Four in 10 Americans Believe in Strict Creationism?

[vortexring 0]

I came across this poll by Gallup:

http://www.gallup.com/poll/145286/four-americans-believe-strict-creationism.aspx

I was bit surprised, especially as in 1993/97 it was 47%. But my issue here isn't my surprise or what Americans believe in, but on the accuracy of the poll.

They interviewed 1019 people in a country with a population of 300 000 000+ and claim the accuracy is within + or - 4%?

Do other recent polls agree with the Gallup one?

Do Americans here on SC generally agree the poll to be a fairly accurate measure of their countries religious beliefs?

Edit: So any opinions on the actual poll and the number of Americans believing in creationism will be greatly received, as opposed to arguments over whether you believe in creationsim or not.

Thanks

'for it's Tommy this, an' Tommy that, an' "chuck 'im out, the brute!" But it's "saviour of 'is country" when the guns begin to shoot.'

[RonD1120 62]

I took statistics in another lifetime and wasn't a very good student. However, isn't there some formula for determining the adequate sample size for a given population?

Look for the shiny things of God revealed by the Holy Spirit. They only last for an instant but it is a Holy Instant. Let your soul absorb them.

[shropshire 0]

It maybe that the 'Pollsters' were trying to find the single most stupid members of a community ...

If that was indeed their goal, it seems to me that they almost succeeded in their endevours.

(.)Y(.)
Chivalry is not dead; it only sleeps for want of work to do. - Jerome K Jerome

[wmw999 2,602]

There is, however the sample has to be truly random. I'm not sure that a random enough sample on something so complex can be generated with that few people.

You probably could on uncontroversial binary things like "do you dye your hair," but evolution has a lot of flavors.

Otherwise you end up with ridiculous polls like "are in in favor of or against abortion?" What's wrong with it? Well, there are lots of steps in that poll -- what about to save the life of the mother? What about in the case of rape, incest, etc? What about a woman who simply exercises a lot when she knows that she has an incompetent cervix (i.e. guaranteeing a miscarriage).

Simple polls on complex subjects are stupid in the first place.

Wendy P.

There is nothing more dangerous than breaking a basic safety rule and getting away with it. It removes fear of the consequences and builds false confidence. (tbrown)

[Andy9o8 3]

Quote

Simple polls on complex subjects are stupid in the first place.

J'ever notice there are lots of polls in Speakers Corner?

[BIGUN 1,507]

What a horrible methodology... one would expect more from Gallup.

Nobody has time to listen; because they're desperately chasing the need of being heard.

[Remster 30]

Quote

What a horrible methodology... one would expect more from Gallup.

What do you base that on? After reading their methodology and sampling technique, I feel it's an adequate poll.

Remster

[jclalor 12]

Quote

Quote

Quote

Simple polls on complex subjects are stupid in the first place.

J'ever notice there are lots of polls in Speakers Corner?

The ones that fit my agenda seem to be pretty much right on the mark.

[sacex250 0]

Quote

Quote

What a horrible methodology... one would expect more from Gallup.

What do you base that on? After reading their methodology and sampling technique, I feel it's an adequate poll.

The poll's methodology is a joke! It's not a scientific survey, period!

It's all been said before, no sense repeating it here.

[Remster 30]

Quote

Quote

Quote

What a horrible methodology... one would expect more from Gallup.

What do you base that on? After reading their methodology and sampling technique, I feel it's an adequate poll.

The poll's methodology is a joke! It's not a scientific survey, period!

Again. That's what you claim. Saying "it's a joke" doesnt say why you think it is. Care to explain why you think it's a joke?

Remster

[jclalor 12]

Now there's a shocker!!! those with the lowest education level (on average) believe in creationism.

[sacex250 0]

Quote

Quote

Quote

Quote

What a horrible methodology... one would expect more from Gallup.

What do you base that on? After reading their methodology and sampling technique, I feel it's an adequate poll.

The poll's methodology is a joke! It's not a scientific survey, period!

Again. That's what you claim. Saying "it's a joke" doesnt say why you think it is. Care to explain why you think it's a joke?

First of all, if you read the methodogy information on the website it doesn't even make sense.

From what I can tell they had a total sample of 1,019 respondents of which at least 850 were contacted by landline phone and at least 150 were contacted by mobile phone. The overall sample is too small.

Then they took that too small sample and broke it down by gender, age, race, education, region and phone lines.

So assume for a moment that gender was split 50/50, that means that approximately 510 men/women were asked to speak for all of the men/women in the United States, That makes it a joke right there - "95% certain of a +/- 4% error"? Yeah, right! It sounds like they're just giving the margin of error for the sample group not the population.

Now apply the same thing to race which wasn't even reported. How many whites, blacks, hispanics, native americans? The sample sizes are getting smaller and more useless which is probably why they weren't reported.

What about region? Could they have contacted any more than 200 people in each region? Could you do an effective poll in any one state with a sample of just 200 people, let alone an entire region? Would it really be that difficult for a college student doing a survey of fellow students to come up with 200 respondents regarding the college parking policy?

It's all been said before, no sense repeating it here.

[Remster 30]

Quote

The overall sample is too small

You base that on what? Your gut feeling truthiness?

A sample size of 600 is actually sufficient for their confidence levels and interval.

Take some stats classes.

Remster

[sacex250 0]

Quote

Quote

The overall sample is too small

You base that on what? Your gut feeling truthiness?

A sample size of 600 is actually sufficient for their confidence levels and interval.

Take some stats classes.

