Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Survey: Churchgoers more likely to back torture

  1. Apr 30, 2009 #1

    djeitnstine

    User Avatar
    Gold Member

  2. jcsd
  3. Apr 30, 2009 #2
    No kidding

     
  4. Apr 30, 2009 #3
    I am not surprised and my reasoning beyond that statistics is that:
    1) Majority Religious people might be conservatives
    2) Religious people have evil definition and they might be supporting "Us against them"
    3) Majority liberals might be non religious
    4) God tells them that it's okay to kill your enemies (witches/etc)
    It can be either all of them or none of them.

    Religions are the reflection of ancient societies.. I am not against them but I also don't have blind faith in them. They will hardly ever meet with the present needs.


    But, one thing to point out:

     
    Last edited: May 1, 2009
  5. May 1, 2009 #4
    Party affiliation.

    “The nationalist not only does not disapprove of atrocities committed by his own side, but he has a remarkable capacity for not even hearing about them.” -George Orwell
     
  6. May 1, 2009 #5

    Vanadium 50

    User Avatar
    Staff Emeritus
    Science Advisor
    Education Advisor

    I'm amused that people are drawing conclusions from this - particularly conclusions that I suspect support what they "knew all along". RootX touches on the key point: 742 people were tested. If 8 of them change their answer from "sometimes" to "rarely" (or the other way), the purported effect goes away.

    Everything depends on the answers of those 8 people.
     
  7. May 1, 2009 #6
    The question was about torture of suspected terrorists:

    "The survey asked: "Do you think the use of torture against suspected terrorists in order to gain important information can often be justified, sometimes be justified, rarely be justified, or never be justified?"

    This complicates matters greatly. What kind of "important information" do people imagine the question pertains to? For Americans right now it probably calls to mind an impending attack on the scale of 911: "Is it OK to torture a few people if that means a few thousand would be spared getting rammed by a hijacked airliner? What about sparing the grief of the many more people represented by their surviving families?" 54% of churchgoers called up a mental scenario such that they agreed it could be justified and 42% of non-churchgoers called up a scenario such that they agreed to it.

    The main trouble is that it's a fairly vague hypothetical question that will never actually neatly fit any real world situation. To the extent anyone gives an answer it is an answer to the rohrschach scenario they happen to imagine in response to the question. It could be, I'm not saying it is for sure, that churchgoers, being family oriented, do a lot more brooding over the thought of wives whose husbands were suddenly blown away, and children whose fathers suddenly stopped coming home.

    So, imagine yourself in an emergent scenario: you're some churchgoing small town sheriff who happened to stop and arrest a fleeing terrorist. You called the FBI and they're on their way, but in the meantime you know there's a hijacked jet in the sky planning on crashing into some densely populated area. You have the hijacker's accomplice in the back room, but he's just grinning and spitting at you.

    "Time's a wastin', Clarice. Tick toc tick tock."
     
  8. May 1, 2009 #7

    Hurkyl

    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    To rephrase -- you know that the numbers are inadequate to draw any conclusions, yet you still wish to draw conclusions. :tongue: It sounds like you're just looking to bash religion....
     
  9. May 1, 2009 #8

    Astronuc

    User Avatar

    Staff: Mentor

    Seems like a very small sample.

    From the United Methodist Church of which my parents are members:
    http://www.umc-gbcs.org/site/c.frLJK2PKLqF/b.5128831/k.521B/United_Methodists_Do_Not_Torture/apps/ka/ct/contactus.asp?c=frLJK2PKLqF&b=5128831&en=cuJTK5NLJmISL1PTIhJMK9OSKnI0I5MMJfIUIbPXKuJbF [Broken]

    National Religious Campaign Against Torture
    http://www.nrcat.org/index.php

    http://www.pcusa.org/peacemaking/
    http://www.pcusa.org/peacemaking/actnow/actnow.htm#stoptorture [Broken]

    I'm quite sure the majority of Unitarian Universalists oppose torture as well as war.
     
