What you're talking about is subjective data, where the deliverer of the data has a vested interest in being less than honest, in having a vested interest in delivering an underestimate of infidelity. You cannot conduct a worthwhile study using data that is that flawed.

But one last thing regarding "vested interests". The researcher in question and the data collector are independent of each other, have no association, and the data collector has no reason to sway the data one way or another. Your statement supports my conclusion that you don't know what you're talking about.

You cannot conduct a worthwhile study using data that is that flawed.

Wrong. Data have problems. Some more than others. The more problematic the data, the less reliable the conclusions -- or the weaker the conclusions that can be drawn at the same reliability (actually, I shouldn't be using reliability here, since it's a term of art with a precise definition, but you get the point). However, many studies have been done to benchmark the reliability (here used correctly) and accuracy of self-reported data against more objective measurements. All of the problems you cite -- and many others -- have been looked at in detail by methodological researchers. Social science data analysis explicitly includes sources of bias and measurement error in the modelling process. Well, for people who do good work anyway -- I can't comment on any particular studies but the rate of bad methodological practices in social science is pretty high.

But the important point is this: just because the data has problems doesn't mean it's worthless. There's plenty of gray area between perfect and worthless -- some might call this area "useful."

Scotto, you're missing the point. You're talking about objective information there. There are independent means of determining how many years someone went to school. We're talking about asking someone how many times they ###### someone else they weren't married to. And apparently, there was no corroborating objective data to badk up the survey data, like DNA analysis of vaginal swabs or soemthing like that.

So, no matter how many time you say the authors are aware of the problems of survey data, and try to control for it, unless those controls can be independently validated by objective means, they don't mean anything and we're back to square 1.

Again, I will remind you of the history of AIDS epidemiology. Male Haitian homosexuals or heterosexual male prostitutes catering to homosexual clients were out-and-out lying to their interviewers about their sexual practices. The interviewers were professional statisticians who were well aware of the methodolical biases but weren't able to sort out the true picture and eliminate Haitians as a high-risk group until an assay (an objective source of data) was available.

So don't try to tell me data I don't believe is reliable is reliable because the statisticians who did the stury were trying to control for biases. Unless they can show me they weree able to do so effectively, then I am not going to believe the results of the study.

With this statement, you have demonstrated such a fundamental misunderstanding of the topic being discussed simultaneously coupled with a lack of willingness to try to grasp what you so clearly don't understand that I'm concluding that I'm wasting my time. It's like arguing with a 9/11 conspiracy theorist or an ideologue.

No, YOU won't admit to the inherent weaknesses in the methodologies the social "sciences" often use in conducting studies.

I'm sorry but you're going to have to do better than "Well, the authors were aware of the weaknesses in the data and tried to control for it". That's BS.

Kevin: As I said yesterday, the more you speak on the subject, the dumber (less credible) you sound. scotto nailed it on the head.

The interviewers were professional statisticians who were well aware of the methodolical biases but weren't able to sort out the true picture and eliminate Haitians as a high-risk group until an assay (an objective source of data) was available.

Well, maybe that's the problem. Professional statisticians are not trained interviewers.

And how was it discovered that a subgroup of the Haitian population was lying to the interviewers? Most likely, by doing social science research.

Wrong. Data have problems. Some more than others.

Data that has problems is different from data that is flawed. Flawed data is unreliable data. You cannot use flawed data, data that contains excessive misinformation, and make any useful conclusions from it. That's it. You can't.

Again, I will remind you of the history of AIDS epidemiology. Male Haitian homosexuals or heterosexual male prostitutes catering to homosexual clients were out-and-out lying to their interviewers about their sexual practices. The interviewers were professional statisticians who were well aware of the methodolical biases but weren't able to sort out the true picture and eliminate Haitians as a high-risk group until an assay (an objective source of data) was available.

I rather doubt professional statisticians did the interviews on an epidemiological study. In any event, doing face to face interviews is probably the wrong way to collect sensitive data of this sort. That's when you use a paper questionnaire or computer to collect the data, so that the respondent has no incentive not to report accurately, which I've mentioned before and which you've ignored because it contradicts the fundamentalist position that you hold.

DCA, good posts at 402 and 406.

Data that has problems is different from data that is flawed. Flawed data is unreliable data. You cannot use flawed data, data that contains excessive misinformation, and make any useful conclusions from it. That's it. You can't.

