Where’s the Backlash?

Jerry Coyne has a post up reporting on new polling data on science and religion coming out of Virginia Commonwealth University. Jerry notes that the numbers for the evolution questions are broadly consistent with what past surveys have found. I mostly agree, but there was one number that jumped out at me. Here was the question:

Which of these statements comes closest to your views on the origin of biological life: biological life developed over time from simple substances, but God guided this process, biological life developed over time from simple substances but God did not guide this process, God directly created biological life in its present form at one point in time?

The option, “biological life developed over time from simple substances but God did not guide this process,” was selected by 18% of the people. That struck me as higher than normal.

Then I remembered the numbers from the long-running Gallup poll on this question. In the Gallup poll the option was phrased, “Human beings have developed over millions of years from less advanced forms of life, but God had no part in this process.”

The percentage of respondents selecting this option has grown slowly but steadily since 1999. It was 9% in 1999, then 12% in 2001, then 13% in 2004 and 2006, then 14% in 2007 and 2008.

Now we have 18% in the VCU survey.

These polls are fraught with peril, since it is very difficult to capture people’s religious beliefs with simple poll questions. For example, the phrasing in the VCU question said simply that God did not guide the process. A person could well believe in God and still select that option. The Gallup phrasing (God had no part in the process) is far more stark. What if you believe God created the initial conditions in which evolution unfolds? Doesn’t that constitute having a part in the process?

So caution is always needed. But with this number growing steadily over all of these polls, to the point of doubling in ten years, it is certainly possible that there is a real trend here.

I do not know the explanation for these numbers, and I have no interest in speculating.
It does seem interesting, though, that while many people are wringing their hands over the supposedly pernicious effect of the New Atheists on evolution acceptance and education, the numbers show not the slightest evidence of a backlash. To the extent that the numbers are moving at all, they seem to be going in the right direction.

Comments

  1. #1 NewEnglandBob
    July 13, 2010

    Of course there is a backlash. It is in the horde of accommodationists. Those milquetoasts who are afraid of the big bad theistic wolf who will come blow down their houses of straw.

    The accommodationists number in the 10s. Meanwhile the so-called ‘New” Atheists who vastly outnumber the accommodationists are getting the word out using critical thought and logic, backed by scientific information and are getting the job done in the US which is raising those numbers.

  2. #2 Zach Voch
    July 13, 2010

    Well, Jason, there’s a positive trend in spite of anti-theism.

    They’ll still drive away all of the moderates! Just you wait and see!

    The onerous Dawkins, crowned King Godless in 2012, riding a great, anti-clerical dragon, will descend upon the Unitarians, Catholics, and progressives, belching fire toward every Bible held by community organizers, who, painfully, trudge with tremendous difficulty through the alligator-ridden swamp of extremists.

    “We just want to have cookouts and talk about Jesus and feed poor kids in Africa,” they plead, trembling before King Godless, who responds with a maniacal, quiet laugh, building into outraged momentum.

    “But it isn’t true, is it? You have no evidence, and… I AM THE KING OF TRUTH, AND TRUTH IS GODLESS!”

    After they recover from their burns, they will surely become creationists.

  3. #3 James Sweet
    July 13, 2010

    I’m sure many people have faith that the backlash is still coming eventually. Ba-dump.

  4. #4 I.Strange
    July 13, 2010

    The percentage of respondents selecting this option has grown slowly but steadily since 1999. It was 9% in 1999, then 12% in 2001, then 13% in 2004 and 2006, then 14% in 2007 and 2008.

    Now we have 18% in the VCU survey.
    .

    The 2005 VCU survey has 17%, however.

    http://www.vcu.edu/lifesci/images2/survey2005.pdf

  5. #5 Jason Rosenhouse
    July 13, 2010

    Zach -

    I knew someone was going to say that!

  6. #6 Zach Voch
    July 13, 2010

    I took the pleasure of preempting any future comments which take a similar line.

    I’ll wait for somebody to say something similar with a straight face.

  7. #7 JimR
    July 13, 2010

    Maybe the accommodationists are aspiring authors who hope to write a sufficiently muddled book that will win a Templeton award and leave them awash in cash.

  8. #8 Stephenson Billings
    July 13, 2010

    I find it quite sad that there is no room in the evolution “debate” today. There is ample evidence that Darwin’s theories were fraught with technical and moral errors, yet our schools still pump out graduates who take his theories completely at face value. And yet where is the bigger picture in all of this? Does anyone have a complete and spiritual understanding of what life is? The Bible has been offering a very well supported answer to all these questions for thousands of years but a few liberals ensconced in our ivory towers seem to disregard the faith of billions of people so they can push their philosophy that apes are our brothers. It’s incredibly offensive and wrong and I worry for the souls of all these children who grow up to be atheists.

  9. #9 GAZZA
    July 13, 2010

    The thing I find most amusing about the accomodationists is the claim that the “New Atheists” are divisive.

    As if the accomodationists weren’t.

  10. #10 GAZZA
    July 13, 2010

    Stephenson@8: There’s plenty of room in the debate. Here’s how you enter it: go get yourself a degree in zoology, evolutionary biology, or similar. Accredited universities all over the world offer them, they’re generally pretty comprehensive.

    Can’t be bothered to do that? Yeah, neither can I. But if we can’t be bothered to learn the subject ourselves, then the logical thing to do is to accept the evidence that the experts tell us – you know, the gentlemen and ladies that DID go to all that trouble.

    Ready a fairy tale book, unfortunately, doesn’t count as cliff notes.

  11. #11 NJ
    July 13, 2010

    There is ample evidence that Darwin’s theories were fraught with technical and moral errors

    [Cite]

    It seems that you didn’t learn much from getting pummelled over at PZ’s place. The tone may be different here, but you’ll still get torn a new one…

  12. #12 Valhar2000
    July 13, 2010

    here is ample evidence that Darwin’s theories were fraught with technical and moral errors

    “Darwin’s theories” are close to 150 years old; of course you are going to find errors! Biologists have been working during that time, and have extended and corrected. Darwin’s work was astounding in its time, which is why he is still remembered and respected, but we have moved on. Pointing out niggles in a 150 year old book while ignoring the state of the art does not paint you in a good light.

    As for the moral errors thing: science is amoral. It is not prescriptive, but rather explanatory. The results and methods of Science may often be invoked in discussions of morality, but only to the degree that one is interested in using knowledge of the real to make moral decisions that affect the real world. One sure sign of pseudoscience is a theory that claims to provide moral guidance through scientific means.

    In short, science tells us what it is and how it works. To figure out whether “it” is good or bad we have to use something else.

  13. #13 natual cynic
    July 13, 2010

    I find it quite sad that there is no room in the evolution “debate” today. There is ample evidence that Darwin’s theories were fraught with technical and moral errors, yet our schools still pump out graduates who take his theories completely at face value.

    There is no debate. It’s over. The reason that there is no scientific debate is because the evidence is far beyond sufficient to confirm Darwin’s basic ideas. There certainly are areas where Darwin was in error, but that’s due to the lack of knowledge. And moral errors are irrelevant. I don’t like the Theory of Gravity because I occasionally fall down and it is becoming progressively harder to move my ass. Big deal. And your idea that any scientist in the field takes the ToE at face value demonstrates your lack of understanding of the scientific endeavor. Glory awaits the one who comes up with a better theory that also takes into account the mountains of data that have supported the ToE.

    And yet where is the bigger picture in all of this? Does anyone have a complete and spiritual understanding of what life is? The Bible has been offering a very well supported answer to all these questions…

    Is there a bigger picture? Show the evidence. And what is a spiritual understanding of life? Show the evidence that it exists.