    Last edited by a moderator: May 4, 2017
  10. May 1, 2009 #9

    LowlyPion

    User Avatar
    Homework Helper

    It seems to me to be a weak correlation. The sample size is large enough I suppose to come to a meaningful conclusion, but I don't think the skew in responses between the extremes of non-church going to frequent attendee is sufficient to characterize it as anything but a potential tendency, certainly not a strong enough conclusion to warrant the article.

    And this takes no account I think of the bias carried in the question itself.

    Maybe ask people if they think that enhanced interrogation performed on untried detainees denied legal rights is fair and reasonable?
     
  11. May 1, 2009 #10

    DaveC426913

    User Avatar
    Gold Member

    Perhaps, but one can do that with any set of statistics. The question is: is the study rigorous enough to be valid?

    If the answer is yes then the results (even if it's just 8) is statistically significant.
     
  12. May 1, 2009 #11

    Gokul43201

    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    I second phrak's opinion on this. I think this is more about party affiliation, and the tendency to condone or denounce a position taken by someone that you favor or reject. The correlation is stronger when directly tabulated against party affiliation1.

    1. http://people-press.org/report/510/public-remains-divided-over-use-of-torture

    So what? What does it matter that "everything depends on ... 8 people"? It's completely irrelevant, without analyzing the statistics, and moreover, unless I'm making a mistake below, it also appears to be wrong.

    Of the 336 religious respondents, 54% chose 'often' or 'sometimes'. While of the 186 hardly religious respondents, 42% chose 'often' or 'sometimes'. For the "purported effect" to go away (just between these two groups; there's also a third in-between group), 12% of either group (or some combination making up the 12% difference) - that's 40 religious people or 23 of the other group - will have to switch sides. Where did you get the number '8' from?

    On samples sizes2 of 336 and 186 respondents, the error bars at the 95% confidence level are +/- 6% and +/-8%. For the "purported effect" to go away, you are asking for people to switch sides, which we know will not happen with a confidence level in the high 90% range.

    2. http://people-press.org/report/?pageid=1506 [Broken]

    Are you (the sheriff) churchgoing, or is the town? And you (the sheriff) can have only arrested a suspected terrorist. And then, you should multiply the probability of your responses to this emergent scenario with the frequency with which you expect to encounter such a scenario (and accordingly for any other scenarios you can think of) to determine if your final answer belongs in the 'sometimes' (or 'often') category or the 'rarely' (or 'never') category.

    True ... if by 'rephrase', you mean 'rewrite without regard to the meaning conveyed'. djeitnstine only said that the sample size is "not large", not that it was inadequate to draw conclusions. In this case, you can not draw the conclusion stated in the thread title at the 3sigma confidence level, but you can draw it at the 2sigma (around 95%) level.
     
    Last edited by a moderator: May 4, 2017
  13. May 1, 2009 #12

    Vanadium 50

    User Avatar
    Staff Emeritus
    Science Advisor
    Education Advisor

    8 people only have to "switch sides" from "sometimes" to "rarely".

    But if you like, take the raw data from the Pew Forum. Looking at the graphics for the case where they compare the torture opinion to the churchgoing rate allows one to make a 3x4 table of number of respondents. One can then calculate the Chi-square of the experimental table under the assumption that there is no correlation between one's opinion on torture and one's rate of churchgoing. I get a chi-squared of 8.5 for 6 degrees of freedom.

    That's not enough to reject it at the 90% confidence level, much less 95% or 99%.

    The 8 is saying the same thing, just another way: ~700 people (some declined to answer) and 12 bins means you have ~60 people per bin. A change of 8 in a counting experiment of 60 is only a one standard deviation effect.
     
  14. May 1, 2009 #13
    It may also depend on what each person considers to be torture. I think it's relevant that conservatives, who tend to have a religious faith, would be more inclined to beat the snot out of someone if they believed they had information about an impending act of terrorism. I know I wouldn't have an issue with it.
     
