NationStates Jolt Archive

..... that does not mean that their opinion is held by the majority of America.

I often hear about polls performed by newspapers, by universities, and by the government itself that present their findings as those held by the majority of America. They say they polled 100 people, or 300 people, or even up to 1,000 people, but they never seem to realize that those numbers are a fraction of a fraction of the U.S. population. Even if you polled 1,000,000 people (which is never done) you wouldn't have even reached the number of people who inhabit Manhattan island.

America has nearly 300,000,000 people (I believe the exact figure is in the area of 291,000,000 people). Polling 100,000,000 people is only asking 1/3 of the population. Polling 1,000 (which is barely ever reached) is so small of a percentage that I'm not even going to bother mentioning it.

There is no way to accurately gauge the mentality of the entire nation except by asking the entire nation. This has never, and will never, be done. Nobody knows the opinions held by the nations majority, and we probably never will. Get used to it.

The US is rather conformist, if you ask me. 99.5% of people are between the republicans and democrats in the political spectrum.

-----------------------------------------
R j00 b45h|n9 m3j3 6r4mm4r, ph45c|57?
Free your mind! (http://dwardmac.pitzer.edu/Anarchist_Archives/bright/berkman/comanarchism/whatis_toc.html)
I like big butts!
http://img63.photobucket.com/albums/v193/eddy_the_great/steatopygia.jpg

im guessing you havent taken a college level course in statistics

im guessing you havent taken a college level course in statisticsExactly.

It is likely that you will get a representative mix of views if you poll as many people as you can, in different areas, randomly.

..... that does not mean that their opinion is held by the majority of America.

I often hear about polls performed by newspapers, by universities, and by the government itself that present their findings as those held by the majority of America. They say they polled 100 people, or 300 people, or even up to 1,000 people, but they never seem to realize that those numbers are a fraction of a fraction of the U.S. population. Even if you polled 1,000,000 people (which is never done) you wouldn't have even reached the number of people who inhabit Manhattan island.

America has nearly 300,000,000 people (I believe the exact figure is in the area of 291,000,000 people). Polling 100,000,000 people is only asking 1/3 of the population. Polling 1,000 (which is barely ever reached) is so small of a percentage that I'm not even going to bother mentioning it.

There is no way to accurately gauge the mentality of the entire nation except by asking the entire nation. This has never, and will never, be done. Nobody knows the opinions held by the nations majority, and we probably never will. Get used to it.

If the poll is constructed well (i.e. the sample is truly random and large enough to reflect a wide spectrum of opinion), then you can say that it roughly reflects the opinion of the larger group. True, you might get outliers, such as a poll of 1000 people who all happen to be Republicans. But you might also flip a coin 1000 times and have it land on heads every time. Usually, you can get useful information out of a poll.

And the larger your sample size, the higher the likelihood you'll be accurate--the margin of error drops to a couple of points. As long as the methodology of the polling is sound and the questions aren't rigged to get a particular answer, you can get a fair sense of what a group of people are thinking, regardless of the size of the group.

The problem is that it's so easy to manipulate polls to get the answer you want by fiddling with the questions that they're not very trustworthy unless you can see the raw data and methodology (and know what you're looking at).

its very similar to evaluating medical studies. you cant just look at a study's conclusions and change your life accordingly. researchers do some stupid things. you have to look at how they ran the study and what statistical tests they did on the results before you can begin to trust the conclusions

same with polls. if you really want to believe the results you have to look at their methods. how did they choose their respondants? how did they frame their questions? look at what was actually asked, even a novice can recognize a loaded question. look at their stated margin of error. esp in political polls the MoE is often wide enough that the results could be reversed as to which opinion held the majority.

its a complicated business even when the pollsters are impartial.

You're right, Ashmoria--but it's also a fairly transparent business if you're willing to make the effort to actually look at the data and the questions. Problem is that most people don't care enough to do the work and just look at the results. The proliferation of statistically invalid internet polls has also hurt the credibility of polls as a whole. For instance, in an MSNBC internet poll a couple of days ago, Kerry led Bush and Nader something like 60-37-3. Mush as I wish that were the case, there's no way in hell Kerry is at 60% nationwide, but if someone who knows nothing about polling saw that, they might come to a different conclusion.

yeah its much easier to expect that kind of thing from the polls done by the parties themselves. we tend to expect our news organizations to do a good job. i expect it to be a very close election again this time although perhaps not as close as the last one. i do so hate it when the winner doesnt get the most votes.

Whenever you read a poll; always remember Benjamin Disraeli.

"There are three kinds of lies: lies, damned lies and statistics"

..... that does not mean that their opinion is held by the majority of America.

I often hear about polls performed by newspapers, by universities, and by the government itself that present their findings as those held by the majority of America. They say they polled 100 people, or 300 people, or even up to 1,000 people, but they never seem to realize that those numbers are a fraction of a fraction of the U.S. population. Even if you polled 1,000,000 people (which is never done) you wouldn't have even reached the number of people who inhabit Manhattan island.

America has nearly 300,000,000 people (I believe the exact figure is in the area of 291,000,000 people). Polling 100,000,000 people is only asking 1/3 of the population. Polling 1,000 (which is barely ever reached) is so small of a percentage that I'm not even going to bother mentioning it.

There is no way to accurately gauge the mentality of the entire nation except by asking the entire nation. This has never, and will never, be done. Nobody knows the opinions held by the nations majority, and we probably never will. Get used to it.

then I suggest you stop taking medication. or paying for insurance.

there is not garuntee that the medication is safe for you. only that it was effective and safe for X number of people.

so tough...do u have any practical suggestions?

im guessing you havent taken a college level course in statistics

Actually, I have. Just had the last class last Friday.

And I'm not saying you never get an accurate impression of the majority will with a limited poll, just that it's not definite. You can't say "America feels this way", you can just say "America probably feels this way"