    …so they can push their philosophy that apes are our brothers. It’s incredibly offensive and wrong and I worry for the souls of all these children who grow up to be atheists.

    It’s not a philosophy, it’s hard scientific evidence that we are related to apes. Why worry about the soul when you cannot show that such a thing exists. So I find it offensive that you and your ilk want to perpetuate your mythology as truth.

  14. #14 Owlmirror
    July 13, 2010

    As noted in his url, “Stephenson Billings” is from “christwire.org”, and it was pointed out that Christwire parody of fundamentalist Christianity similar to Landover Baptist.

    Or was they used to say in the old Usenet days,

    YHBT!

    HTH, HAND.

  15. #15 Owlmirror
    July 13, 2010

    Um:

    Christwire is a parody, etc.

    Fixed.

  16. #16 Hamilton Jacobi
    July 14, 2010

    There is another major difference in the phrasings: Gallup refers to “human beings,” while VCU refers only to “biological life.” Many people are unfazed by the concept of “lower animals” evolving, but freak out when told that humans are just another animal. Others aren’t bothered by humans evolving, but insist that God must have zapped in a soul at some point.

  17. #17 Crusader
    July 14, 2010

    The problem with “experts rule” is that has been used to silence AGW skeptics. Instead we are told to hand over 20% of the GDP because “experts” say AGW is an imminent threat.

  18. #18 tresmal
    July 14, 2010

    we are told to hand over 20% of the GDP…

    20%? Citation needed.

  19. #19 Deepak Shetty
    July 14, 2010

    Stephenson Billings

    But a few liberals ensconced in our ivory towers seem to disregard the faith of billions of people so

    You should read
    http://www.jesusandmo.net/2009/09/04/four/

  20. #20 hoary puccoon
    July 14, 2010

    Stephenson Billings–

    “…moral errors….”, in science, are actions such as falsifying data or taking credit for other scientists’ work. Darwin was known to be extremely scrupulous in that regard. If by “moral errors” you mean something like, “I, personally don’t like the conclusions of Darwin and/or modern biology, and think they can be misused,” well, that says nothing about the truth or falsity of the science.

    If you’re worried about misuse of ideas, before you diss Darwin, you might want to study the thirty year’s war and the Spanish Inquisition, not to mention Martin Luther’s writings re Judaism. How well is Christianity doing on the score of “moral errors”?

  21. #21 Pseudonym
    July 14, 2010

    Surely from the point of view of the accommodationist vs new atheist thingy, this is a completely irrelevant number.

    18% of people believe that deities had no hand in evolution. Fine. Ask them if everyone else should believe that deities had no hand in evolution and you’ll get a much smaller number.

  22. #22 Zach Voch
    July 14, 2010

    Stephenson,

    Hello! First, I want to point out that your writing could serve as a case study in memetics. Let’s look at some examples of things which are memes, that is, things (true or false) which are units of information repeated throughout culture. Take lolcats, or Rick Astley. Now, take this:

    There is ample evidence that Darwin’s theories were fraught with technical and moral errors, yet our schools still pump out graduates who take his theories completely at face value.

    The first: completely true! Darwin’s original theory has been extensively modified, overhauled, and edited. However, the core features focused on in public schools, such as descent with modification via natural selection acting on variation in a population, are still relevant and central to modern evolutionary thought. So, whether or not a particular schoolteacher claims that Darwin was right about absolutely everything is a separate question to whether or not Darwin was right about many things.

    Of course, that’s not what you meant, is it? You are probably referring to things like the common ancestry of man and other primates, or more generally, the common ancestry of all cellular life. But sorry, those things have stood the test of time, greatly strengthened, in fact, since Darwin’s time.

    The Bible has been offering a very well supported answer to all these questions for thousands of years but a few liberals ensconced in our ivory towers seem to disregard the faith of billions of people so they can push their philosophy that apes are our brothers.

    What does it mean for something to be an answer? For example, suppose that you were to ask me how baseballs are made. I answer, quite simply, that a baseball producer made them.

    Well, you might wonder, that’s a tautology. If I just say “it’s a baseball, therefore it was produced by a baseball producer,” then all that I have done is hand you a definition, not an explanation.

    This is what the Bible does. It says “this is a creation, and the creator made the creation.” No demarcating mechanism is provided, so it fails to be an answer. So, this answer cannot be, by the very nature of it, well-supported, because it isn’t even an answer.

    But we run into other problems as well. Suppose the Genesis account is divinely inspired… Even if the creation part is left hollow, what about the other explanatory bits, such as Noah’s flood?

    Well, if it’s true, as many creationists claim, that all of the geological column was deposited by the flood, why then do we have layers? For, suspensions in water settle into a single, graded layer, not distinct layers of distinct compositions. Further, the column should be ordered by density of material, with all the largest chunks of rock at the bottom.

    But that’s not what we find.

    Further, we wouldn’t expect to find fragile structures, such as stream beds, animal burrows, coral reefs, and animal tracks all throughout the column.

    But that’s what we find.

    Lastly, if all of dinosaur fossils are the result of the flood, we should expect to find them dispersed throughout the sediment, or at least in no obvious order. Instead, we find that certain types of fossils simply do not appear below certain layers.

    So no, the flood story isn’t well supported. The creation story isn’t even supportable.

    As for the “moral errors,” you would have to point them out. First, a brief note in metaethics, which you should keep in mind in your response.

  23. #23 Zach Voch
    July 14, 2010

    Haha crap, I got poe’d.

  24. #24 Valhar2000
    July 14, 2010

    The problem with “experts rule” is that has been used to silence AGW skeptics. Instead we are told to hand over 20% of the GDP because “experts” say AGW is an imminent threat.

    1 trillion dollars to ratify Kyoto? That will be catastrophic! The world economy will disintegrate! Billions of people will be left eating grass and tree bark!

    1 trillion dollars to bail out Wall Street? Sure, we can do that no problem.

  25. #25 Stephenson Billings
    July 14, 2010

    Well thank you for the response and it surprises me that some of you are sincere enough to take me seriously while others display the type of arrogance one would expect from today’s crop of academics– quick to label me a false prophet and not take anything I say seriously. I am pleased to inform you that you, my antagonists, really underestimate my passion and devotion to the Word. I will forever fight for the righteous and the saved.

    As for your other arguments, yes you are experts at twisting words and creating structures that prop up your lack of faith in an almighty power determining our fate but where is your passion? Your words are dry and mundane and could easily be turned into arguments supporting the concept that all life descends from tomato soup. I’m not buying and most people in America will not find you convincing. You have to have the power of persuasion, the passion of a deeply held belief to drive you. Once you live and breathe that love and willpower honestly, you will get others to follow and believe. In many ways our faith in Jesus is a self-fulfilling prophecy. We need to believe for him to continue to survive in our hearts. In that sense, you all sorely fail. I do pray and I do love you all.

  26. #26 hoary puccoon
    July 14, 2010

    Stephenson Billings @25;
    “… our faith in Jesus is a self-fulfilling prophecy. We need to believe for him to continue to survive in our hearts.”

    In other words, all our attempts at logic will fail, because you cannot emotionally face the truth. There was a writer named Arthur Koestler who wrote a book called, “The Case of the Midwife Toad,” about a scientist who faked data and eventually committed suicide because he couldn’t accept evolutionary theory. Koestler was very much on the scientist’s side. Eventually, Keostler himself committed suicide, for the same reason. I fear you may be going the same way.

    But I think I speak for many here when I say that I love life, and one of my pleasures in it is understanding the power and beauty of science, and especially of evolutionary theory. As Darwin said, “there is grandeur in this view of life.” I sincerely hope that someday you will realize that.

