Yesterday, I linked to a piece in the National Review that advanced the following point:
Northeastern states, which make up the geographical heart of liberalism, are considerably wealthier than the upstart Sunbelt states, and that
by John Cole| 26 Comments
This post is in: Open Threads
Yesterday, I linked to a piece in the National Review that advanced the following point:
Northeastern states, which make up the geographical heart of liberalism, are considerably wealthier than the upstart Sunbelt states, and that
Comments are closed.
Aaron
People tend to vote on the basis of their situation. But it can cut so many ways it’s hard to tell.
Mainly, I’d be interested to see the percent of government employees that vote Dem.
My private theory is that those LEAST at risk of income swings vote Dem. If I am salaried goverment employee, I’m not too worried about higher taxes. Either is a doctor really.
If I own a small business, some of the tax issues are make or break and the income is just not guaranteed.
It might even be the same for employees who understand that this could affect them as well in the long run.
The Jim Dandy
Any kind of single issue analysis like this is wrong, if for no other reason than they focus on one aspect at the expense of the thousands of others. The bottom line is Republicans are better salesmen than Democrats. That’s why they win more, and that’s all-encompassing enough to be adequate.
Darrell
Also, Blue states with larger metropolitan concentrations, typically will have higher incomes because the cost of living is so high in the cities and because of the high tax/union strength, you will see people paid way more than they could hope to make anywhere else. I remember when I lived in the San Francisco Bay area in the late 1990’s, it came out that there were some BART transit employees working behind the token booths being paid over $110,000/year.
good post John.. you were correct not to trust 1 quarter of economic data to “prove” anything
Sojourner
Actually, the data I’ve seen show the opposite. Single women are more likely to vote Dem because they feel most at risk. The government employees I know are largely Republican, which is admittedly ironic given their stance against big government. Nor does that prevent them from taking advantage of Democratic policies such as using the unions to deal with employment issues all the while looking down their noses at those who utilize government services.
Rick
As the transgessor who forwarded the link for discussion, I must admit I fail to see the point about Washington State. As I read it, that state’s income statistics, the outlier, does much to proper up what is already a trailing figure.
The 1Q-only “analysis” is correct, but a .3% growth differential would, if it could be extrapolated, compounds greatly.
So there is a slight edge to “Red States” over the posted snapshot, and it accords with the usual reporting on GOP strengths–the growing regions as the population shifts south and west (to some Blue benefit on the left coast, but GOP from the Redwood Forests to the Gulf Stream waters).
COrdially…
John Cole
No, it doesn’t.
It was -7.68% this quarter, dragging down all red states. The formatting of the post makes it look like that is a +7.68% for the quarter.
John Cole
Ahem- dragging down all blue states…
Rick
But, but, it also boosting the 2004 4Q results even more “outlier”-ly fashion than it depresses the ’05 1Q. And surely you mean “blue states.” And I’ll stop calling you Shirley.
Sticking with the thrust–the people, the money and the income growth are flowing S to W in this great land.
Cordially…
MC
John and Rick:
For reasons I explained in the earlier post, percentage of increase/decrease can only be used to explain trends relative to itself.
California’s quarterly personal income is 1.3 trillion. This is larger than the plains and the rocky mountain region combined (as classified by the BEA). There is only one red state that even comes close – Tezas, with QPI of roughly 750 billion. That’s less than 75% of California’s total. Combine California, New York, Pennsylvania, and New Jersey, and Illinois, and that’s over 1/3 of the entire nation’s quarterly income. Only Texas, Florida and Georgia even come close to these states (surprisingly, not Arizona). Population shift in this country has been going on since the early 1960s, so it’s only logical that some sort of multicollinearity would occur with income growth.
A .2% in California is equivalent to a 1% increase in Mississippi in terms of actual value, so even though it looks like Mississippi is growing the economy, they’re starting from so far back, it’s really not all that significant. There would have to be orders of magnitude difference for a sustained period of time before it would even begin to make a difference. When growth rates are close, in terms of real figures, it’s astronomically higher in large blue states.
The other thing, and why you should probably use a median instead of a mean, is that either group can benefit significantly by large growth, even though it doesn’t really mean anything. For instance, if Mississippi (again, I’m picking on them), grows three percent and Texas grows one, that counts as two percent average growth. If PA and CA both grow two percent, the average is the same. Texas was brought up by the smaller real growth in MS, and as I stated earlier, in real numbers, the PA/CA growth is more significant
MC
Sorry about the double-post…actually, don’t use a median, but determine real growth by summing all income figures and increases which would properly weight the percentage increases derived…median is just plain lazy…
Mr. Kahn
Lies, damn lies and statistics.