Again, wrong. Social science is not hard science as you understand it. If you're thinking like a hard scientist you're not understanding the data. You can't use the methodology in one to do the other. The appropriate comparison for social science stats are ecology and atmospheric science, where there's more noise than signal in the data and bias and uncertainty in measurement.

Again, most importantly, there's lots of space between perfect and worthless. All good social science data lies in this space. Some social science data *is* worthless. That you can't tell the difference doesn't mean that the rest of us can't.

DMN, how would you feel about your 15 year old daughter hanging around with a 28 year old married man?

I don't have a 15 year old daughter, but if I did, I suspect I wouldn't want her ever having contact with any individual of the opposite sex over the age of about 12. Maybe 10, just to be safe. I think this attitude of mine would probably not change when she reached 18, but might cease at the point where I was dead of old age.

I don't really see the relevance of the attitude of a father towards his daughter's possible sexuality, which isn't normally the type of situation in which we expect dispassionate, balanced thinking.

And another thing. I tried to read the entire article JC referenced. I have an NIH library account and the The Journal of Family Issues isn't available. There are a bunch of family-related social science journals available but not that one. And since the NIH library stocks just about every journal worth reading, I'm not so certain the article JC has cited is "highly touted".

Let it go, Kevin. No one's trying to convince you of anything.

With this statement, you have demonstrated such a fundamental misunderstanding of the topic being discussed simultaneously coupled with a lack of willingness to try to grasp what you so clearly don't understand that I'm concluding that I'm wasting my time. It's like arguing with a 9/11 conspiracy theorist or an ideologue.

But he's a scientist! Just ask him!

Remember, this is the same guy that didn't understand why you need to do tests on separate samples before one can conclude that a drug test is positive.

If you're thinking like a hard scientist you're not understanding the data.

Wrong. If you're NOT thinking like a hard scientist, you're not understanding the data. "hard"=solid/reliable. "Soft"= equivocal/unreliable.

Let it go, Kevin. No one's trying to convince you of anything.

Yes they are. They're trying to convince me that biased and unreliable data can be used to make objective assumptions. I'm never going to buy that one.

I don't really see the relevance of the attitude of a father towards his daughter's possible sexuality, which isn't normally the type of situation in which we expect dispassionate, balanced thinking.

Maybe it isn't relevant. I was only asking to hear what I presume to be another human's pov would be.

I have a soon-to-be 10 year old in the house. Regarding her sexuality, I half kiddingly wish for her to be a lesbian. (She doesn't know this, and won't, until she's an adult and we can all laugh about it together.)

I rather doubt professional statisticians did the interviews on an epidemiological study.

Well then, you're wrong about that too.

That's when you use a paper questionnaire or computer to collect the data, so that the respondent has no incentive not to report accurately, which I've mentioned before and which you've ignored because it contradicts the fundamentalist position that you hold.

I haven't ignored it. I've considered it, then blown it off as BS. As anyone trained in the scientific method would.

Being a stickler for accuracy is not being a fundamentalist. In the sciences, NOT being a stickler for accuracy is the same as not being a scientist at all.

Kevin:

As a "hard" scientist who often veers into drawing conclusions about human beings on less than "hard" evidence, what would you conclude about the following person:

(1) Makes repeated false claims about his military service;
(2) Makes claims suggesting he has read a book he has not; and
(3) Makes definitive claims about fields he exhibits no competence in whatsoever.

Would you join me in drawing the conclusion that his opinions are unreliable?

JC, when you're in a hole, it's best to stop digging. That study you cited is horseshit.

Which makes your elitist claims about persons of religious faith being somehow more virtuous than persons of secular orientation horseshit too.

The "study" that's horseshit is the General Social Surveys, which as scotto pointed out, is widely used and esteemed. The analysis is done by a cracker-jack young lead social scientist. In your favor you've got your pattern of lies, exaggerations, and incompetence. I have no doubt you'll continue to think you won the argument. The clincher being back when you placed scare quotes around "sciences".

Kevin: Since you've got no qualms about demeaning others' jobs and products, you tempt me to do the same about yours with a line a nuclear engineer once used when talking to me about chemists. I'll refrain, however.

The "study" that's horseshit is the General Social Surveys, which as scotto pointed out, is widely used and esteemed.