  27. #27 386sx
    July 14, 2010

    I hate it when today’s crop of academics go around all the time quick to label Stephenson Billings a false prophet. I hate it when that happens!

  28. #28 Science Avenger
    July 14, 2010

    Crusader said: The problem with “experts rule” is that has been used to silence AGW skeptics.

    Baloney. Do you not see how self-contradictory it is to complain about being silenced on a public forum?

    As for the AGW skeptics, they are just creationists in a different forum, using the same anti-science tactics, and of course doing no science themselves. Tehy aren’t silenced, but they are ignored, and for good reason.

  29. #29 Larry Spencer
    July 14, 2010

    Stephenson Billings @25:

    Your words are dry and mundane and could easily be turned into arguments supporting the concept that all life descends from tomato soup. … You have to have the power of persuasion, the passion of a deeply held belief to drive you.

    I was a devoted, Bible-believing Christian for 40 years. At age 52, I left my faith behind because I learned that the truth lay elsewhere. In fact, the conservative Christians I had so admired had shown utter lack of integrity when it came to truth. Evolution was only one example.

    What convinced me that my faith was untrue were logical, dispassionate arguments. Yes, some of us do listen to arguments and facts, and even try to ignore the irrelevant and deceptive “passion” with which they are delivered.

    …most people in America will not find you convincing.

    Sadly, I think you are correct. Most peoeple are only interested in data that confirm their beliefs. If you do not wish to be one of them, may I suggest that you read some science books by evolutionists, if you have not already done so? I found that there was a big difference between the case for evolution as presented and then “refuted” by creationists, and the case for evolution as presented by evolutionists themselves.

  30. #30 Tulse
    July 14, 2010

    There is another major difference in the phrasings: Gallup refers to “human beings,” while VCU refers only to “biological life.”

    Actually, as I read it the question differences are even more profound, as the VCU talks about life developing from “simple substances”, or in other words, abiogenesis. Agreement on this issue are in a very real sense far more “scientific” than just saying that humans evolved, as it really pushes against a major Creationist touchstone. Moderately-informed creationists will often say that “life doesn’t come from non-life” (or make arguments about hurricanes and junkyards). If nearly a fifth of Americans are willing to say that not only is godless evolution true for humans, but godless abiogenesis as well, that is a huge advance.

  31. #31 Josh Rosenau
    July 15, 2010

    I agree with Hamilton Jacobi. Poll questions referring to human evolution consistently perform worse than those referring to all life. For instance, this Harris poll found:

    In reply to one question almost half (45%) of adults say they believe humans were created directly by God and only 29% say they evolved from other species. In reply to another question 53% of these same people say they believe that “plants, animals and human beings have evolved over time,” and only 21% say they do not believe this, with fully 25% who are not sure or decline to answer.

    So you have an 8 point spread in that poll when the only difference is whether people are asked about human evolution or evolution of all life. Comparing between the Gallup and VCU polls gives us a 9 point spread, which is statistically the same difference, even without accounting for the difficulties of comparing two different pollsters.

    In other words, there’s no change, not a 10 point shift or “doubling of support” for evolution. In fairness, you should correct the erroneous claim, and perhaps revise the dubious conclusions about accommodationism built on that flawed premise. The numbers show no sign of moving at all, nor have they for the 30 years Gallup has been polling the issue.

  32. #32 Jason Rosenhouse
    July 15, 2010

    Josh -

    The Gallup poll all by itself shows a slow but consistent rise in support for the option of evolution with no explicit mention of God, from 9% to 14%. Furthermore, the idea that the Harris poll is some sort of infallible metric for telling us how to interpret the 18% figure in the VCU poll is itself a flawed premise.

    You also need to read my post more carefully. The thing that doubled in ten years (10 points? WTF?) was the percentage choosing the option of non-guided evolution when given the choice of three options that we can loosely characterize as atheistic evolution, theistic evolution and creationism. And the only conclusion I drew was that that the recent poll numbers show no evidence of a backlash against the New Atheists. I said nothing about accommodationism at all.

    Even if the numbers are flat that is still enough to provoke the question in the title of this post. I keep hearing that the New Atheists are hurting the cause of promoting science education. I am looking for evidence that that is the case. The NA’s have been around long enough that there has been time for the poll numbers to reflect any pernicious effect they are having. But no such effect is turning up.

  33. #33 386sx
    July 15, 2010

    If I were a creationist and someone were trying to “accommodate” me, I would recognize it for the total baloney that it is. I would play along with their coy accommodation game and mess with their heads, and take every inch I could get. And then I would talk to invisible people and pretend like I could “poof” stuff with magical powers. And my god would be able to fly up up into the sky like a tweety birdie and walk on water and stuff. (If I were a creationist, that is.)

  34. #34 Matti K.
    July 15, 2010

    #32: “I keep hearing that the New Atheists are hurting the cause of promoting science education. I am looking for evidence that that is the case. The NA’s have been around long enough that there has been time for the poll numbers to reflect any pernicious effect they are having. But no such effect is turning up.”

    You are not taking into consideration the benevolent effect provided by the accommodationists. Without their selfless work, there could very well be a backlash.

    If only the New Atheists would shut up, the Acommodationist Marketing Machine could do its work unhindered and soon no american would doubt evolution.
    :-)

  35. #35 Josh Rosenau
    July 15, 2010

    I responded to you and Jerry at greater length at TfK.

    I’m not sure where you get the prediction of a backlash detectable in the polls (maybe someone said something at some point that works, I just don’t know and the details of the prediction would matter). It strikes me as unlikely that any effects of the “New Atheists” would be visible on short time scales.

    The comparison between Gallup and VCU is completely flawed, and I don’t know how you could leave it sitting there, uncorrected. As a mathematician who’s written a book on statistics, people will expect that your statistical analysis will meet basic standards, and this just doesn’t. Anyone who knows anything about polling knows better than to attempt an apples/apples comparison between questions asked by different pollsters using different wording (at least not without serious statistical work, which you didn’t do).

    The poll question you’re looking at changed 1 point between VCU’s 2005 poll and their 2010 poll. That apples to apples comparison wipes out your claimed “doubling,” since Gallup’s polls flanking that 2005 survey both had 13%, and VCU’s was at 17%. Knock at least 4 points from your claimed doubling between Gallup ’98 and VCU ’10, and that leaves only a 5 point shift, which is within (but at the limits of) the overlapping margins of error.

    Finally, the option you categorize as “theistic evolution” was categorized by the VCU pollsters as “intelligent design.” I think that’s an error on their part, because the question encompasses both ID and TE. It’s just as wrong for you to leave ID out of your description of it as it is for them to leave TE out. Note also that that category is the only one significantly different (much larger) in VCU ’10 than VCU ’05.

    Which brings us back to the question of “backlash.” The rise of “New Atheists” isn’t the only thing that’s varied in the last 5-10 years. These data are nowhere near sufficient to tease out the effects of a series of books that were read by maybe 1% of the population, or of blogs that, in aggregate, are read by still fewer people (though both are surely discussed more widely). Especially not when there’s also an active creationist movement and active science education efforts of various sorts. 2005 brought the Dover verdict, 2009 brought a lot of evolution (and general science) outreach. I’d guess more people know about those things than have read a book by one of the Four Horsemen. So how do we attribute causality with multiple potential factors acting at once? There’s no randomization of treatments, and no data on which to perform statistical controls.

    I know that, and you know that, but do all of your readers? Could this have been a teachable moment about the responsible use of statistics and the real difficulties of interpreting social survey data, rather than a chance to make unjustified digs at people who you disagree with?

