The kerfuffle over the measles vaccine is off-topic for the blog. It is also ultra vires (not virus) for CJLF. I have watched the discussion with some interest, though, particularly the propensity of media outlets of a given orientation to ascribe anti-vaccine beliefs to the other side. Crunching some actual numbers was a fun project (in a nerdy sense of fun) for a soggy Saturday here on the Left Coast, and I found some interesting things about how political orientation relates to an essentially nonpolitical subject.
A few background points are needed before jumping in. First, too many people, and particularly those in "mainstream media," act as if everyone in America resides in one of two buckets labeled "liberal" and "conservative," and everybody in each bucket shares all the same views on political, social, and other issues. Also, each of these "sides" views the other as more monolithic than it really is. Cracks in coalitions are more apparent close up.
Even when the label-attachers recognize degrees of liberalism and conservativism and place people along a one-dimensional spectrum instead of two buckets, they are still working from a faulty model. A two-dimensional model with an individual v.collective axis along with a conservative v. liberal axis would come closer to the truth.
Now a bit about the measles flap. Vaccines are one of the great advances of modern medicine and have nearly wiped out many formerly dreaded and widely fatal diseases. Yet there are a few people who refuse to let their children be vaccinated. Some have individual medical reasons. Some have religious beliefs against any form of medical treatment. (Whether imposing such a belief on a child is proper parenting or child abuse is a separate controversy we won't get into here.) What I say below does not apply to the folks described in this paragraph.
Then there are people who are just against vaccines, generally because they think risks of vaccines outweigh benefits. At one point, there was a study in a peer-reviewed journal claiming a link between vaccines and autism, so there was a time when such a theory had some degree of respectability. That study has been completely debunked, and the fact it was ever published in the first place is part of a scandal that requires a reevaluation of peer review.
So at this point the anti-vaccine crowd is mostly crackpots, endangering the lives of their own children and others. Unvaccinated children who become infected spread the disease not only to the children of the other crackpots but also to those who are unvaccinated for the reasons noted above and to the few who contract a disease despite having been vaccinated. Anti-vaccine crackpotism is particularly strong in California, so it should be no surprise that if a measles outbreak were going to break out, Disneyland would be a prime candidate for it. And that is what happened.
Is there a political dimension to this? The New York Times thinks so. In "Measles Outbreak Proves Delicate Issue to G.O.P. Field," Jeremy Peters and Richard Perez-Pena write:
So how does antivaxx really relate to political orientation? I decided to find out. First, I had to find a readily available data source online. I found data for the Personal Belief Exemption for vaccination by county in California, and voting data for the last presidential election is also available by county. Now, county isn't the best unit for this study. Some of California's flakiest enclaves are buried within massive Los Angeles County. But the data are quick, free, and easy. You often see this in research. The ideal data set does not exist, and it would be difficult and expensive to gather it, so researchers make do with what has already been gathered. This goes in the "limitations" section of the article that few people read at all but everyone should read first.
So, we have the percentage of kids whose parents claim the PBE in the 2014-15 school year and the 2012 Obama-over-Romney margin (a negative number in counties "carried" by Romney) for each county in California. We fire up the trusty R statistics package to correlate the two and get r = -0.144 and p = 0.28. The p tells us to forget about the r because it is meaningless. Small p is good, and the standard rule of thumb is that a result with a p over 0.05 is not "statistically significant." We can argue about that rule of thumb, but 0.28 is outside the range of argument. There is nothing worthwhile here. A scatterplot confirms that. Click on the graph for a larger version.
That seems to pretty much debunk the NYT view that there is any relation to mainstream conservative values. But can any correlation be found in these data? Just for grins, let's take a look at minor parties. Bear in mind that in 2012 there was no doubt in the mind of anyone paying attention that President Obama would carry California and get all its electoral votes. So anyone inclined to vote for a minor party could do so "safely." Let's run a correlation between the PBE percentage in a county and its Green Party vote.
Paydirt. We get r = 0.379 and p = 0.0034. Social scientists break out the champagne when they get a p that low. The correlation coefficient is substantial for this type of research. Let's look at the scatterplot.
Now we get r = 0.683 and p = 3.44 x 10-9 . For those who don't do exponential notation, that's 0.00000000344. Wow. Should we break out some ridiculously expensive wine? Sorry, I don't own any.
