Reality check: 95 percent of Americans had premarital sex

[Captain Chaos]

So, they lied. They didn't really mean 95 percent of Americans (and they knew they didn't mean that), they only meant 95 percent of the people they interviewed. Then they extrapolated and just ASSUMED that this result was true of the entire American population.

There really is a helluva lot of research and mathematical methodology into generalizing from a sample to the whole. Seriously, this is a huge area of science and math. I am sorry you do not believe any of it, but if you studied it well enough to understand it, it would make sense to you.

I don't consider asking people a bunch of questions and then drawing conclusions about a larger group of people based on the answers "science."

That is because you do not understand it. May I take it that you are both unwilling to study the matter in depth, and also unwilling to just believe the entire world of mathematics?

[Epistemologist]

Quote by: ByaKya oh sweet jesus, i know where this is going

Wow, we're still debating statistical validity? We need to know a bit more about the statisticians' data collection methods and how they did their hypothesis test; otherwise, we shouldn't really trust their data. The figures do seem somewhat nonsensical, but at least I could use them to make my point about your corruption.

Quote by: Captain Chaos There really is a helluva lot of research and mathematical methodology into generalizing from a sample to the whole. Seriously, this is a huge area of science and math. I am sorry you do not believe any of it, but if you studied it well enough to understand it, it would make sense to you.

Don't condescend on people just because you also took some statistics. Most of these studies seem awfully biased and don't consider numerous confounding variables that disallow them to make their conclusion by rejecting or accepting their null hypotheses. Even though they used sufficient samples, they still probably screwed up.

[PatrickHenry]

Quote by: Chancellor There is absolutely no truth whatsoever in polls or similar "studies." There is nothing scientific about asking someone a bunch of questions to a group of people and drawing conclusions about a larger group of people based on the subjective answers given.

That a pretty absolute response to something that most folks agree is a rational concept. Your position is very fringey, according to the polls...:rolleyes:

[Epistemologist]

Quote by: Chancellor I don't consider asking people a bunch of questions and then drawing conclusions about a larger group of people based on the answers "science."

Well, although many of these observational studies and pseudo-experiments are flawed, biased, and unreliable, some do provide acceptable results. It's just a matter of finding out which ones do; usually, there are some credible statisticians and polling agencies that exercise as much caution as possible in order to get rid of confounding variables.

[ByaKya]

Quote by: Chancellor The article said "More than nine out of 10 Americans, men and women alike, have had premarital sex." Since only 38,000 people were interviewed, that means there were 36,100 people who had pre-marital sex. Since the article claims that 95 percent of Americans had pre-marital sex, they're saying that 36,100 people is 95 percent of Americans. So, they lied. They didn't really mean 95 percent of Americans (and they knew they didn't mean that), they only meant 95 percent of the people they interviewed. Then they extrapolated and just ASSUMED that this result was true of the entire American population.

No, the data shows clearly that indeed 95% of americans according to this scientifically conducted study have had pre-marital sex.

Now i understand conservatives have a problem with science, but it must be nipped in the bud, before it becomes more serious.

[ByaKya]

Quote by: Epistemologist Don't condescend on people just because you also took some statistics. Most of these studies seem awfully biased and don't consider numerous confounding variables that disallow them to make their conclusion by rejecting or accepting their null hypotheses. Even though they used sufficient samples, they still probably screwed up.

He can condescend, espeically in this case where it was warranted.

MOST studies seem awfully biased? Back that up
Probably screwed up?
They are a bit more educated they do this for a living, so im going to go with the "they probably didnt screw up" theory if you dont mind

[Epistemologist]

Quote by: ByaKya No, the data shows clearly that indeed 95% of americans according to this scientifically conducted study have had pre-marital sex. Now i understand conservatives have a problem with science, but it must be nipped in the bud, before it becomes more serious.

Well, the sample could have been just college students, for instance, and that is one possible confounding variable. As I said before, unless we know more about the data collection methods and hypothesis test, then the conclusion is rather specious. I'm just saying. There are some reliable polls, but without more information I guess this one's reliability is questionable.

[Epistemologist]

Quote by: ByaKya MOST studies seem awfully biased? Back that up

Well, I was afraid to use the word "most," but I must say that that's the impression. Once a few studies are unreliable, people are led to believe that the whole field is unreliable. I'm not saying that it is, though.

Then again, the contradicting conclusions of some scientific studies kindles this impression. Of course, there are some reasonable explanations for these, and scientific methods are often used, but often more bias than is necessary is included.

Here's an example of confounding . If we look at the increased number of drownings that occur at the beach during summer and the increased number of ice cream sales, then we see a correlation. Some unfortunate fellow might also see a causation. However, I think it's pretty clear that ice cream does not cause drowning. There was confounding variable, and the increased drownings were really caused by increased number of people swimming since there are more people at the beach. However, someone intent on forming a certain conclusion will ignore confounding variables that have no negligible effect and employ too much bias in making a conclusion. Their hypothesis test would be flawed. This is often what occurs in studies, such as those saying that chocolate is good for health. Of course some would naturally say that because they like chocolate; maybe it is good for health, but their studies often don't prove it adequately.

Quote by: ByaKya Probably screwed up? They are a bit more educated they do this for a living, so im going to go with the "they probably didnt screw up" theory if you dont mind

I didn't see this part of the message the first time I read it. Nonetheless, see above. I think that just because someone is a statistician, it doesn't mean he/she is good or is strictly interested in maintaining the integrity of the practice. Scientists aren't always looking for the truth. Just look at that Korean guy and his pseudo-research on stem cells. They might have other agendas, unfortunately.

[Chancellor]

Quote by: ByaKya No, the data shows clearly that indeed 95% of americans according to this scientifically conducted study have had pre-marital sex.