  36. #36 Matti K.
    July 15, 2010

    Mr. Rosenau:

    Above you imply that the “new atheists” have a marginal contact surface within the population of USA.

    If so, then why do many accommodationists (although not necessarily you) constantly worry about the supposedly harmful effects outspoken atheists have on the acceptance of science by religious people?

    Of course, public manners is a valid topic of discussion, but if the target audience of f.ex. NCSE is in no contact with the “new atheists”, why whine (speculatively) about their harmful effect on this outreach?

  37. #37 Jason Rosenhouse
    July 15, 2010

    You keep saying that I claimed the number doubled in ten years. What I actually said was, “So caution is always needed. But with this number growing steadily over all of these polls, to the point of doubling in ten years, it is certainly possible that there is a real trend here.” That followed a paragraph in which I discussed how the differences in phrasing made comparisons difficult, if not as thoroughly as you would have liked.

    It preceded a paragraph that ended with, “To the extent that the numbers are changing at all…” which is not something I would have written if I were also claiming that the number had doubled. For heaven’s sake, I didn’t even say the Gallup numbers formed a trend, I just said it was possible that they do given the steadiness of the rise. I thought all of that made it clear that the “doubling” remark was meant as the most generous interpretation of the data.

    Since you are in full lecture mode at the moment you must realize precisely how stupid your final paragraph is. The potential effect of the NA books goes well beyond the small number of people who actually read them. They continue to be discussed in virtually every media outlet that discusses cultural trends, which is a great many media outlets indeed. Numerous books on both sides continue to be published. Television sgements on chat shows are hardly infrequent. The internet is filled with videos and blogs discussing them. The books have led to atheist ad campaigns which have reached still more people. To limit the potential effect of the books just to people who read them is idiotic.

    I also very specifically did not attribute any causality to the NA’s. You just made that up, like your earlier remark about accommodationism. My point was very simple. These poll numbers are an obvious place to look for evidence of backlash against the NA’s. The evidence is not there. I asked you for tangible evidence that the NA’s are hurting the cause, and you did not provide any. Instead you told me about how the NA’s, far from being the tactical disaster so many have claimed, are actually irrelevant to this conversation.

    If you really believe that then why do people like you keep claiming they are hurting the cause?

    Also, my book was about probability theory, not statistics.

  38. #38 Josh Rosenau
    July 15, 2010

    I stand corrected on the Monty Hall book’s topic. I don’t think that changes my point, though.

    You wrote: “But with this number growing steadily over all of these polls, to the point of doubling in ten years, it is certainly possible that there is a real trend here.”

    “Certainly possible” is a good and appropriate hedge, but it comes too late. There is no “doubling in ten years,” which is put forward without any such hedge. I wish I could see how that statement could be read other than as a bald statement that the number has doubled since 1998, but … how? How do you want me to be reading that passage?

    The only basis for even intimating a possible doubling is the apples to oranges comparison of VCU’s question (about all biological life) to Gallup’s (about humans, an important distinction your post fails to mention). The legitimate, apples-to-apples comparison does not show any sort of doubling. I don’t see how to read the comment about doubling as being hedged or qualified. Nor do I understand why you chose 1998, when it is an outlier from the other Gallup numbers. Why not go to the previous poll (1996), when it was 11%, not 9? Or the 2001 poll, when it was 12%?

    As to the trend: A GLM (not incorporating autocorrelation) based on the binomial family finds no significant slope all nine Gallup polls, though models using the Gaussian do find a significant slope. I believe that the binomial is more appropriate for poll data. A nonlinear binomial regression finds no significant trend, though it gets closer.

    I’m not disputing that New Atheists have impact on the general public. I’m simply saying that lots of other things impact the public, too, and that trying, even in a handwaving way, to impute causality to one factor without any consideration of other factors is irresponsible. I don’t know how to estimate the various follow-on impact of the New Atheists (how many of those youtube videos are preaching to the converted, for instance?), and no data seem to be at hand from you or Matti. So how can we say that there is or isn’t a backlash if we can’t parse out the effects of NAs on the broader culture from everything else happening in the broader culture.

    As I explain in my blog post about this: “My concern, and what I understand Chris (and his co-authors Matt Nisbet and Sheril Kirshenbaum, at various times) to have principally argued, is that the New Atheism is likely to turn people away from folks they’d otherwise regard as trustworthy experts, and that this could hold back ongoing outreach efforts. These data don’t really allow testing of that hypothesis.”

    In other words, I don’t know that New Atheism would turn lots of people away from evolution, so much as I expect it would slow their acceptance of it, making some folks unwilling to listen to science and others unwilling to listen to some of the most knowledgeable and credible scientists. That’s not irrelevant. It’s not an immediate effect (might not show up in the 5 years since New Atheism took off), but it is a long-term risk to be wary of.

  39. #39 Jason Rosenhouse
    July 15, 2010

    I already covered your points about what I wrote. Since I was so unambiguous in saying that I was not even saying there is a clear trend here it is beyond me how you are attributing to me the definitive view that the number has doubled. You are taking one phrase out of context and getting the vapors over it. Bending over backward to be fair to your point of view, something I am not inclined to do since I am angry at you right now, I could have phrased that line more clearly. But it is very unfair of you to act as though it were my central point.

    The 1999 Gallup number is not an outlier, which would mean a data point located very far from the mean. Up to 1999 the numbers were 9, 11, 10, 9. Those are all trivial fluctuations. But starting in 1999 the number in each poll is greater than or equal to the one before. Each individual step is small and easily dismissed, but the rise is steady and ultimately outside the margin of error. That is a clear change from what had happened previously. You really couldn’t figure out why I started in 1999?

    When you write:

    I’m simply saying that lots of other things impact the public, too, and that trying, even in a handwaving way, to impute causality to one factor without any consideration of other factors is irresponsible.

    I did not, even in a handwaving way, impute causality to anything. Did you miss the part where I wrote, “I do not know the explanation for these numbers, and I have no interest in speculating.” My point, stated so clearly that I am astonished by your stubborn refusal to notice it, was that these numbers show no evidence of backlash. The first clue that that was my point was that my post was titled, “Where’s the Backlash?” The second clue was my final paragraph.

    Nor did I say there wasn’t a backlash. I said these numbers record no evidence of a backlash. Those are very different claims. There is a very common assertion out there that the NA’s are hurting the cause. Michael Ruse calls them a bloody disaster and writes, “I think that the new atheists are doing terrible political damage to the cause of Creationism fighting.” In the chapter on the New Atheists in UA, Mooney and Kirshenbaum describe the NA’s as “strongly counterproductive” and say they are effectively in league with the conservatives. They are the ones making the claim that needs evidence. Not me.

    My take on the NA’s is this. Because of their efforts atheism is part of the public conversation in ways it plainly wasn’t prior to their efforts. Atheism is far more visible today than it was a few years ago, and I regard that as inherently good regardless of whether it heralds a major realignment of public views about atheism.

    If you want something more concrete, try this. I live in one of the reddest counties in Virginia but my local Barnes and Noble now has an atheism section. It wasn’t there a few years ago! When TGSOE came out they had a huge, front of the store display for it. They didn’t do that for The Ancestor’s Tale. They did it because of the success of The God Delusion. The stuff of revolutions? Certainly not. But they are all concretely positive developments.

    I have no idea if this will ultimately raise the public acceptance of atheists, but nothing I have heard from NA critics strikes me as remotely up to the task. But people I respect, including you, keep telling me that they are hurting the cause of the science education. I take that claim very seriously. If it were true, it would dampen my enthusiasm for the NA’s. But I can find no evidence that it is true. When I ask for the evidence I only get excuses for why it hasn’t turned up.