Esteemed by whom? By people who approve of it's conclusions? Or by people who are truly interested in objective information?

I think we both know the answer to that.

And Scotto has made a thoroughly unconvincing case, even going so far as to suggest that making the data collection process of highly embarrassing (or more so, criminal) behavior more "objective" by eliminating a face-to-face interview.

Right.

Kevin: Since you've got no qualms about demeaning others' jobs and products, you tempt me to do the same about yours with a line a nuclear engineer once used when talking to me about chemists. I'll refrain, however.

See, this is why you get off the rails so often. You take criticism so personally, you start to twist what has actually been written. I'm not demeaning jobs and professions. I'm legitimately criticising methodologies, which is part and parcel of the profession of every person dealing with the generation of new information.

If you can't take legitimate criticism, then you have no business doing science.

If you can't take legitimate criticism, then you have no business doing science.

The key phrase is legitimate, which implies an understanding of what's being discussed, which is not the case here from your end.

Scotto, I think it is you who is tone-deaf to legitimate criticism. And my criticisms are certainly legitimate, as you have already acknowledged, but tried to rationalize away.

Data that has problems is different from data that is flawed. Flawed data is unreliable data. You cannot use flawed data, data that contains excessive misinformation, and make any useful conclusions from it. That's it. You can't.

kevin, you know I'll defend you if I think people are being unreasonably dismissive. I can't do it in this case. Brian McNamee's testimony has excessive misinformation; and DMN's (and RDiP's, and others') point for several months is that on that basis we can't make any useful conclusions from it. Yet you have had no trouble doing so. This flies in the face of your claims on flawed data above.

That you might eventually be proven correct does not excuse the fact that you are willing to use flawed data with certainty when it suits you. I'm sure in your day job you have a higher standard than you've displayed here.

DMN, how would you feel about your 15 year old daughter hanging around with a 28 year old married man?
I don't have a 15 year old daughter, but if I did, I suspect I wouldn't want her ever having contact with any individual of the opposite sex over the age of about 12. Maybe 10, just to be safe. I think this attitude of mine would probably not change when she reached 18, but might cease at the point where I was dead of old age.

I don't really see the relevance of the attitude of a father towards his daughter's possible sexuality, which isn't normally the type of situation in which we expect dispassionate, balanced thinking.

Ha! My wife brought up the Clemens-McCready thing last night, using the exact same wquote "how would *I* feel" And I replied precisely as DMN did - it won't stop at 18 or 30, but when I am dead.

Esteemed by whom? By people who approve of it's conclusions? Or by people who are truly interested in objective information?

The GSS does not have "conclusions." It is not a study; it's a survey. It's a giant data set, collected regularly (annually or biannually) for the last 35 years. Because of that, it's the gold standard of social science data.

It's incredibly valuable both because it has been conducted over such a long time (allowing social scientists to study trends) and because the data is made freely available to everyone, allowing anybody to use the data set.

Which makes your elitist claims about persons of religious faith being somehow more virtuous than persons of secular orientation horseshit too.

I missed this. "Elitist?" I don't think so. Further, you may note that I specifically denied that this study means religious people are more virtuous. Add #5 to my list above: (5) Makes false assertions about what people claim that evince either (a) a willingness to distort other people's views or (b) serious deficiencies in reading comprehension.

No; "clinging to religion" is the opposite of elitest.

The GSS does not have "conclusions." It is not a study; it's a survey. It's a giant data set, collected regularly (annually or biannually) for the last 35 years. Because of that, it's the gold standard of social science data.

Thank you for re-stating a point that I made earlier, and it is the second most used dataset for social science research after some US Census surveys for a very good reason.

This is what I meant about kevin not understanding what he's talking about, conflating data collection and survey design with the study, and so on and so forth.

Kevin,

I have the article if you want me to send you a copy. Here's what the article has to say about the issue you mention, in the discussion.

Moreover, research of this kind necessarily relies on self-reports of infidelity,
and some skepticism about the reliability of such data may be warranted.
However, despite some allegations to the contrary, studies have
turned up little clear association between religiosity (especially the kinds of
religious variables considered in this study) and the tendency to give biased,
socially desirable responses. At least one recent, thorough study of this issue
among young adults argues strongly against such a view (Regnerus & Smith,
2005). Nevertheless, it would be helpful for future studies to rule out obvious
sources of response bias in work on marital infidelity among adults.