No, the data only shows that out of 38,000 people 95 percent had pre-marital sex.

Now i understand conservatives have a problem with science, but it must be nipped in the bud, before it becomes more serious.

First, I'm not a conservative and, second, I don't have a problem with science (as long as it is completely free of a priori assumptions such as the philosophies of naturalism or creationism), I have a problem with calling this kind of study "science."

[Chancellor]

Quote by: Epistemologist Well, the sample could have been just college students, for instance, and that is one possible confounding variable. As I said before, unless we know more about the data collection methods and hypothesis test, then the conclusion is rather specious. I'm just saying. There are some reliable polls, but without more information I guess this one's reliability is questionable.

Then, of course, there's the number of people sampled - 38,000 and the vast majority of them women. There are approximately 300,000,000 Americans, of which 38,000 is significantly less than one percent. Further, women only comprise about 51 percent of the American population while the study questioned 33,000 out of 38,000 (significantly more than 51 percent of the people questioned). Consider me stubborn, pig-headed or whatever but I refuse to accept the notion that sampling significantly less than one percent of the entire American population can produce valid information about 95 percent of the entire American population. Then, of course, one has to also take into consideration the percentage of the American population that is pre-pubescent or otherwise theoretically incapable of "pre-marital sex."

[PatrickHenry]

I would like to see the research methodolgy and data. But it costs $65 to subscribe. I guess I can wait for a public source of the research. Public Health Reports | ONLINE

But overall, would anyone contest that large majorities of US men and women had sex before marrying?

[Gods_Mercenary]

Don't think so, but polls as a concept are severely flawed

[PatrickHenry]

Quote by: Gods_Mercenary Don't think so, but polls as a concept are severely flawed

it all depends on the methodoly. They can be very precise.

[Epistemologist]

Quote by: PatrickHenry But overall, would anyone contest that large majorities of US men and women had sex before marrying?

No, that seems correct, although these provided figures are specious. But I think I sort of proved my own point that pedagogues and other people simply use pseudo-statistics to prove their claims. That is, I was previously talking about the corruption of Westerners based on this poll itself even while knowing that its conclusion was flawed due to its bias.

Quote by: Chancellor Then, of course, there's the number of people sampled - 38,000 and the vast majority of them women. There are approximately 300,000,000 Americans, of which 38,000 is significantly less than one percent. Further, women only comprise about 51 percent of the American population while the study questioned 33,000 out of 38,000 (significantly more than 51 percent of the people questioned). Consider me stubborn, pig-headed or whatever but I refuse to accept the notion that sampling significantly less than one percent of the entire American population can produce valid information about 95 percent of the entire American population. Then, of course, one has to also take into consideration the percentage of the American population that is pre-pubescent or otherwise theoretically incapable of "pre-marital sex."

Yes, you've listed other forms of bias present in this poll. One of the biggest mistakes made is having too small of a sample; you need a certain size for certain hypothesis tests. The more you have, the more reliable your results can be, although there comes a point when adding more to the sample won't significantly increase your reliability. Anyway, I think we can argue that this poll had too much bias and mistakes.

[PatrickHenry]

Quote by: Epistemologist ... Anyway, I think we can argue that this poll had too much bias and mistakes.

You could argue that. But without access to the methodology and data, you're just blowing smoke...

[Gods_Mercenary]

You have to be completely random and have a huge base, but even then, bias can be put into the wording in undetectable ways.

[Chancellor]

Quote by: PatrickHenry I would like to see the research methodolgy and data. But it costs $65 to subscribe. I guess I can wait for a public source of the research. Public Health Reports | ONLINE But overall, would anyone contest that large majorities of US men and women had sex before marrying?

I have no doubt that a large majority of men and women had sex before marrying. I also have no doubt that a large majority of men and women who have never married had sex.

[Chancellor]

Quote by: Gods_Mercenary You have to be completely random and have a huge base, but even then, bias can be put into the wording in undetectable ways.

And if the questioning was done face-to-face, even the inflection in the voice as the question is asked can change the results. It is exactly because of these kinds of things that I reject the entire notion of statistical sampling.

[Epistemologist]

Quote by: PatrickHenry You could argue that. But without access to the methodology and data, you're just blowing smoke...

On the contrary, without access to that stuff, I'd just reject their conclusion. Already from what I've seen concerning the poll, there seems to be enough bias to say their conclusion is unacceptable. I mean, they're the ones making new claims and giving new figures.

Quote by: Gods_Mercenary You have to be completely random and have a huge base, but even then, bias can be put into the wording in undetectable ways.

Yeah, besides from the most obvious mistakes in the application of the statistical method itself, bias can still be present.

Quote by: Chancellor And if the questioning was done face-to-face, even the inflection in the voice as the question is asked can change the results. It is exactly because of these kinds of things that I reject the entire notion of statistical sampling.

Yes, this often occurs particularly in verbal polls. But sometimes when we're testing a drug, for instance, we can minimize the effects of confounding variables and bias by making the study/experiment blind, double blind, etc. Also, statistical sampling is used on non-human subjects as well, so that form of bias you mentioned wouldn't really make much of an impact.

So, you shouldn't discard the entire field of statistics just because of the mistakes of numerous statisticians. However, they give a notedly bad impression.

[PatrickHenry]

Quote by: Epistemologist On the contrary, without access to that stuff, I'd just reject their conclusion. Already from what I've seen concerning the poll, there seems to be enough bias to say their conclusion is unacceptable. I mean, they're the ones making new claims and giving new figures.

And the researcher published his study. It is available if you want to see it. I just don't feel like spending the money. However for you to sandbag his study based upon your own ignorance of the research is medieval, I think is the correct term...