2006 Election Analysis
Nihilist provides us an election redux which appears to show being a RINO (Republican In Name Only) was a death sentence in this election. Nihilist gives up plenty of numbers and this conclusion:
Governor Pawlenty has brought the Republicans considerably to the left, fiscally, over the last two years with poor electoral results. Although several strong fiscal conservatives like Representative Phil Krinkie and Senator Brian LeClair were defeated, on the whole, fiscal conservatives were better off than their RINO colleagues. After all, why vote for a RINO when there is a real Democrat just down the ballot.
I don’t think Nihilist came to the right conclusion based on his numbers which I’ll address in a bit. What I found more fascinating were the comment trolls which popped up screaming “correlation doesn’t prove causality.” It is a valid criticism, but that doesn’t mean Nihilist’s numbers are meaningless, nor does it mean all correlations are meaningless. To prove causality you typically need random assignment. Sometimes you can’t do this. Think smoking, it would be cruel to randomly assign smoking to some people (force it upon them) and not others. For the longest time the only smoking/lung cancer data we had was correlative. Even without random assignment we know smoking causes lung cancer thanks to an endless stream of correlative studies. The same goes for physics. We know the equations for relativity match up with and predict the movements of celestial bodies. They correlate perfectly, but we can’t randomly assign planets to prove relativity. Does this make the Theory of Relativity wrong? Do we not know smoking causes lung cancer?
Correlative data can be very useful. Sure, it doesn’t prove causality but if the data can be used to predict events and there’s a lot of data points we can assume safely (or it can imply) causality. One can’t simply criticize numbers soley because they are correlative. Empirical data is often only available in correlations; try genetic inheritance. We know human intelligence and a bunch of other factors are inherited and thus genetic based solely on correlations.
The trolls that immediately popped up and started screaming “cum hoc ergo propter hoc” are always present when a conservative posits an idea based on real world data but they’re never available for discussions about Global Warming (only based on very slim correlative data) or second hand smoke (very slim empirical data). Knowledge of logical fallacies is a good thing to have but they are philosophical in nature. Knowledge of statistics will serve people better but is much harder. Give me enough data points of correlative data from real world events and I predict and express a truth about the world; give me a study with random assignment about human behaviour in a lab and you give me almost nothing useful unless humans happen to live in a lab.
Screaming about logical fallacies and avoiding the subject completely is a way of avoiding a valuable discussion. The trolls over at Nihilist should be ashamed.
As for Nihilist’s conclusion, I think he’s overstepped. Here is how I read his data:
1)In the state of Minnesota in 2006 a legislative move to the center did not correlate with electoral success and in fact had a negative correlation. Moving to the center did not provide a viable electoral defense to incumbants who practiced it in Minnesota in 2006. More data is needed to come to further conclusions, especially recommended comparing only swing districts to each other. This could also be expanded to include national races.
2) Being liberal in liberal districts and conservative in conservative districts resulted in the best electoral results for incumbants. Democrats in the bottom half of the Taxpayer’s League ratings and Republicans in the top half had the best electoral results.
Matter of fact, I think I’m a little overreaching in my conclusions since I make some assumptions about “swing” and “non-swing” districts. A good and careful look at the data might provide an answer to the question “do you appeal to the base or to the middle” but there isn’t enough here to give a solid answer. It is completely possible that going to the middle provided a slim defense to defeat but that effect was drowned out by a giant blue wave. The reverse is also possibly true, that people voted in a real Democrat to replace their RINOs. We just don’t know. I think one thing I would say is that there are probably external factors more likely to lead to electoral defeat (say a rising anti-war tide) and that they have a bigger effect than factors like legislative records. Basically, sometimes you’re gonna lose and there’s nothing you can do about it. In those years, you might as well stick to your principles than try to sell out and whore yourselves to moderate voters.
It’s just really hard to predict when to stay pat and when to whore yourself. I say err on principles. In Minnesota in 2006 that would have been the best option. I think a large study of swing districts over the last 30 years or so (using available American Conservative Union data) would add insight into whether appealing to the base or going to the middle would help. No, I’m not doing the study. At least, not until I’m a paid something or other.