The cited source is Regnerus, M. D., & Smith, C. S. (2005). Selection effects in studies of religious influence. Review of Religious Research, 47, 23-50. I have access to this as well but haven't looked at it at all.

I'm not demeaning jobs and professions.

Yes, you are. You are demeaning everyone who works in social science research, by classifying the type of data they use (and hence everything they do) as worthless. I said earlier that social science is not hard science -- as you understand the term "science," social science is not science. I used "hard science" to represent that distinction, but if that didn't get the message across: social science is not science. It's a fundamentally different type of inquiry and the methods are different, though they overlap quite a lot with the more messy scientific fields. "I'm a scientist" doesn't signify expertise in this area.

Thanks for the cite, DCA.

I have a soon-to-be 10 year old in the house. Regarding her sexuality, I half kiddingly wish for her to be a lesbian. (She doesn't know this, and won't, until she's an adult and we can all laugh about it together.)

Unless she lurks on BTF.

Then again, having her lurk on BTF may well be the most effective means of scaring her away from men.

Then again, having her lurk on BTF may well be the most effective means of scaring her away from men.

No kidding. :)

Although there are only two gay guys I know of on here. The other one doesn't scare me at all, except when he talks about 1986.

kevin, you know I'll defend you if I think people are being unreasonably dismissive. I can't do it in this case. Brian McNamee's testimony has excessive misinformation; and DMN's (and RDiP's, and others') point for several months is that on that basis we can't make any useful conclusions from it. Yet you have had no trouble doing so. This flies in the face of your claims on flawed data above.

Vi, this might be true if McNamee's testimony was the only information out there. But the footprints of clemens juicing were all over the snow years before the Feds busted mcnamee. All his story does is corroborate essentially what everyone with one good eyeball could already see.

At least Clemens has Karl Malone on his side.

You are demeaning everyone who works in social science research

No I'm not. I'm criticizing some of their methodologies as unscientific. That's not the same thing.

Although there are only two gay guys I know of on here. The other one doesn't scare me at all, except when he talks about 1986.

Yeah, I don't think the two gay guys (that I, too, know of) here are the ones Bivens would worry about his daughter hooking up with, whether or not sexual orientation's included in the equation.

Since my example with the black politicians vanished without a ripple, here's another good example of the way that people will lie about things that make them ashamed: The significant gap in results for studies on the average length of the erect male penis, between studies that rely on self-reported data and studies that actually involve an impartial arbiter putting a ruler beside the organ.

I'm glad that social scientists try to correct for error, but without knowing things that they have no way of knowing, they can't get a truly accurate estimate of the magnitude of potential sources of error, either. If they actually knew, they wouldn't need to run the studies in the first place.

It's not just the cheatin' couples. In general, I don't believe in most sociological research, because I think that the approximations and assumptions that are inherent to the process make it useless for anything more than entertainment value. Such is life.

Vlad, the so-called "Bradley effect" is out-of-date anyway, and again, it represents a workaroundable methodological problem. (Note that the issue here is the difference in fidelity, not absolute levels of fidelity, so unless religious people are more likely to lie about it, the fact that people will lie about fidelity is irrelevant.)

There are problems with sociological research, to be sure. The major one is that it's very difficult to control for everything -- indeed, it's very difficult to figure out what needs to be controlled for in any given case. And most studies that don't use the GSS or Census are working with data sets too small to allow for the extensive controlling which is necessary.

But that doesn't mean that all social science research is a waste of time. You can, if clever, work around many problems and find ways to corroborate your results. (Hell, you can get a New York Times reporter to collaborate with you on a bestselling book if you're clever enough about it.)

No I'm not. I'm criticizing some of their methodologies as unscientific.

Of course they're unscientific. That's not a criticism. It would be like criticizing a woman for being unmale. Or a square for being unround. We already knew that, so what's the point?

"so unless religious people are more likely to lie about it"

Which is exactly my point. We don't have any way of knowing whether they are or not.

I have no problem with sociologists doing what they do, and I'm sure they do the best they can with what they've got. I just think that it's foolish to treat a lot of this stuff as being significantly stronger than a wild-ass guess.

I should know better than to get involved here, but I just wanted to remind everyone what started this conversation. kevin wrote:

Hah. Show me one study that suggests religious people are more faithful to their spouses than non-religious ones.

