LifeWay, an evangelical religious research firm, found that most Americans view most denominations favorably and when they feel ready, they would be willing to explore any or all of the churches listed.

“American Perceptions of Denominations,” is the result of a survey of 1,000 U.S. adults last fall who were asked to give a favorable or unfavorable rating on nine traditions, or say they don’t know enough to form an opinion.

RNS:

Many Americans today don’t think they have a place for church in their lives.

“But they believe the church has a place for them, when or if they are interested,” said Scott McConnell, vice president of LifeWay Research, which issued a new survey Wednesday (June 3) on perceptions of religious denominations.

The findings show that as many as 45 percent of Americans will look at the church brand on the sign out front — Catholic or Baptist or Methodist or whatever — and drive past, thinking it is “not for me.”

And yet, McConnell said, the survey reveals an openness in most people — if not a very theologically deep one — to stopping by, even if they declare no religious identity, the “nones.”

“Many people view a church like the ice cream parlor down the road. They think, ‘When I’m in the mood, I can go.’ Church leaders can take it as good news: People haven’t ruled them out. But they have to be a little unsettled at how little people are thinking about this,” said McConnell.

*Posted by Andrew Gerns*

Sticking to the topic (the nones), what I liked about this survey is that it challenged my priors that the nones are simply turned off by religion, and that the nones turned off to all religion by the louder voices of more conservative religionists in politics and in the media.

What this survey suggests (setting aside questions of whether it is done well, or presented with a slant) is that the nones aren't as turned off to religion as I thought. They might even visit from to time to time, and have the sense that the church is there for them in a time of crisis.

The table ranks the denominational groupings by "favorability". It lumps all Lutherans together as one; likewise all Methodists and Presbyterians. If you look the size rankings of denominations you will see that Episcopalians come up next (14th largest, with two flavors of Lutherans ahead of us, etc.).

http://hirr.hartsem.edu/research/fastfacts/fast_facts.html#largest

It looks to me that we're seeing something along the lines of "I know someone who's a ___." It's like the more become aware they know someone who is gay the more favorable they become of gay rights.

Sorry, my mistake. A sample of 1,000 gives a margin of error of 3.1%, not 3.2%. Teach me to work from memory.

In any event, while the population size does have a very small effect in calculating a margin of error (as in, several decimal places in), the sample size is the significant factor. As the sample size increases, the margin of error drops - but the RATE at which the margin of error drops slows down significantly.

A sample of 50, for example, gives a margin of error of 13.9%, while a sample of 100 gives a margin of error of 9.8%. Doubling the sample size by 40 decreases the margin of error by 4.1 points.

But doubling the sample size again to 200 only reduces the margin of error to 6.9%, a reduction of only 2.9 points.

Bear in mind that, for the researcher, there is a cost in every contact made, so doubling the sample size doubles the cost, and the benefit of doing so gets less and less as the sample increases.

Most pollsters will operate on a sample in a range from just below 1,000 to just above 1,500. Occasionally polls will sample over 2,000, but not often. The margin of error for a sample of 1,000 is 3,2%. For a sample of 1,500 it's 2.5% and for 2,000 it's 2.2% Doubling the sample size from 1,000 to 2,000 - and doubling the cost of the research - reduces the margin of error by less than one percentage point.

If one more than doubles the sample size from 2,000 to 5,000, the margin of error drops less than a percentage point to 1.4%, and doubling that sample size to 10,000 only reduces the margin of error to 1.0% - less than half a percentage point.

So no, a sample size of 1,000 isn't small at all, not for the state of Wyoming, not for all of the USA and not for the entire planet. The benefit off doubling the sample size to 2,000 is minimal and anything beyond that devolves into silliness.

Sounds like solid sample.

Sounds like 'Episcopal' as a category no longer is missed for these purposes.

Less and Less people are attending a church. The older I get and the more I see I can understand why. I am very thankful that I was privileged to be brought up in the Episcopal Church and it continues to change and move in more progressive ways and I am also thankful for that. I think though that most people nowadays are focusing on spirituality, not religious tenets. How many times have I heard someone say it seems religions cause more problems than they solve and we wouldn't have so much conflict in the world if it weren't for religion. It seems like that is a valid statement. What's worse is there seems to be many ideologies within each religion. I try to keep abreast of what is going on in the world with regards to the many religions, cults, aberrant groups, etc. It's not easy. One of the wisest people I will ever know was my father. He always said before you start discussing a higher power and what their beliefs are, find out what their concept is of their higher power-and it's true-everyone has a different concept. These days I am around quite a few young adults, and most of them are atheists. Even older persons like myself seem to be drifting toward more meditation and spiritual routes, rather than organised religion. The future of religion is up in the air and be far different I imagine in another 20 years.

1000 is NOT a small sample. A random sample of 1000 will be accurate within 3.2 percent, 19 times out of 20. One can question how effective the sampling methodology is at achieving randomness, but a 1000 sample is not small.

In fact, the majority of polls work on samples in the order of 1000 because the rate at which the margin of error shrinks gets slower the higher the sample size. That is to say that the difference between a sample of 500 and a sample of 1000 is significantly larger than between a sample of 1000 and a sample of 2000.

Public opinion research has a whole lot of challenges - one of them being achieving authentically random samples. But one of the problems is that most people have no clue how polling works.

But you have to admit that 1000 surveys seems a small number to get an accurate idea of what a nation of 320 million people thinks. At least is does to me.

Bro David

Given our average Sunday attendance (about 1/3 that of the Assemblies of God) we're basically too small to measure accurately. Kind of like the Church of the Nazarene, or Church of Christ (evangelical denominations which are also not included in the survey).

Malcolm is basically correct.

With 1000 samples, if 1% of the population belongs to a denomination the survey will have about a 90% chance of lying between 0.5% and 1.5%. In other words, pretty inaccurate. With 10000 samples there's about a 90% chance you'll get an answer between 0.8 and 1.2%.

For 3%, about 90% of the estimates will lie between 2 and 4%.

Not so. A sample of 1,000 gives a margin of error of 3.2%, 19 times out of 20 whether the sampled population is 320 million, 320 billion, 320 trillion or 320 thousand.

1000 is a very small pool for a very large, pluralist country.