A conundrum about evolution acceptance

Looking through the raw data from Pew's surveys on science from last summer, I cannot make heads or tails of certain findings.

I started out looking at the correlation between rejection of evolution and rejection of global warming. As one would expect, it was significant. But the survey handily also asked not just about people's personal belief regarding these sciences, but also their beliefs about the views of the scientific community. And again, people who think scientists reject evolution are more likely to think scientists reject global warming.

Then I got interested in whether people who think evolution is rejected by scientists also personally reject it, and that's where things get odd.

Here's the table from R:

Do scientists agree that humans evolved over time?No
Don't know
YesHumans and other living
things have/have not
evolved over time
Have not
294
76
251Don't know
40
62
49Have
249
103
877

You'd expect all the responses to be grouped along the diagonal, with people who know scientists accept evolution also accepting it themselves, those who think scientists rejecting evolution rejecting it themselves, and maybe some spread among people who don't know or refuse to answer one question or the other.

Instead, the vast majority of people who accept evolution know that scientists do as well (and vice versa). The "don't know" answers are distributed remarkably uniformly, as are the answers from people who either personally reject evolution or who think scientists do.

To a degree, this makes sense. Lots of creationists know evolution is well-supported by scientists. They think the scientists are wrong, biased as they are by evil sciency things. I suppose it even makes sense that creationists should be equally likely to think scientists accept or reject evolution, though the nearly identical numbers in those cells are surprising.

What makes no sense is that, among people think who think scientists reject evolution, equal numbers personally accept and reject the science. If you think scientists dispute the claim, why would you accept it?

The only explanation I can conceive is that people in the survey were choosing answers somewhat at random, or at least with no clear conception of what they were answering. And surely all the clueless people too proud to say "don't know" didn't choose the same combination of answers. So how many people from other groups should be moved into the don't know column? If we assume they distributed randomly, it requires us to take 249 from each other group, leaving 2 people agreeing that scientists back evolution but who disagree personally, which is clearly too small. And reducing the group who reject evolution and think scientists do too to a mere 45 people doesn't seem reasonable. Perhaps this difference is driven partly by a slight asymmetry in the questions, where people are asked about scientists' views on the evolution of humanity, while they are asked their own views about humans and other living things. Acceptance for evolution of non-humans generally runs higher than acceptance of evolution alone, but simply mentioning human evolution should get people's dander up either way.

My guess is that most people haven't got carefully reasoned opinions about the issue. Most know that scientists back evolution, and so they back it as well. Outside that group who back the scientific consensus, there's no particular coherence. And maybe that's an important enough lesson, but I'd like a bit more. I'm checking with Pew for clarification, but thoughts from readers in the comments are welcome.