  40. #40 Michael Fugate
    July 15, 2010

    Josh Rosenau – on a related point – do you have any evidence for the “The Clergy Letter Project” having a positive effect? Is any one following up by doing surveys in churches?

  41. #41 Josh Rosenau
    July 15, 2010

    I’m sorry that you’re angry at me, and I don’t know why you would be. I would never intentionally do something to hurt your feelings and I apologize for having done so. I’m pursuing this not out of malice, nor in self-defense of myself or of the ill-defined “accommodationism,” but because I think you are not reporting the data correctly, and I don’t want people misled.

    The biggest problem I have is your conflation of the VCU and Gallup questions without ever mentioning that one is about humans and the other about all life, even though it’s well-established that such a difference has a major effect on poll results. You acknowledge other differences in wording, but not that glaring and very important difference. I’ve raised this, as have other commenters, and you make no acknowledgment of that crucial omission.

    The only reason to even raise the prospect, however sketchily or hand-wavily, of a “doubling” is based on that misleading conflation. I keep coming back to that because it’s either a poor phrasing that you could fix in the original post with no loss of meaning and an increase of clarity, or it’s a genuine error that deserves correction and revision to any conclusions based on that error.

    You say that the data through the ’90s give no trend, which is fair enough, since there’s no statistically significant trend anywhere in the data. The jump from 9 in 1999 to 12 in 2001 is within the margin of error. The jump from 9 in 1999 to 14 in 2008 is barely outside the margin, but making that jump requires ignoring the long-term underlying average, which is bogus. Yes, it strengthens your unclaimed claim of a trend, but it isn’t rooted in any objective standard. It seems arbitrary to start from the lowest value rather than from the average through (say) 2004, when Harris published End of Faith. If it were someone I knew less well and did not respect as highly, I’d say it was cherrypicking.

    In my GLMs, the residual from the 1999 datapoint consistently stands out from the other data. That’s why I call it an outlier. If you think it isn’t an outlier, give me a data-based reason for doing so. If you’re going to restrict your analysis from 19982-1999, give me an a priori reason for that, as well.

    Writing “I have no interest in speculating” is only compelling if you don’t then speculate. And yeah, the next two sentences do speculate about what it means that “the numbers show not the slightest evidence of a backlash,” and that “they seem to be going in the right direction.”

    They are, statistically speaking, going in no direction. It’s statistically erroneous to claim, on the basis of an eyeball estimate, that there might be a trend, when a statistical analysis finds no trend.

    I understand that you think these numbers show no evidence of a backlash. I don’t know if you’re talking about the real numbers, or the erroneous account of them that you offered in your post (where VCU and Gallup are asking about the same thing, and VCU’s result hints at a possible doubling). But setting that aside, I’m not sure why these numbers and this analysis would be the best place to see the result you’re after.

    Let’s take Michael Ruse’s claim as an example. Is a decline in support for unguided evolution the only plausible interpretation of “terrible political damage to the cause of Creationism fighting”? I don’t think so. In politics, opinions matter less than intensity. Opinions, as they say, are like assholes, everyone has one. Most people, when asked, want to stop global warming, but when asked to rank their priorities, it falls well below everything else. So tell someone that fixing global warming will hurt anything they really care about (the economy, terrorism, taxes, jobs, healthcare, etc.) and support evaporates.

    It’s easy to see how Ruse’s remarks fit into that context. Maybe no one’s mind changes, but creationists get more engaged. Politicians feel like they have to pander more to creationists, passing bills and banning evolution. Even if opinion polls change not one iota, a rise in intensity of creationist feeling could have serious political consequences.

    Or you could interpret Ruse’s remarks in another context. You could take him as arguing that the rise of NAs will make things harder politically than they would have been otherwise. To test that claim, one needs either a control group, an America demographically identical but unexposed to NAs, or you need a well-validated model of public opinion to which you can compare polling after the rise of NAs. Doing otherwise is like the teabaggers who insist that the stimulus bill did nothing because unemployment is high. The valid comparison is not between today’s unemployment and unemployment before the stimulus, it’s between unemployment today and well-validated models of unemployment that do not include the stimulus. I know of no well-validated models of how public opinion on evolution would have changed absent the rise of NAs, so this hypothesis cannot be tested at this time.

    Have there been other things going on parallel to the rise of New Atheists? Why, yes! The Clergy Letter Project (est’d: 2004) kicked off Evolution Weekend in 2006. There was Kitzmiller, and Tiktaalik, and movies about Kitzmiller and Tiktaalik. There was the 2009 Year of Science and the Darwin bicentennial and Origin of Species sesquicentennial. Before 2003, there were no books by evangelical Christians defending evolution books, but there have been at least 10 since then. All of those things could counterbalance any effect you’d see from the NAs.

    You cannot talk about NAs in isolation from other societal trends. Noting those trends is not shifting goalposts, nor is it making excuses. It’s the way I would think about any other challenge in hypothesis testing. Can you tell me honestly that you’d inflate the (purported? potential? “steady”) trend by choosing an outlying datapoint as a start, then further inflate the trend by making an apples to oranges comparison of polls, ignoring crucial differences in wording poll phrasing? It is a measure of my respect for you that I very much doubt that you would do those things in any other context, and I don’t know why you’re doing it here.

  42. #42 Josh Rosenau
    July 15, 2010

    While my previous post sits in moderation, Michael, I don’t know of any follow-up studies on the CLP. I do know the clergy who signed have also organized Evolution Weekend, where they talk science from the pulpit, and that the members have been effective activists in several instances. Note also that it’s been running about as long as the New Atheist movement (started in 2004, same year End of Faith came out), so any effect from CLP could well mask some or all of the effect of New Atheists in national polling!

  43. #43 Jason Rosenhouse
    July 15, 2010

    Josh, I am angry at you because I don’t think you represented my argument correctly, either in your post or in your comments here. I’ll get over it, of course, but for now, so sorry, I remain miffed.

    As for your last comment, I regard it as mostly unresponsive to my earlier remarks. Yes, of course, it is possible that there is a backlash that is not reflected in these numbers. That is why I emphasized in my last comment that my claim is that there is no evidence of a backlash and not there is not a backlash.

    But I do think that if people are going to use strong language like “bloody disaster” and “strongly counterproductive” then the burden of proof is on them. It is not my job to prove there is not a backlash. In my last comment I pointed to what I regard as obviously good things that have come about as the result of the NA’s. If I am to believe they actually do more harm than good I will need evidence to that effect.

    Concerning the polls, you may be right that the results in the Gallup poll are statistical noise. That is why I used cautious phrasing like, “it is certainly possible there is a trend” and “to the extent that the numbers are moving at all,” that last leaving open the possibility that they are actually not moving. But the fact remains that for the last ten years the statistical noise is producing results that are all in the same direction. I find that interesting. It is different from the pattern that held in the first four Gallup polls.

    It is still unclear to me what speculation about causes you think I engaged in. Flat poll numbers are not evidence of a backlash. Period, full stop. If you want to argue that the numbers are not telling the whole story because forces working against science education are being balanced by forces friendly to it then go right ahead. In principle such an argument could be compelling, but the ones you are offering are not. Seriously, Tiktaalik and the CLP are the great counterbalancing forces to the bloody disaster of the NA’s? I think I’ll stick with my explanation, thanks. Specifically, that there is no evidence of a backlash because there is no backlash.

    Enough. If you want the last word you are welcome to it, but that is it from my end.

  44. #44 Josh Rosenau
    July 15, 2010

    I’m not arguing that there is or is not evidence of a backlash in these numbers. I don’t really care about that.

