Bloodless Coup

Fun With Statistics

Via Pharyngula, we get some initial results of the 2006 General Social Survey (GSS). The GSS is done every few years (two? four? I don't remember), and is an attempt to study the general characteristics of the population of the US. No one claims that it is perfect, but the GSS has been giving pretty much the same battery of tests every iteration for decades. The advantage here is obvious: since the same questions are given every time, we can track how the population has changed by the changing answers to the questions.

But that's not what this is about. Pharyngula points to correlations between religion and lack of scientific knowledge:

(click for larger image)

Let's deal with the depressing news first: by a slight (and likely statistically insignificant) amount, more Americans disagree with evolution than agree with it (that's the pair of columns on the right). That's not good, considering that evolution is the foundation for biology, and is taught (or should be) fairly early on in science. Given this result, one doesn't really have to wonder why other countries are beating the pants off us in when standardized test scores are compared across states.

However, Pharyngula wants to beat up on fundamentalism. As he notes, in this table the clear correlation is to fundamentalism: almost three times as many "fundamentalist" reject evolution as accept it (almost twice as many "moderate Christians" reject evolution as accept it). In every other religious category the number that accept evolution outweighs the number that reject it. If there were any doubts about the consequences of the rise of fundamentalism, this should leave no doubt: science and religion are antithetical (at least the version of religion accepted by the majority of Americans). Notice the numbers of respondents in each category: of the 1849 people surveyed, about 800 fall into either "fundamentalist" or "moderate Christian" (and, given the responses about evolution, I'm hesitant to call those people genuinely moderate). That's close to 50% (technically, 43.1%), which is a scary number of people, if they have this belief system. These are certainly depressing results.

(It's worth noting a significant, and perhaps fatal, caveat here: the religious categories are "self identified," meaning every person who took the GSS is asked to choose their own category for religious belief. People choose where they want to put themselves. Thus, it is possible that the data does not accurately reflect the reality of people's beliefs. In other words, given that "fundamentalist" has a certain connotation (as opposed to a defined denotation), some people do not see themselves as fundamentalist (even though they meet a definition) and won't self-identify into that category. This doesn't make the data wrong, but does add a note of caution to sweeping conclusions.)

Pharyngula goes on to argue that the following chart is scarier:

(click for larger image)

He argues that this is even worse news: that "some college" makes everybody accept evolution, except for fundamentalists, who do not change their beliefs. While the chart does show that "some college" increases the percent of people who accept evolution in every religious group (except fundamentalists), I think Pharyngula jumps to conclusions here: the key is that this is a chart of people who have had "some college," not "college graduates." "Some college" includes everyone who enrolled for a semester (and, perhaps even people who took courses for college credit in high school; technically, they have had "some college" education) up to people who have Ph.Ds. Thus, the data in the second chart seems more suspicious to me than the first one, and I'm reluctant to draw as firm a conclusion. Still, interesting, isn't it?

While ignorance is never a good thing, I'm not sure we can blame religion for it all. The problem with the two charts on evolution is that evolution has become a buzz-word, or at least a hot-button topic today. Evolution has been distorted, and is rarely reported accurately in the media. Thus, polling on evolution runs the risk of being distorted by the media and debate surrounding the issue. Pharyngula wants to draw conclusions about the compatibility of fundamentalism and science from a potentially distorted pool of data. I'm not sure I buy it, and two additional sets of data (one a table, the other just numbers) call into question the idea that fundamentalists are more ignorant than everybody else. (via Marginal Utility):

Yes, you read it right. About 30% of fundamentalists either believe the Sun revolves around the Earth, or don't know that the Earth revolves around the Sun. This isn't a good number: this is a basic fact of science that everyone should know. The problem here, however, is that about 30% of every religious group believes the same. The numbers for fundamentalists may be slightly higher (they look to have over 30% with this belief system, where some of the other religious groups are under 30%), but the difference does not look to be statistically significant (and even if it is, its a small difference).

Lastly, a purely scientific question was asked:

Update: After I posted this, it occurred to me that I should look at a less religiously charged scientific knowledge question, e.g., on the experimental method. The question asks how best to test a new drug: with or without a control group. (This is explained in the question.) Ns are the same as in the other graphs.

Total: 79% w/control, 16% without, 4% DK

Fund prot: 76% w/control

Mod prot: 81%

Lib prot: 84%

Catholics: 75%

Jews/lib others: 82%

None: 85%

So, Fundamental Protestants are a bit below the other groups on this form of scientific knowledge, too, but the difference is less extreme.

So, in a question about the basic scientific method, about 4/5ths of Americans got it right. Being a fundamentalist reduces that ratio (to about 3/4th getting it right), while being a "moderate" or "liberal" protestant (or "none" for religion) raises you above the average. While being a fundamentalist make you more likely to reject basic science, the difference is not overwhelming (the maximum change for this question is 10 percentage points: the worst group (fundamentalists) had 25% get it wrong, while the best group ("none" for religion) had 15% get it wrong).

In the end, I think Pharyngula is engaging in a bit of baiting here. While it is true that the question on evolution produced some very discouraging results (see the first table, where about three times as many fundamentalists rejected evolution as believed in it), questions about more prosaic scientific methods produced much less flamboyant results (yes, fundamentalism scored worse, but not that much worse than other religious groups). I think, but cannot show (no data in the GSS), that evolution has become a "buzz word" more than a debate, and that many fundamentalists are reacting to the word (and the debate in the media) rather than the reasoned science. In other words, the GSS results are tracking beliefs about the debate on evolution, and many self-identified fundamentalists may feel the need to reject evolution in the GSS as a political statement (or something like that). Overall, the results to clearly show that fundamentalism puts you at a disadvantage with respect to science, though how much of one is debatable (some of the results may not be statistically significant, and if they are produce only a marginal problem for fundamentalism.

In the end, the biggest problem these GSS results point to is the depressing state of science for all Americans. Only 80% know that the Earth revolves around the Sun? Only 80% know what a control group is for? That's really depressing.

Posted by baltar at June 10, 2007 09:29 PM | TrackBack | Posted to Crunchy Nutbars | Culture | El Infierno de kansas | Religion | Science | The Academy