Yeah, you got me there on the size of the whole sample. Did you bother to read the rest of my post?

They would still need a sample of 600 respondents for each subgroup. That means at least 600 men and 600 women - they didn't have both.

Did they have 600 people from each region?

Did they have 600 people with postgraduate degrees?

Did they have 600 people who attend church weekly?

Did they have 600 Republicans, 600 Democrats, and 600 Independents?

The overall sample of the poll is large enough as a whole, but as soon as it gets broken down into subgroups the smaller sample size creates a larger margin of error.

To quote Gallop, "For results based on the total sample of national adults, one can say with 95% confidence that the maximum margin of sampling error is ±4 percentage points."

The subgroups most likely have a much larger margin of error, which means that comparisons between the subgroups are going to be less accurate.

It's all been said before, no sense repeating it here.

[kelpdiver 2]

Quote

From what I can tell they had a total sample of 1,019 respondents of which at least 850 were contacted by landline phone and at least 150 were contacted by mobile phone. The overall sample is too small.

The fact that they got some cell phones already corrected for the most common error of phone based surveys - a bias against younger people who are more likely to not have a land line.

The sample size is sufficient, your beliefs not withstanding. But yes, the margin of error is greater for the sub categories, though the trend lines seem pretty clear.

This is basic statistics - get the intro textbook for more information. Random sampling is far more important than sample size.

The only fault I see in it is that the question forces Christians to choose between strict creationism (God did everything in last 10000 years) versus God created evolution and it took a really long time. I would think this forced many to pick strict even though they don't believe it was a recent event. I'd prefer the question not have the 10000 year bit, and just ask if they believe God created the world with animal variety pretty much as it is now. (not sure how the Dinosaurs fit, but you can't cover everything here).

[sacex250 0]

The funny thing is I agree with the poll's conclusions, although it shouldn't take a rocket scientist to figure it out.

People who attend church regularly are more likely to believe in Creationism. Well, duh!

People with more education are more likely to believe in evolution. Well, duh!

I also agree with you about the possible responses. What would an Agnostic who believes that the creation of the universe is outside the comprehension of humans answer? Other? It could theoretically be evolution with or without God.

It's all been said before, no sense repeating it here.

[Calvin19 0]

Quote

Four in 10 Americans Believe in Strict Creationism

And about that percentage in the US as well as on this forum seem to think that the three WTC towers were hit by holograms ten years ago today. (You're all the same on the inside).

I call it hopeless.

[BIGUN 1,507]

Quote

Quote

What a horrible methodology... one would expect more from Gallup.

What do you base that on? After reading their methodology and sampling technique, I feel it's an adequate poll.

They lost me at RDD.

Nobody has time to listen; because they're desperately chasing the need of being heard.

[rhaig 0]

It's been 20 years since I took a stats class, but we talked about polls and I remember someone asking why many polls listed +-3% as error margin. The answer was because of the diminishing returns beyond 1500 sample size where the error is +-3% based on some formula I forgot 19.5 years ago. So with 1019, I'm guessing that formula spit out +-4%.

+1 to what everyone else said about complexity of the issue and random sample.

--
Rob

[kallend 2,183]

Quote

It's been 20 years since I took a stats class, but we talked about polls and I remember someone asking why many polls listed +-3% as error margin. The answer was because of the diminishing returns beyond 1500 sample size where the error is +-3% based on some formula I forgot 19.5 years ago. So with 1019, I'm guessing that formula spit out +-4%.

+1 to what everyone else said about complexity of the issue and random sample.

Indeed. Sampling error depends on the size of the sample, NOT on what fraction of the total population the sample represents. (always assuming that the sample is not chosen with some bias, of course).

...

The only sure way to survive a canopy collision is not to have one.

[TriGirl 349]

Quote

Quote

Four in 10 Americans Believe in Strict Creationism

And about that percentage in the US as well as on this forum seem to think that the three WTC towers were hit by holograms ten years ago today. (You're all the same on the inside).

I call it hopeless.

How about the number who think there were 3 towers?

See the upside, and always wear your parachute! -- Christopher Titus

Shut Up & Jump!

[billvon 3,132]

>Do Americans here on SC generally agree the poll to be a fairly accurate measure of >their countries religious beliefs?


Yes. Unfortunately, many americans do not "believe" in evolution.

[shropshire 0]

Quote

>Do Americans here on SC generally agree the poll to be a fairly accurate measure of >their countries religious beliefs?


Yes. Unfortunately, many americans do not "believe" in evolution.

But that will change over time

(.)Y(.)
Chivalry is not dead; it only sleeps for want of work to do. - Jerome K Jerome

[quade 4]

Quote

Yes. Unfortunately, many americans do not "believe" in evolution.

I believe that's partially due to the way some scientists have historically waffled on the subject of god to appease the masses.

Even you, billvon, continue to say a physical god is a possibility rather than simply a metaphor. Or at least you did the last time I checked.

quade -
The World's Most Boring Skydiver

[Southern_Man 0]

Quote

It's been 20 years since I took a stats class, but we talked about polls and I remember someone asking why many polls listed +-3% as error margin. The answer was because of the diminishing returns beyond 1500 sample size where the error is +-3% based on some formula I forgot 19.5 years ago. So with 1019, I'm guessing that formula spit out +-4%.

Yes, to 1/2 the margin of error you have to quadruple your sample size. It reaches diminshing returns at relatively low levels. The subgroup margin or errors would be higher but that all depends on the size of each subgroup.

"What if there were no hypothetical questions?"