    I care that your post still incorrectly compares the VCU result to the Gallup result when there are serious reasons to do no such thing. To take the Gallup numbers, to ignore the 2005 VCU results, and to just casually slap up the claim that “Now we have 18% in the VCU survey,” as if that 18% meant the same thing as the 14% from Gallup is just wrong! I’ve said this repeatedly, and you’ve yet to even address that simple point.

    I care that you use that error as a basis for suggesting that these numbers could be “on the point of doubling.” They aren’t. You’re a serious person who takes math and science seriously, and I don’t know why you are being so flip with stuff that I learned (and taught) in intro. stats classes. If you think there’s a trend, do the statistics and show me the numbers.

    Saying “the fact remains that for the last ten years the statistical noise is producing results that are all in the same direction” means exactly nothing. It has no statistical meaning, and has no place in a discussion of statistics (such as trends in poll results). I’m confident that you know better. I’m confident you could fix these errors in your post, and I can’t fathom why you won’t do it.

    I’m not posting here, or at TfK, because I have a problem with your hypothesis. I’m posting because you’re violating basic standards of hypothesis testing, mangling basic concepts along the way.

    You might be right about the lack of a backlash. You might be right that Ruse and others shouldn’t have used the language they did. You might be right that NAs are doing a lot of good. You might be wrong. But the analysis you’ve done cannot be compelling, cannot convince anyone, because it is not a proper analysis. And because of the great and genuine respect I have for you, I expect better.

    If we can fix those errors, there’s a conversation to be had about what factors might countervail the NAs. But if you can’t even represent the data accurately, why bother?

  45. #45 Tyler DiPietro
    July 15, 2010

    “I care that your post still incorrectly compares the VCU result to the Gallup result when there are serious reasons to do no such thing. To take the Gallup numbers, to ignore the 2005 VCU results, and to just casually slap up the claim that “Now we have 18% in the VCU survey,” as if that 18% meant the same thing as the 14% from Gallup is just wrong! I’ve said this repeatedly, and you’ve yet to even address that simple point.”

    Jason included a caveat about comparing the results of the two polls, so I don’t know exactly what you’re bellyaching about. At best the point your making is a minor nit-pick that is more or less irrelevant to Jason’s main point, which is that there is no evidence of a backlash against the New Atheists. Perhaps you are a little too invested in the idea that there is for you to deal with evidence and Jason’s argument honestly.

  46. #46 Josh Slocum
    July 15, 2010

    Rosenau, you are incredible. Very simple question, very direct – where is the evidence of a backlash? Where? No speculation, no “I suspect,” just the evidence. Where is it?

  47. #47 Josh Rosenau
    July 16, 2010

    Tyler, Josh: I don’t care if there’s a backlash. I don’t think I’ve ever predicted that there would be this backlash, and I feel no need to defend a phenomenon I have not advocated.

    It isn’t obvious to me why, a priori, you’d expect NAs to drive atheists away from atheistic evolution, rather than that TEs would switch to ID or creationism, or that creationists who might otherwise have been won over to evolution have stayed creationist (the latter being harder to measure, but closer to what I worry about). Jason’s premise that the unguided evolution option is where we should find this effect seems post hoc and poorly justified. This is problematic as a matter of statistical methodology (post hoc hypotheses have to be tested differently than a priori hypotheses), and absent evidence of anti-NAs advancing that particular hypothesis, it could be a straw man.

    Since my first post, my issue has been that Jason failed to accurately describe and account for the differences between VCU’s poll and Gallup’s. Correctly characterizing the differences between VCU’s question and Gallup’s would require reworking the section beginning “Now we have 18% in the VCU survey.” That statement implies a degree of comparability that does not exist. Yes, Jason mentions that these comparisons are fraught, but ignores the biggest difference between VCU’s poll and Gallup’s. The only reason to say anything about anything being “to the point of doubling,” even cautiously, is because Jason didn’t acknowledge this simple, obvious, well-documented source of bias.

    Here’s pollster George Bishop discussing the 2005 VCU poll:

    …results resembling earlier Gallup polls on human origins, … A notable portion … thought that “biological life developed over time from simple substances, but God did not guide this process.” So, by characterizing it as “biological life” rather than “human life,” the VCU experiment may have reduced, somewhat, resistance to the theory of evolution.

    All of this goes to show how easily what Americans appear to believe about human origins can be readily manipulated by how the question is asked. As we have seen, depending on the wording of the question the percentage of apparent biblical creationists can vary from as little as 42 percent to as high as 64 percent; the percentage of theistic evolutionists or believers in “intelligent design” from as much as 41 percent to as little as 10-18 percent; and the percentage of Darwinist or naturalistic evolutionists, from as low as 10-13 percent to as high as 33-46 percent.

    …American public opinion on this matter would seem to be a lot more malleable than we have heretofore suspected.

    People like me and Jason know this, and know there is always, consistently, more acceptance of evolution when it is posed in terms of all life, than when it is in terms of humans specifically. Jason and I know this because we work with these polls regularly. He can’t assume that his readers know this, and he certainly can’t ignore or omit to mention this effect, because doing so misleads his readers who do not know the lore of creationism polls.

    This is such a well-established principle that not mentioning it as a caveat and a reason to treat that 18% as an overestimate is simply inexplicable. It’s not a “minor nitpick.” Jason comments on the difference between “God did not guide the process” and “God had no part in the process,” which wording changes have no established bias, but doesn’t mention the difference that does cause a consistent bias. Unfathomable. If it’s really so minor, why won’t he just fix the error and shut me up?

    When one looks at the VCU poll now and in 2005, you find no change in the unguided evolution option, and a jump outside the margin of error in the theistic evolution/ID option, mostly shifting from the undecided group. Jason didn’t know about or didn’t report the results from VCU ’05, which would have made it easier to trace the trends he’s interested in. Instead he tried (unsuccessfully) to draw a parallel to the Gallup poll. 2005 was before the New Atheists really took off, and works fine as a baseline for the comparison he wants to make.

    There’s no statistical evidence for any increased acceptance of unguided evolution (not with all the hedges in a shrubbery), and Jason knows better than to substitute an eyeball estimate for statistical analysis. Plus, there is evidence of a boost in the TE/ID option, which certainly doesn’t speak to any positive effect by NAs. He could probably rescue his central point by looking at VCU ’05 and by dropping the error-laden attempt at combining the VCU and Gallup polling.

    If he’d have done that, I wouldn’t have made a fuss over this post. I don’t think I ever predicted this backlash, so it doesn’t matter to me if it doesn’t exist. As it stands, this post makes claims that might be right, but that he’s getting to for reason that are erroneous and misleading. And I think it matters that people get the science and statistics right.

  48. #48 Peter Beattie
    July 16, 2010

    » Josh Rosenau:
    I’m not sure where you get the prediction of a backlash detectable in the polls (maybe someone said something at some point that works, I just don’t know and the details of the prediction would matter). It strikes me as unlikely that any effects of the “New Atheists” would be visible on short time scales.

    So you’re saying the NAs are having a negative effect, but it’s not detectable? Or are you saying the NAs don’t have a negative effect? Or are you taking Zach’s spoof line that there will be an effect, we’ll just have to wait longer?

  49. #49 Josh Rosenau
    July 16, 2010

    I’m not a sociologist, and I’m somewhat loathe to speculate about the speed and magnitude of any effect from NAs on big national surveys without that background. As I indicated above, the effect could be principally on intensity of political action, stirring up the creationist base but not changing anyone’s views and thus not changing national polls while changing the political landscape. Or, as I indicated above, the effect could be masked by other changes in the creation/evolution sphere (increasingly vocal pro-evolution evangelicals, scientific advances, pro-evolution clergy doing more to talk about this with their congregations, etc.). Looking at the data available, I don’t know how to disentangle the effects of NAs on national surveys from the many other things that happened contemporaneously with the rise of NAs. I think NAs have had relatively modest reach outside their base, which, yes, means that it will take a while to see any big, easily detectable, effect, and the effect may be obscured by other influences. This might make the effect undetectable in practice, though not untestable (it would be detectable if proper data existed).

