Evolution-doubting and illiteracy, part 3

Last week's Science has an article about "public acceptance of evolution" by Jon Miller, Eugenie Scott and Shinji Okamoto. The article covers results of polls that demonstrate that a low proportion of Americans believe that humans evolved, compared to relatively higher proportions in Europe and Japan.

Beginning in 1985, national samples of U.S. adults have been asked whether the statement, "Human beings, as we know them, developed from earlier species of animals," is true or false, or whether the respondent is not sure or does not know. We compared the results of these surveys with survey data from nine European countries in 2002, surveys in 32 European countries in 2005, and a national survey in Japan in 2001 (5). Over the past 20 years, the percentage of U.S. adults accepting the idea of evolution has declined from 45% to 40% and the percentage of adults overtly rejecting evolution declined from 48% to 39%. The percentage of adults who were not sure about evolution increased from 7% in 1985 to 21% in 2005. After 20 years of public debate, the public appears to be divided evenly in terms of accepting or rejecting evolution, with about one in five adults still undecided or unaware of the issue. This pattern is consistent with a number of sporadic national newspaper surveys reported in recent years (6-10) (Miller et al. 2006:765).

There's no question that the central point of the article is correct -- a large proportion of Americans reject the idea that evolution explains many of the central facts of life. And I think that Miller, Scott and Okamoto hit upon most of the essential reasons why: the strength of American fundamentalism, the incorporation of creationism into political platforms, and lack of information about "modern genetics". I wonder whether there are additional factors that might be explored, such as a greater skepticism of pronouncements from "experts", or wider awareness of frauds -- side effects of the American political experience and its post-Watergate distrust of authority. But certainly the overwhelming majority of American attitudes toward evolution are influenced by religion, one way or another.

However, I noticed another thing reading this article: just how badly written these poll questions apparently are. Take the one cited above:

"Human beings, as we know them, developed from earlier species of animals."

Human beings have never "developed" from earlier species of animals. We evolved from them. Adult human beings develop from zygotes, embryos, fetuses, infants, children, and adolescents. And of course, some of these latter categories are themselves human beings (teenagers being the main exception!). That's why "evolutionary developmental biology" is called evo-devo instead of devo-devo!

Now, I can understand why a pollster might substitute "developed" for "evolved". "Evolved" is a "charged word", and people might react to it strongly. "Developed" has an everyday meaning that people understand, which isn't necessarily connected to ontogeny.

But really, who exactly is going to "react strongly" to the word evolution, but is going to agree with the notion that human beings developed from earlier species of animals? Somebody who would say, "Oh, well, if we didn't evolve, then I guess I can accept that we developed from monkeys." It's an empty set!

The article presents another set of questions from one of these polls:

For example, only a third of American adults agree that more than half of human genes are identical to those of mice and only 38% of adults recognize that humans have more than half of their genes in common with chimpanzees. In other studies (1, 14, 15), fewer than half of American adults can provide a minimal definition of DNA. Thus, it is not surprising that nearly half of the respondents in 2005 were not sure about the proportion of human genes that overlap with mice or chimpanzees.

Or just maybe, they failed to recognize these "facts" because they are nonsensical!

For example, "more than half of human genes are identical to those of mice"? Let's start by defining some terms. "Identical" generally means "exactly the same". Two identical genes should have the same nucleotide sequence, no? Now, we haven't found all the genes in the human genome. And further, we don't know what segments of DNA should really be included in any given "gene", since each may be regulated by sites that are very far from the coding sequence. So we will want to be very cautious no matter what we claim about similarities between humans and other species.

But let's consider the known coding sequences alone. A sample of some 13,000 genes in the human and draft chimpanzee genomes shows over 39,000 amino-acid coding differences between the two species. This means that a given human will differ from a chimpanzee by an average three amino-acid coding substitutions per gene. Certainly these are not equally distributed -- some genes are more different than others. But far fewer than half of this sample of genes are identical in their amino acid sequences. Even fewer -- only around one in ten -- are identical in their nucleotide sequences, including synonymous substitutions. If we include intronic sequences as part of each gene, then none of the 13,000 genes have identical sequences in humans and chimpanzees. Mice, of course, are more different from us than chimps.

OK, let's be generous and assume that the poll intended something that makes sense, like "more than half of nucleotides are shared between human and chimpanzee genomes." On the one hand, the obtuseness of the question would seem to vindicate those Americans who can't provide "a minimal definition of DNA". I mean, the poll doesn't understand DNA, so why should they?

On the other hand, this looks like what we in the professoriate would call a "trick question". As in:

"Hmm.... The book says that humans and chimpanzees have ninety-eight percent sequence identity, but this question says 'over half'. Now I know that ninety-eight percent is more than half. But why would the question just make up a number that was so far off? Is the question asking about something I didn't study? And it doesn't say 'nucleotide sequence', it says 'genes'. And I know that there aren't any full gene sequences that are identical. And this other question says "in common with." What the hell does that mean? Is it about genes versus pseudogenes? Oh, crap, why didn't I take economics instead?"

Now, sure, there probably aren't very many people who answer these poll questions the "wrong" way because of objections like mine. But when scientists can't seem to get their facts straight, just how exactly are nonscientists supposed to become "literate"?

I actually think 38 percent is pretty impressive penetrance for the human-chimpanzee factoid. After all, only twice that many know that the Earth revolves around the sun.


Miller JD, Scott EC, Okamoto S. 2006. Public acceptance of evolution. Science 313:765-766. DOI link