JC in DC cited such a study, at which point kevin summarily dismissed it because it was a social science study. Which is ridiculous, because even if you think the data used in the study is "flawed," at the very least this is a study that "suggests" religious people are more faithful to their spouses.

My question for kevin is, when you asked to be shown "a study that suggests religious people are more faithful to their spouses than non-religious ones," what kind of study were you expecting to see? What would you consider a satisfactory study? If nothing would satisfy you, why ask the question in the first place?

Of course they're unscientific. That's not a criticism. It would be like criticizing a woman for being unmale. Or a square for being unround. We already knew that, so what's the point?

They may not be scientific in the strictest sense of the word, but opinion research does reflect what people think, and how different groups think differently about things. The vast majority of the surveys that are done, mostly funded by the government, are about non-controversial, non-sensitive topics. Are you going to lie about whether your kid has been vaccinated? I don't think so.

In addition, these same techniques are used by huge corporations to inform their business decisions, marketing strategies, and so forth, not to mention how campaigns use them to figure out how to win. If it were an epic failure, you'd figure they'd be doing something different.

Vi, this might be true if McNamee's testimony was the only information out there.

So, flawed data is useful if it agrees with prior subjective observation (the "one good eyeball" test), but useless on its own?

That's not even wrong.

"I have a soon-to-be 10 year old in the house. Regarding her sexuality, I half kiddingly wish for her to be a lesbian."

been there, done that ; ) , and mine i 4 and a half. i suspect you already understand her position on the matter, if you dare.

my boy, on the other, what a robust, able fella! you should see him kick a football (i mean soccer ball, of course)

jeez. roger, i hope you get whatever it is you deserve. you made your own bed (regardless of the occupants).

David Nieporent Posted: April 30, 2008 at 01:22 PM (#2764222)

DMN, how would you feel about your 15 year old daughter hanging around with a 28 year old married man?
I don't have a 15 year old daughter, but if I did, I suspect I wouldn't want her ever having contact with any individual of the opposite sex over the age of about 12. Maybe 10, just to be safe. I think this attitude of mine would probably not change when she reached 18, but might cease at the point where I was dead of old age.

- grinning

EXACTLY what my daddy said. i mean thought. most all daddys get daddyitis over their Little Grrrls and dial he gonna be the worst EVAH bout LED

my brother he starting to seriously repeat my daddy's line about how God give men daughters to get even with them for what they did/tried to do to every other man's daughter...

Funny. It's said that Dad wants his daughter to be a lesbian. But if his son is gay, he wonders where and how he screwed up. I say this only half in jest.

I told Answer Dad "If having a lousy father is what turns a boy gay, then the ghettos in Baltimore would gayer than P-Town." It doesn't seem to register.

"Are you going to lie about whether your kid has been vaccinated? I don't think so."

Some people probably do. My dad hates survey-takers, and lies to them about everything.

Social science is science. This is why I earlier brought up philosophy of science and scientific methodology. THere's no reason - except prior ideological commitment which in Kevin's case is clear - to reserve science for the particular kind of science practiced by hard scientists. Kevin spouts off as though that kind of science is simply absent bias, but as I'm sure many here know, that's not true. The existence of biases doesn't negate science, it's a part of science even while scientists of all kinds - hard and soft - do their best to limit bias and make biases known and accounted for.

I had a nice exchange w/the author of the study Kevin so cavalierly and ignorantly dismissed. She mentioned the article DCA mentioned and pointed to a book as well ("The Social Organization of Sexuality") which treats survey methodology pertaining to sex.

Dave rightly points to the genesis of this particular occasion of the nudity of Kevin's ignorance. What we've learned is that Kevin wasn't asking for a study, but simply implying that social science is bunk through and through.

Which is like a nuclear engineer saying of chemistry that it's just morons playing with beakers, that chemistry is the science failed scientists get into: dismissive and insulting.

It may be dismissive and insulting, but in this case it's not wrong. Until we can actually drill holes in people's heads and pull out thoughts and look at them, there simply isn't any way to know whether people are telling the truth or not. Doing your best to limit and account for biases is good, but it just isn't the same thing as delivering demonstrably accurate results.

Social science, as currently practiced, requires too many leaps of faith to meet the same standards as hard science, and therefore should deservingly be treated as less reliable. Full stop.

