# Do the majority of Americans take the Noah's ark story from the Bible literally?

An ABC News poll released Sunday found that... ...Sixty percent believe in the story of Noah’s ark and a global flood

This claim is often repeated, including in a number of lists such as The 10 Most Ridiculous Things People Believe, and I recall this or similar statistics being noted in Bill Maher's Religulous.

Is it possible, in spite of the current scientific knowledge, that 60% of the 307 million Americans – some 184.2 million people – actually believe the Biblical account of Noah's ark to be literally true?

• Take a look at Christianity.SE for a very subjective impression, there are quite some people there that believe in a literal flood. Sep 19 '11 at 14:25
• I'd just like to post this from the article: "The poll, with a margin of error of 3 percentage points, was conducted Feb. 6 to 10 among 1,011 adults" 1011 people does not a representative sample make. Sep 19 '11 at 15:14
• @Darwy well it says it has a margin of error of 3%. So the adult creationists are between 58% and 64%; those figures would make absolutely no difference to the question. 1011 people would be a large enough sample to generate that level of error. Sep 19 '11 at 20:36
• @Darwy: 1011 people may or may not be a representative sample, depending on how they are drawn from the population, but if it is... If a coin is 60% fair and you toss it 1000 times, the number of heads is a binomial distribution with mean = 600, sigma = sqrt(.6*.4*1000) = sqrt(240) = 15, which is 1.5% of 1000. If you want a 95% confidence, you would take +/- 2 sigma, or +/- 3%. So that's roughly what a sample of 1000 can tell you. If you go to 100 000, you can shrink the uncertainty by a factor of 10. Sep 20 '11 at 1:40
• @Darwy You have a misconception about random polling. The sample size is plenty. Furthermore, the margin of error is already given at 3%, as others have noted. Making the sample size larger would make this margin smaller, but not by very much. The only bias we need to worry about is non-sampling bias – but once again, that would be scarcely affected by a larger sample. Sep 20 '11 at 14:20