A recent NSF poll on science knowledge, in which 25% of Americans failed to correctly answer answer that Earth orbits the Sun, is getting a lot of press; you can find articles on NPR, Discovery, and phys.org. This is depressing, and highlights the concerning state of science knowledge in the U.S. However, most of these stories pretty much stop there. But there’s (always) more to it than that. Ignoring issues of sample size (only 2,000 survey respondents or so), there are multiple levels to this survey.
First, the survey asked 10 questions and collected data from all around the world. We can then compare the state of scientific knowledge in the U.S. with that of other countries. With respect to the orbit question, the United States scored higher than every other surveyed country (including the EU) except South Korea.
Interestingly, over 80% of survey respondents knew about continental drift (i.e. that the continents have been moving around for millions of years) but less than half acknowledge that humans descended from earlier species. This seems paradoxical to me. The U.S. scored better on the continental drift question than all but the EU, Japan, and South Korea, but scored worse on the evolution question than every surveyed country but Russia. There’s more to the survey than this, but I won’t get into a play-by-play of the results. The curious can see the results here in the original report.
A New Angle
What is perhaps more interesting, and overlooked by all of those other articles, is that the NSF has conducted this same survey twice before in the U.S., once in 2001 and again in 2004. This lets us look at trends over time (and I love time trends!).
For example, only 80% and 78% of respondents correctly answered the continental drift question in 2001 and 2004, respectively. So it seems like there’s been some improvement (although it’s hard to tell because there’s no information on the margin of error in the survey, which is critical to extrapolating to a population).
In 2001 and 2004, 44% and 53% correctly answered the evolution, respectively. Compared to the 48% from the current survey, this looks like a pretty stable trend. That’s probably the most upsetting part of all. In over ten years, we haven’t made any progress on that front (same with a question about the origins of the universe, which is consistently somewhere in the 30 – 40% range).
Overall, I’m not sure that this report is worth the hoopla. Sure, science knowledge in the U.S. isn’t up to snuff, but it’s a lot better than other countries. And what is up to snuff? Will we ever hit the 80-90% range and is that even a reasonable goal? Sure, the majority of the population doesn’t buy evolution, but we knew that already. Probably the most depressing point is the fact that we haven’t been able to improve over the past decade. With more and more Americans being college educated, it seems odd to me that our science knowledge isn’t improving. Is there a failure of the education system here?
Further, the sample size in these surveys is remarkably low (something the Discovery article completely and surprisingly brushes aside). Extrapolating the results of 2,000 respondents to over 315 million (that’s 315,000,000 compared to 2,000)… well… that’s tenuous at best. The NSF report appears to concatenate results from numerous surveys, but its unclear to me whether all of those results went into this specific questionnaire or not. Regardless, I guess I’m not surprised and not really all that disappointed, either.