i.e. the vote view model.  The two metrics range liberal to conservative and one is economic, one is social.  There is a mystery, at least to me, about these two metrics.  You don’t really need the social metric to explain most congressional votes.  Yes, the accuracy of the model improves some by adding the social metric; but not tremendously.  So why do we expend so much time ranting about the social dimension?  It’s a disconnect between the talk and the action.

These charts show an average adult (sic) for these two metrics; one point per state.  This data is new to me, I’ve only seen two dimensional vote-view scatter plots for individual legislators.  It would be fun to have error bounds on the points, it would be fun to see the spreads between the voters and the population at large, it would be fun to have dots scaled to electoral votes.

One way to frame up the mystery is to say that we talk about social issues, but legislators act on economic ones to such a degree that we have the same disconnect seen between of signal-value use-value for in sexy but unreliable sports cars.  The sexy bits sell the product, but the actual product delivers something else.  That kind of thing isnt’ unusual, i.e. say/do ~ signal-value/use-value ~ social-values/economic-values.

No doubt there is a framework that explains when the signal-value and use-value of a product tend to diverge.  For example when the feedback loop delivering information about the use value is broken, or when the buy decision is emotional and impulsive.  I bet there is something here that is interesting to be dug out about political activism.

It’s interesting to see charts showing these metrics for the buyer side of the electoral process.