Everyone should read a little bit of Edward Tufte.
http://www.edwardtufte.com/tufte/books_vdqi
He does an excellent job of explaining how charts often lie.
Mr. Kahn
Lies, damn lies and statistics.
Everyone should read a little bit of Edward Tufte.
http://www.edwardtufte.com/tufte/books_vdqi
He does an excellent job of explaining how charts often lie.
neil
Sticking with the thrust, even when it has been proven to be premised on a lie. That’s the Bush/Rove way if I’ve ever seen it.
In Q4 2004, the average personal income growth of all states besides Washington was 3.044%. In Q1 2005, the average of all states besides Washington was 0.948%.
In Q4 2004, Washington was an outlier by 8.27%. In Q1 2005, Washington was an outlier by 8.457%. In other words, Washington affected the 2005 Q1 results in a more “outlier”-ly fashion than the 2004 4Q.
But of course, nobody could outlie you, Rick…
Rick
Not even my Master, Karl Rove?
Cordially…
Rick
In essence, it was a twofer, attempting to dispel the ‘myth’ that Republicans are the party of the rich, and to show that there is evidence that in reality, the GOP was the party of those aspiring to better their economic positions.
And pretty much a success. The Democrats have an entirely healthy proportion of the millionnaire Senators and tycoons you can think of (Richard Mellon Scaife being a singularly notorious Republican malefactor of great wealth).
If crude divisions are sought, then the Democrats are the party of the rich, and the “little guy” (poor, or working poor), while the GOP is the party of the middle class. Which is where the real money is.
Cordially…
Kimmitt
Look, if you examine Cole’s bottom graph, it’s pretty clear that the piece was written as an attempt to mislead. At that point, there are no ideas to address; the writer has forfeited the privilege of my consideration.
I mean, Rick was convinced, but we’ve already established that the verity of a claim is irrelevant to Rick’s embrace of it.
Rick
At that point, there are no ideas to address; the writer has forfeited the privilege of my consideration.
I shall try to contact the miscreant, and sugarcoast this devastating news. It’s another case of TANG and damning DSM “documents,” that’s for sure.
Cordially…
Rick
I’ve sent a couple emails to individuals making assertions purporting to dispose the NRO guys argument. These emails ask for restatement of the supposed demolition. I haven’t yet heard back, which proves nada.
But just to show that I can learn something here, I’ll venture the guess that in re: Washington State fluctuations (affecting little of the national or Red/Blue picture), our host HISOWNSELF fell victim to the nefarious, left wing apologists’ “Chewbacca Defense.”
Cordially…
Bernard Yomtov
Rick,
Give it up. It’s been demonstrated here, beyond any doubt, that this is a stupid and dishonest article.
Did you try to understand the objections? Did you look at the exit polls?
Rick
It’s been demonstrated here, beyond any doubt, that this is a stupid and dishonest article.
Bernard,
I’ll give it up when it IS demonstrated, beyond any doubt, that this is a stupid and dishonest–call it Krugmanesque–article. It’s a partisan article, or else it wouldn’t have run on NRO. And data can be sliced and picked in numerous ways.
But in the scope of what was examined, who has laid what glove on it? As I suggest above, John’s Chewbacca Defense was thrown out, and victory declare. I’m just saying “huh?”
So there’s your opening to drive through, Bernard.
Cordially…
Bernard Yomtov
Who’s laid a glove on it?
Just about everyone who’s commented, and John too. I’m not going to repeat the criticisms. You haven’t refuted one yet.
You’re going to believe this article no matter what. Go ahead.
Rick
Bernard,
Criticism and comment are different from showing *error.* If you scrutinize Jphn’s official Green=Blue/Purple=Red big bar graph, it merely confirms the simpler, kind-of-Powerpoint graph at the top. That is, the 2005 1Q figure is confirmed.
Chewbacca defense, I say. You wish to join the home run trot, but everyone has just been popping back of home plate. John made the point that a mere quarterly consideration isn’t a good measure; not longitudinal enough. But his longer spanned chart shows a 2+ year trend of greater income expansion in Red states.
Yet it’s feigned here that it doesn’t, or shows the opposite. Sure, I could be wrong, but no one has made the case in the English language on this thread. The Washington State figures work *against* John’s statement.
At the risk of repeating myself: “huh?”
Cordially…
MC
Rick:
Everything the author said may or may not be true, we don’t know, because he demonstrated nothing. The author made two claims. In the first one, a correlation between personal income and state political affliation. Explain how that has been demonstrated, and even it has, what does that mean? There is a correlation between violent crime and ice cream sales and no causal link (it’s temperature – common response). We would need some sort of regression analysis to even begin to discuss causality. The same holds true in the second case – that being Republican results in higher personal income growth. Again, explain how the causal link has been demonstrated at the exclusion of confounding variables? Explain the signifigance of independent variables and the overall correctness of the model presented. This has nothing to do with ideology, it’s simple methodology and GLM. The author is working with multiple variables and used no statistical tools to demonstrate a relationship.