The scatterplot confirms that PBE rises steadily with Libertarian vote.
In case you are wondering why the dots are arranged in neat columns, the quick and easy data set I am using has vote percentages rounded to 1/10 percent. If I were doing really serious research, I would go back to the raw votes and recalculate.
Conservatives and libertarians are often allied on political issues, but they come to their conclusions by different routes based on different values. They disagree on some important issues, especially crime. I recommend Jonathan Haidt's invaluable work for a more complete explanation.
In this case Green and Libertarians make even odder bedfellows. Many Greens are reds in disguise and just love state control of just about everything, anathema to what Libertarians believe in. Yet they seem to have common ground in the antivaxx folly.
For the three people who are still with me to the end of this long post, I hope you have enjoyed this statistical frolic and detour. The morals of the story are: (1) attitudes are much more complex than simple, one-dimensional liberal v. conservative; and (2) don't believe anything the New York Times says about conservatives or Republicans.
Update Note (2/9): I have tweaked the post from its original form to more accurately reflect Sen. Paul's statement and position, to provide a link to the video for those who wish to see it for themselves, and to make a few minor edits.
Even when the label-attachers recognize degrees of liberalism and conservativism and place people along a one-dimensional spectrum instead of two buckets, they are still working from a faulty model. A two-dimensional model with an individual v.collective axis along with a conservative v. liberal axis would come closer to the truth.
Now a bit about the measles flap. Vaccines are one of the great advances of modern medicine and have nearly wiped out many formerly dreaded and widely fatal diseases. Yet there are a few people who refuse to let their children be vaccinated. Some have individual medical reasons. Some have religious beliefs against any form of medical treatment. (Whether imposing such a belief on a child is proper parenting or child abuse is a separate controversy we won't get into here.) What I say below does not apply to the folks described in this paragraph.
Then there are people who are just against vaccines, generally because they think risks of vaccines outweigh benefits. At one point, there was a study in a peer-reviewed journal claiming a link between vaccines and autism, so there was a time when such a theory had some degree of respectability. That study has been completely debunked, and the fact it was ever published in the first place is part of a scandal that requires a reevaluation of peer review.
So at this point the anti-vaccine crowd is mostly crackpots, endangering the lives of their own children and others. Unvaccinated children who become infected spread the disease not only to the children of the other crackpots but also to those who are unvaccinated for the reasons noted above and to the few who contract a disease despite having been vaccinated. Anti-vaccine crackpotism is particularly strong in California, so it should be no surprise that if a measles outbreak were going to break out, Disneyland would be a prime candidate for it. And that is what happened.
Is there a political dimension to this? The New York Times thinks so. In "Measles Outbreak Proves Delicate Issue to G.O.P. Field," Jeremy Peters and Richard Perez-Pena write:
The politics of medicine, morality and free will have collided in an emotional debate over vaccines and the government's place in requiring them, posing a challenge for Republicans who find themselves in the familiar but uncomfortable position of reconciling modern science with the skepticism of their core conservative voters.Huh? Since when is anti-vaccine crackpotism related to conservative values? As James Taranto puts it in the WSJ: "How in the world did Republicans manage to get themselves cast as defenders of liberal antivaccine (colloquially 'antivaxx') wackos?" Fox News quoted an expert as saying that it is the people who shop at Whole Foods and drive Priuses that are the "antivaxx." (A side note. I very nearly bought a Prius the last time I needed to replace my car. I didn't, but it was a close call.) I like that term "antivaxx." I'll use it for the rest of the post.
So how does antivaxx really relate to political orientation? I decided to find out. First, I had to find a readily available data source online. I found data for the Personal Belief Exemption for vaccination by county in California, and voting data for the last presidential election is also available by county. Now, county isn't the best unit for this study. Some of California's flakiest enclaves are buried within massive Los Angeles County. But the data are quick, free, and easy. You often see this in research. The ideal data set does not exist, and it would be difficult and expensive to gather it, so researchers make do with what has already been gathered. This goes in the "limitations" section of the article that few people read at all but everyone should read first.
So, we have the percentage of kids whose parents claim the PBE in the 2014-15 school year and the 2012 Obama-over-Romney margin (a negative number in counties "carried" by Romney) for each county in California. We fire up the trusty R statistics package to correlate the two and get r = -0.144 and p = 0.28. The p tells us to forget about the r because it is meaningless. Small p is good, and the standard rule of thumb is that a result with a p over 0.05 is not "statistically significant." We can argue about that rule of thumb, but 0.28 is outside the range of argument. There is nothing worthwhile here. A scatterplot confirms that. Click on the graph for a larger version.