    Look, Gallup gives us 4 datapoints since the rise of the NAs. How do you propose to do serious statistical analysis of four datapoints? And that’s the survey with the highest data-density. We could toss in VCU and Pew and NORC and other surveys, but then we have to find a way to account for house-effects, etc, and smoothing out those differences in the context of such low data-density will mean that those corrections could wipe out any small signals that would emerge.

    I would like this to be easy to test, but it isn’t. Rather than looking at national polls, which can miss shifts within small subpopulations, I’d think that it would be useful to do lab work, and to look at the broader literature on communications. Daniel Loxton did a nice roundup of a few useful studies in this realm, and Mike McRae looked at a wider sampling in the context of the “don’t be a dick” discussion.

    Someone grounded in that body of research could develop some testable hypotheses about how folks might respond to NAs. Then you could do lab work, bringing in a large and representative sample of folks with views across the c/e spectrum. Do a pretest, then have them read a selection from TGD, then do a post-test. Follow up a month later, and see how their views on science generally, evolution specifically, and the relationship between science and religion have changed. Follow up a year later.

    That’s science, and I’d be interested in the results. What isn’t science, and what I’m not interested in, is slapping together disparate polls and ignoring significant and well-documented differences between two polls, then cherry-picking a starting point for the comparison, all to construct a narrative. As I say, Jason could be right about there being no detectable backlash, but the analysis above misrepresents the polls and flubs basic statistical principles, and thus could not possibly demonstrate anything at all. If he fixed those errors, it might demonstrate something, and we could talk about the result. But now, the methodological problems are too overwhelming for such a discussion.

  50. #50 Jack Weiss
    July 16, 2010

    Josh’s statistical analysis is clearly bogus. When you plot the data there is an obvious trend, which, unless there are hidden problems with the data, any sensible analysis should be able to detect. Josh is right to say that the poll data are intrinsically binomial, but in order to analyze them as such you need to know what the binomial denominator is, i.e., how many people were polled. Binomial data are always the number of “successes” out of a fixed total N. If you don’t know N you cannot analyze them as binomial data.

    The linked Gallup data do not report N. In lieu of knowing N (but see below) we can treat the percentages as the response variable. Probabilities if bounded away from 0 and 1 can be treated as normally distributed. Alternatively a logit transform can be used and the result can be assumed to be normally distributed. This is equivalent to assuming that the probabilities have a logistic-normal distribution. (There are other approaches but they all yield essentially the same results for these data.)

    In truth Gallup typically surveys 1000+ randomly selected individuals. (See e.g., http://media.gallup.com/PDF/FAQ/HowArePolls.pdf). If we assume N=1000 in each of the sampled years then it is legitimate to use a binomial probability model with a logit link.

    The linked figure compares these three approaches and superimposes the results on a scatter plot of the raw data on a probability scale. All three models yield statistically significant trends that are virtually indistinguishable from each other when graphed. I’ve superimposed a nonparametric smoother for comparison as a graphical assessment of the linearity assumption.

    It’s worth noting that the rounded percentages reported by Gallup sum to less than 100 in three of the nine years. I ignore this problem in the above analysis, but correcting for it in obvious ways has no effect on the results. One can argue whether the observed trend in the belief in human evolution is of any substantive importance, but there is no argument as to whether it is statistically significant. It is!

    R code for the analysis.

  51. #51 Peter Beattie
    July 16, 2010

    Josh, you’re belabouring a peripheral, actually very probably irrelevant, point. Jason’s point, as he has pointed out to you and as is readily apparent from the headline and the closing paragraph, is that those in the accommodationist camp have never shown any evidence for their claims that the Spiffy New Atheists’ approach is destructive, counterproductive, or otherwise negative in some way or other. One possible place to look for such evidence would be the poll numbers Jason is talking about. And it is about those that he makes one simple observation: there is no consolation in them for the accommodationists. He is saying: there is, other things being equal, no prima facie evidence of negative NA effects in these numbers. I.e., we’re still waiting for you to produce something that might make your assertions debatable.

  52. #52 Josh Rosenau
    July 16, 2010

    Jack Weiss:

    Thanks for posting your code. I see where I went awry in my coding (I converted the percentages to decimal form and treated it as a logistic regression, and yes, something about that felt odd). There’s a significant trend in the unguided evolution line. Thanks for the correction.

    Two things about that still stand out (and I noted these before, so the goalposts are constant!): a) this trend began well before the rise of the NAs, so there’s no evidence that NAs are having any positive effect, and b) there’s no reason a priori to truncate at 1999, as Jason did. 1999 still comes out as a borderline outlier, making it doubly dubious to truncate there. I also am not sure what it should mean that there’s a significant trend here but no significant trend in either of the other answers. One would expect that a significant rise in acceptance of unguided evolution would take support from something else, and it doesn’t. Something odd is going on.

    Peter: It’s not peripheral, because the data Jason is trying to use here are a conflation of two different polls, and he’s ignoring an established and well-documented source of bias between those two polls. That’s why it matters that he’s ignoring the shift from human evolution in Gallup to evolution of all life in VCU.

    There’s no reason to take the 18% in VCU as directly comparable to 14% or 9% in Gallup. The wording issues he does talk about are not known to introduce a bias into these sorts of comparisons (they’d reduce precision in such comparisons, but not necessarily accuracy) while the wording differences he is still ignoring introduce errors in accuracy, but not necessarily in precision. So he ought to be able to correct that bias by controlling for this known effect. But he doesn’t, and doesn’t flag that known bias for his readers.

    The premise of the post is that VCU’s numbers and Gallup’s are comparable in some way. They aren’t. That’s not peripheral or irrelevant.

    Furthermore, no evidence has been given that any critic of New Atheists has ever predicted that NAs would have a measurable effect on these sorts of national surveys. As I point out above, there are other effects that I would think are more likely, which are consistent with the comments Jason cites as his basis for this analysis, and some of could be tested without too much trouble (though I, not a sociologist, am not in a position to do so). Since I haven’t made the prediction Jason is testing, it isn’t fair to call it “my assertion[]” nor do I feel the need to “produce something” to make it debatable.

    To quote my old Ecological Methods textbook: “Before-after comparisons are statistically powerful… But unless there is a contemporaneous control, all before-after comparisons must assume homogeneity over time, a dubious balance-of-nature model that has been found invalid time and time again.”

    In good conscience, I could not claim that these sorts of polls showed a “backlash” unless I had such a control, presumably a statistical control based on a well-validated model. I haven’t got such a model and don’t think one exists, thus I don’t think these polls are the right place to look for evidence either way.

    I do, however, think that scientific, statistical, and mathematical integrity obliges an author to acknowledge significant sources of bias in any dataset they use. Jason did not do that in his OP, and that’s been the focus of my comments throughout.

    BTW, in comments at my blog, a reader notes that I’d mistranscribed the data from VCU in the graph I made. I’ve fixed the error, and doing so eliminates the rise in “guided evolution” between ’05 and ’10 that I’ve mentioned in comments here. Please disregard those comments.

  53. #53 James Sweet
    July 17, 2010

    I think this just goes to show how pointless it is to use guarded language and careful reasoning… Would Josh’s response have been any different if Jason had concluded the post with, “And this proves that there is no backlash. Eat it, Michael Ruse!”? If people are going to interpret you as overstating your case anyway, you might as well overstate it to begin with.