  15. May 1, 2009 #14
    You're right. My scenario is nonsense in any case where it's stipulated we only have a suspect.
     
  16. May 1, 2009 #15

    mgb_phys

    User Avatar
    Science Advisor
    Homework Helper

    It also depends on how the question was phrased - and what conclusion the questioner wants to draw from it:

    Q: Do you believe the US should be allowed to kill terrorist suspects.
    Turns into either "Majority calls for US military presence in Afghanistan" or "Majority support concentration camps"

    Or from the "Yes minister" guide to polls:

    Q: Are you worried about the rise in crime among teenagers?
    A: Yes.
    Q: Do you think there is lack of discipline and vigorous training in our Schools?
    A: Yes.
    Q: Do you think young people welcome some structure and leadership in their lives?
    A: Yes.
    Q: Do they respond to a challenge?
    A: Yes.
    Q: Might you be in favour of reintroducing National Service?
    A: Yes.

    Q: Mr. Woolley, are you worried about the danger of war?
    A: Yes.
    Q: Are you unhappy about the growth of armaments?
    A: Yes.
    Q: Do you think there's a danger in giving young people guns and teaching them how to kill?
    A: Yes.
    Q: Do you think it's wrong to force people to take arms against their will?
    A: Yes.
    Q: Would you oppose the reintroduction of conscription?
    A: Yes.

    Sir Humphrey Appleby: There you are, Bernard. The perfectly balanced sample.
     
  17. May 1, 2009 #16
    You seem to be implying that being statistically significant would mean that it would be reasonable to assume the headline is accurate, which indicates a misunderstanding of the meaning of "statistical significance," which is simply an arbitrary thresholding on the difference between two sample means, and merely indicates whether or not those means come from the same distribution or not. In other words, it has nothing to do with the validity of making a generalization based on the sampling.

    A statistical inference is only valid to the extent that the sample is representative of the entire population to which the result is generalized. If the sampling is not representative of the population, then being statistically significant, or insignificant, is irrelevant.

    Vanadium's point was that the opinion of 8 people were enough to change the results in the sample size, which makes it intuitively obvious that the sample size was not large enough to be representative of the American (or world) population, which involves hundreds of millions or billions of people.
     
  18. May 1, 2009 #17

    Gokul43201

    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    From what group?

    That's a confidence level of about 80%. So you can reject (at the 90% level, or even the 81% level) the hypothesis that religious regularity is correlated to the tendency to choose between one of the 4 choices for allowability of torture. But let's say you're not looking to be able to determine the finer shades and nuances in opinions and want to look only at the coarser tendency.

    So, if instead, you club 'often' and 'sometimes' into one group (those that believe torture is justified at least sometimes), and put 'rarely' and 'never' into another group (those that believe torture is justified no more often than rarely), then you get a table with 6 cells and a chi-squared value of 4.99 with 2 DoF. With this test, the null hypothesis that the choice between allowing torture more often than "sometimes" and less often than "rarely" is independent of religious regularity can be rejected at a 92% confidence level.

    If further, you throw out the middle group of not-so-regular churchgoers, and rewrite your hypothesis to test only the difference in tendencies of the two extremes of the churchgoers (the very regulars and the nearly nevers), you can reject the independence at the 97% confidence level.
     
  19. May 1, 2009 #18

    Gokul43201

    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    I am not possessed of such a strong intuition that this is obvious to me. Can you provide a rigorous proof for this intuitively obvious truth?
     
  20. May 1, 2009 #19

    Hurkyl

    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    I think you're putting words in his mouth.
     
  21. May 1, 2009 #20
    http://www.socialresearchmethods.net/kb/statinf.php

    http://cnx.org/content/m10185/latest/

    http://owl.english.purdue.edu/owl/resource/672/05/

    See pages 13 and onward for ways to determine the minimal sample size necessary for the results to be generalized to a larger population.

    http://books.google.com/books?id=0k...=X&oi=book_result&ct=result&resnum=8#PPA13,M1

    Without even doing the math do you honestly think that a sample size of 742 is large enough to generalize to a population size of 300,000,000 ?
     
Know someone interested in this topic? Share this thread via Reddit, Google+, Twitter, or Facebook




Similar Discussions: Survey: Churchgoers more likely to back torture
  1. Ear torture (Replies: 17)

Loading...