Forget about the metrics used and the interpretation, they’re open to debate. Forget about the aggregated figures. There is no link demonstrated even with the data the author used. Period. It’s junk statistics – they failed to use appropriate tools to demonstrate causality or even significance of correlation and used a predicative model with a single variable. It’s not wrong, it’s not right, it’s useless.
Rick
MC,
Thanks, but that same critique is directed equally at the supposed “demolition” or Cole’s “false picture.” Or Paul Krugman’s articles, as I noted in an earlier comment.
I didn’t read the article as claiming GOP=income growth, but rather that faster growing personal income was a hallmark or predictor of GOP tendencies. To the graphs, the author noted that something like 97 of the 100 fastest growing counties voted for Bush.
My big beef was John’s stating that the graphs painted a different picture (when in fact is confirmed it), then all the acolytes gathering round to acclaim the emporer’s finery.
Thanks agin, you certain do much better than most, who content themselves with lemming-like responses.
Cordially…
MC
John’s graphs neither confirm or deny findings or demonstrate causality or correlation. Neither did the authors original work. If you don’t want to accept John’s work as demonstrative of anything, that’s certainly valid, for the same exact reasons the work of Bowyer is discredited: Inappropriate use of statistical tools to demonstrate variable relationships and failure to provide evidence of model strength and variable significance. The same argument can be made about some of Krugman’s work, although I imagine his academic stuff is pretty rigourous. I’m not an economist, but I would guess his critics don’t find fault with his use of tools, they find fault with his models.
http://www.bepress.com/forum/vol3/iss1/art2/
This is what I would consider a good piece of statistical analysis with a conservative-favorable bent. It claims that Republicans, or more generally conservatives, are at a disadvantage for hiring or advancement at “elite” universities. Most people can understand the author’s findings or with a little effort get an understanding. One can make arguments about the author’s metric for elite and some of the other metrics, but they don’t misuse the metrics they chose, which Bowyer was guilty of, and they conduct smple multivariate analysis on a model to isolate their independent variables and attempt to demostrate causality and the strength of the model, they test to show variance is not random but actually significant, and they provide a measure of the effect of change in independent variables.
A serious question: are you just playing devil’s advocate to get a rise or do you really believe there isn’t a problem with Bowyer’s work? I’m being serious, because I don’t consider myself to be particularly adept at statistics and maybe I’m just way off-base and not considering something. My knowledge is simple univariate and multivariate analysis type-things, common stuff that undergrads in physical science or engineering (I guess econ too) pick up. Honestly, I’m a mathematical moron and maybe I’m making a wrong assumption about things.
Rick
A serious question: are you just playing devil’s advocate to get a rise or do you really believe there isn’t a problem with Bowyer’s work? I’m being serious, because I don’t consider myself to be particularly adept at statistics and maybe I’m just way off-base and not considering something.
MC,
Thanks for being so polite and all, and it’s obvious that everyone on this thread (including yours truly)–and John Cole–are playing an away game on your field.
As stated in several comments above, I have no problem with criticizing Bowyer’s article on the basis of narrow measurement, but in a–what–750 or 1000 word tour d’horizon, I wouldn’t expect detailed embroidery. So, as with Krugman, issue may be taken with method. Let alone even grander considerations of what motivated voting patterns and results. Bowyer suggests economic well-being or satisfaction; many colleagues here suggest brainwashing, snake handling and fraud. Oh, you should see THEIR metrics!
There’s a touted liberal book in opposition to Boyer’s quick take. And a very quick look at the reviews picks up some similar criticisms. And this is a book-length treatment, not a knock-off column.
My OCD outburst is based on the expressed, self-satisfied conclusion contained in the sweeping declaration that the Green-Purple graph “paints a much different picture.” I’m calling BS, because it paints the same picture. The Washington State detail is as interesting as all statistical outliers, but if it demolishes Bowyer by giving a “true picture,” I don’t see how, and my plaintive requests for a demonstration of how go ignored.
I suspect Bowyer’s article was motivated in part by post-election whining that the Blue states were richer, taller, smarter, better-looking and had more symphonies, so that just shows how wrong the Red states are. Two can play that game.
Thanks again, and I look forward to your reply. I guess we’re only reading each other by now, so my real email addy (as opposed to spycraft dead-drop accounts so many sleuths here adopt) is a click away.
Cordially…