    And this one interaction between Josh and Jason proves it!

  54. #54 Shirakawasuna
    July 17, 2010

    Thank you, Josh, for solidfying my opposition to your evidenceless claims against New Atheism. Before this exchange, I was eager to think that you simply disagreed and were going off of anecdotal personal experience (because the evidence sure wasn’t there). Now, it’s clear you have some kind of ideology to defend, going on almost incoherent rants to dither about anything but the point being made.

    If it’s a scapegoat you want, there are plenty of other minority positions which offer more convenient scapegoating and have actual statistics. If you’re interested in a consultation my services are offered pro bono.

  55. #55 Aquaria
    July 18, 2010

    You know, I’d find the accommodationists less despicable if they exerted any of their pearl-cluthching hysterics for the creationists and other religion poisoned idiots who really are making science less accepted. But they don’t. Like the Step-n-Fetchits and Uncle Toms they are, they suck up to power and the status quo like the sniveling cowards they are.

    These smug little toads have zero gratitude for the wiggle room the New Atheists give by pulling the Overton Window further to the science side.

    Nope, they work with the fundie nutbars to pull it further to the anti-science side. They will lie, distort and misrepresent New Atheists, then have the gall to shriek, Wah! Look at the meanie New Atheists! Wah! Bad words! Wah! Incivility! Wah! How does that help?

    What a bunch of cowards, backstabbers and drama queens, all rolled into one.

  56. #56 Josh Rosenau
    July 18, 2010

    Aquaria: I work full-time fighting creationists. That is literally my day job. My blog is full of posts written in my personal time doing the same. Which is to say, you don’t know what your talking about.

    Which is fine. Lots of people don’t know what they’re talking about, and they come to Scienceblogs to learn. Which is why I’m disappointed that, 4 days after first Hamilton Jacobi and then I pointed it out, Jason’s post still wrongly asserts the direct comparability of the VCU poll and Gallup’s, ignoring the known bias introduced by asking about evolution of life vs. humans. People who come to this blog not knowing what they’re talking about will go away from it thinking they’ve learned something, but having been misinformed. I’ve been quick to correct errors in my posts, even when those errors don’t affect the central point I’m making, because I don’t want people misinformed. I don’t know why Jason has yet to correct the several errors in this post (direct comparison of Gallup and VCU, erroneous assertion of a trend “to the point of doubling in ten years”, unjustified truncation of the Gallup data which overstates the trend, failure to note that VCU’s polling shows no trend, attribution to “accommodationists” of a hypothesis about a backlash in national polling that none of them seem to have made – at least not about national polls).

    Here’s what he could have written:

    Jerry Coyne has a post up … [introduction, quotation] … The option, “biological life developed over time from simple substances but God did not guide this process,” was selected by 18% of the people. That’s in line with results from an identical question VCU asked 5 years ago.

    Then I remembered the numbers from the long-running Gallup poll on a similar question. In Gallup polls the option is phrased, “Human beings have developed over millions of years from less advanced forms of life, but God had no part in this process.” Mentioning humans tends to decrease support for evolution, so we expect Gallup’s numbers to be lower, and they are.

    Nonetheless, the percentage of respondents selecting this option has grown slowly but steadily since the question was first asked in 1982. It was 9% in 1982, 11% in 1993, 10% in 1997, back to 9% in 1999, then 12% in 2001, then 13% in 2004 and 2006, then 14% in 2007 and 2008.

    Comparing VCU’s results (17% in 2005, 18% in 2010) to Gallup’s polls is fraught with peril, since it is very difficult to capture people’s religious beliefs with simple poll questions. For example, the phrasing in the VCU question refers to “biological life” rather than humans, it refers to evolution from “simple substances,” and it says simply that God did not guide the process. Each of those could introduce a bias. For instance: a person could well believe in God and still select the “did not guide” option. The Gallup phrasing (“God had no part in the process”) is far more stark. What if you believe God created the initial conditions in which evolution unfolds? Doesn’t that constitute having a “part in the process”? Similarly, referring to “biological life” focuses respondents on the science of biology, rather than theology, and takes focus away from humanity. The latter tends to cause a 5-10 point increase in acceptance for evolution, and focusing people on science tends to boost their acceptance of evolution, too.

    But still, there is a statistically significant rise in acceptance of unguided evolution in the Gallup poll, though not in the VCU data.

    I do not know the explanation for these numbers, and I have no interest in speculating. [could end the post right here, but…]

    Someone interested in understanding the role of New Atheism, of accommodationism, and of other science education efforts, should take a hard look at these numbers before claiming their ideas work or someone else’s are harmful. I’ll note that there seems to be no evidence in national polls for any backlash against New Atheism’s insistence that people understand evolution being in contrast to, and not compatible with, theism. To the extent that the numbers are moving at all, they seem to be going in the New Atheist’s direction, and have been since before New Atheism was given that name. There are many ways this could be explained, and there is not enough resolution in these polls to measure all the effect any of the groups active in this effort, but if accommodationists want to claim New Atheists are doing harm, they should address the growing acceptance of unguided evolution.

    If he’d written that, I would have no complaints. I think it captures the meaning Jason was after, it’s only slightly longer, it doesn’t misrepresent the polls, and it models good practices for reading polls. At least right now, the original post fails on those last two counts. The post could then inspire a thoughtful and informed discussion of what the various parties expect to find in these national surveys, and could generate some useful hypotheses to test, and maybe even inspire research.

  57. #57 Shirakawasuna
    July 18, 2010

    “Which is why I’m disappointed that, 4 days after first Hamilton Jacobi and then I pointed it out, Jason’s post still wrongly asserts the direct comparability of the VCU poll and Gallup’s, ignoring the known bias introduced by asking about evolution of life vs. humans.”

    And this is why I’m disappointed that, 4 days after you three pointed that out, you’re still missing the point of this post despite having it explained to you several times. People are explicitly telling you that you missed the point, that your claim is tangential, etc, and you even had the gall to write your own blog post accidentally explaining how much you miss the point. This is something I’ve seen many times in discussions about New Atheism (or framing, of course), but this is one particularly silly example and one where I really can’t give you the benefit of the doubt.

  58. #58 Owlmirror
    July 18, 2010

    If people are going to interpret you as overstating your case anyway, you might as well overstate it to begin with.

    And this one interaction between Josh and Jason proves it!

    “I’m generalizing from one example, here, but everyone generalizes from one example.

    At least, I do”

    (/Steven Brust via Vlad Taltos)

  59. #59 Sharl
    July 19, 2010

    Woops, neever mind my last comment attempt, since the comment it referenced has been deleted (appropriately so, IMO).

    No need to publish this one either, obviously. That’ll larn me to first refresh an “old tab” before reading/commenting on it.

  60. #60 benjdm
    July 20, 2010

    An anecdotal story of the new atheists NOT hurting science:

    http://blogs.orlandosentinel.com/features-the-religion-world/2010/07/18/in-praise-of-athiests/

    The End of Faith by Sam Harris and The God Delusion by Richard Dawkins made me squirm. I am, after all, a preacher. My wife and I have addressed more than a thousand audiences as we’ve crisscrossed North America these past eight years evangelizing evolution. We initially steered clear of questioning the widespread practice of seeking guidance and inspiration in ancient texts. Now, however, we see the absurdity of revering unchanging holy books while modern discoveries take a backseat or have no seat at all.

    And so my Sunday sermons now begin, “Today’s scripture reading is from cosmologist Carl Sagan.” (Pause for laughter.) “Science is, at least in part, informed worship.”

The site is currently under maintenance and will be back shortly. New comments have been disabled during this time, please check back soon.