That seems to pretty much debunk the NYT view that there is any relation to mainstream conservative values. But can any correlation be found in these data? Just for grins, let's take a look at minor parties. Bear in mind that in 2012 there was no doubt in the mind of anyone paying attention that President Obama would carry California and get all its electoral votes. So anyone inclined to vote for a minor party could do so "safely." Let's run a correlation between the PBE percentage in a county and its Green Party vote.
Paydirt. We get r = 0.379 and p = 0.0034. Social scientists break out the champagne when they get a p that low. The correlation coefficient is substantial for this type of research. Let's look at the scatterplot.
Sure enough. The low PBE counties are mostly clustered in the lower left corner, having low Green votes as well. The high PBE counties are generally above the state average on Green vote and mostly well above. This is quite strong, as research like this goes.
But wait, there's more. The most irresponsible statement in the recent kerfuffle was made by Sen. Rand Paul, M.D., who most certainly should know better. Although saying that vaccines are generally good, he also said on CNBC, "I have heard of many tragic cases of walking, talking, normal children who wound up with profound mental disorders after vaccines." That appears to be an endorsement of the utterly discredited vaccines-cause-autism view. Paul is a libertarian, not a conservative. His main point is that he is personally pro-vaccine but thinks other people should have the right not to be vaccinated and to not vaccinate their own children if they choose. Okay. But how does not wanting to vaccinate your own children correlate with voting for the Libertarian Party?
But wait, there's more. The most irresponsible statement in the recent kerfuffle was made by Sen. Rand Paul, M.D., who most certainly should know better. Although saying that vaccines are generally good, he also said on CNBC, "I have heard of many tragic cases of walking, talking, normal children who wound up with profound mental disorders after vaccines." That appears to be an endorsement of the utterly discredited vaccines-cause-autism view. Paul is a libertarian, not a conservative. His main point is that he is personally pro-vaccine but thinks other people should have the right not to be vaccinated and to not vaccinate their own children if they choose. Okay. But how does not wanting to vaccinate your own children correlate with voting for the Libertarian Party?
Now we get r = 0.683 and p = 3.44 x 10-9 . For those who don't do exponential notation, that's 0.00000000344. Wow. Should we break out some ridiculously expensive wine? Sorry, I don't own any.
The scatterplot confirms that PBE rises steadily with Libertarian vote.
In case you are wondering why the dots are arranged in neat columns, the quick and easy data set I am using has vote percentages rounded to 1/10 percent. If I were doing really serious research, I would go back to the raw votes and recalculate.
Conservatives and libertarians are often allied on political issues, but they come to their conclusions by different routes based on different values. They disagree on some important issues, especially crime. I recommend Jonathan Haidt's invaluable work for a more complete explanation.
In this case Green and Libertarians make even odder bedfellows. Many Greens are reds in disguise and just love state control of just about everything, anathema to what Libertarians believe in. Yet they seem to have common ground in the antivaxx folly.
For the three people who are still with me to the end of this long post, I hope you have enjoyed this statistical frolic and detour. The morals of the story are: (1) attitudes are much more complex than simple, one-dimensional liberal v. conservative; and (2) don't believe anything the New York Times says about conservatives or Republicans.
Update Note (2/9): I have tweaked the post from its original form to more accurately reflect Sen. Paul's statement and position, to provide a link to the video for those who wish to see it for themselves, and to make a few minor edits.
...and this is why you don't want to get in an argument with Kent Scheidegger.
I made it to the end. American political parties are to some extent strange coalitions that often defy the "liberal" versus "conservative" conventional wisdom. I do not even know what I'd call myself, so I've quit trying.
And being the father of 16-month old just recently old enough to be vaccinated (and living in the LBC), I can't even describe my contempt for those who don't get their kids vaccinated. They really shouldn't have a choice. I'll personally hold their bratty kids down and shot them up myself. For free.
PS - personally I think the epicenter of anti-vaccination idiocy in Southern California is a bit further south, precisely in Mission Viejo - Orange County (I used to live there).
Entirely possible. I do wish I had a more finely grained dataset to check out those kinds of hypotheses.