Doing your best to limit and account for biases is good, but it just isn't the same thing as delivering demonstrably accurate results.

Actually, it is pretty much the same thing.

"Actually, it is pretty much the same thing."

No, it's not. If the Pirates signed me tomorrow, I could do my best to deliver quality left-handed relief pitching, but the mere fact that I did so wouldn't make me a quality left-handed reliever. I could make a good-faith effort to dig myself out of Attica with a spoon, but that wouldn't make me a master escapologist. Etc.

If you don't have the tools, you don't get the results, and right now the tools that the social sciences would need to deliver demonstrably accurate results just don't exist. Once we get brainwave scanners and throw away the rules about informed consent for experimental subjects, maybe it'll be different.

"Funny. It's said that Dad wants his daughter to be a lesbian. But if his son is gay, he wonders where and how he screwed up. I say this only half in jest."

Regarding sexual orientation, my two cents is that I want my kids to be happy. The sooner they figure that out, the better.

For religious people, it is much easier to consider something like adultery to be totally out of the question, because it's a sin.

Hah. Show me one study that suggests religious people are more faithful to their spouses than non-religious ones.

Since I objected to what I saw as chick's bigotry, I have to object to this as well. Please don't tell me that you really think religious people have stronger morals than nonreligious people. Didn't the Inquisition, the jihad of your choice, the Quaker who bombed Cambodia, the latest parish scandal, teach you anything?

.... Arkitekton, my conservative friend, please tell me you're not this daft....

The entirety of Crispix's paragraph went:

"For religious people, it is much easier to consider something like adultery to be totally out of the question, because it's a sin. But for non-religious people...well, it's a victimless sin. Unless the "victim" becomes aware that it has occurred."

There is a claim of superiority for religious people in Crisxix's paragraph, and that claim has an implicit moral component since he specifically uses the word, "sin". (If you can find one theologian who considers each of sin and morality to be entirely distinct, one from the other, I will be astonished and impressed.) If we assume for the purposes of this discussion that adultery is wrong, whether we describe it in religious terms as "sinful", or in secular terms, as "unjust", to claim that religious persons are less likely to commit adultery inevitably equates, for a religious person, that greater ability to resist as a moral superiority, at least in the case of adultery. Following this, giving examples of how people professing faith are regularly capable of committing atrocities may be rhetorically crude, but not at all out of line with the preceding.

So, no, JC, not daft. At least not here and not yet. If you tried "Easily confused", or "given to rhetorical bludgeons when scalpels would prove better", I'd have to cede the point.

ok -

a serious question - for just a minute, let's forget that kevin is the one who is originally asking the question (to get rid of THAT error - or is it bias)

i want to know how anyone knows whether or not a person who fills out any anonymous survey is telling the truth about his/her answers?

i mean, there is a big difference between asking people about getting kids their shots and asking them whether they had gay sex or ever had sex with anyone besides their spouse during their marriage

the probability people would lie about shots is a LOT lower than they would lie or "misremember" sex they just might could feel ashamed about.

a lot of people don't want to admit to them self that they cheated so why would they admit it on a survey?

how can you "correct for error" when you got no idea what the error actually IS?

Arki -

The article JC references also has in the discussion an alternate possibility ... that is, the relationship between religious belief and infidelity may be (is likely) not one-way causal. I actually disagree somewhat with this section, in that I believe that a spurious relationship is likely -- with my lens, I see religious belief driven by social norms more than vice versa.

Several important limitations of this research underscore the need for
caution in interpreting these findings and the need for further research into
the links between religion and marital infidelity. First and most important,
the cross-sectional nature of the GSS data and the wording of the item tapping
infidelity make it impossible to establish the causal direction of these
empirical associations. Rather, our study has identified significant patterns
or associations between religious involvement and extramarital sexual
activity. Although it seems unlikely that the associations reported here are
spurious, it is plausible that they are bidirectional in nature. For example,
consistent with research on the links between religion and cohabitation
(Thornton et al., 1992), it is conceivable that religious factors influence the
likelihood of adultery, which in turn affects religious involvement (e.g., perhaps
diminishing religious attendance or biblical literalism). Such a causal
process could result in overestimation of the possible religious influence on
infidelity in cross-sectional data. Although we are aware of no existing
cohort-based data that will allow adjudication of these issues, it should be
an urgent priority for future research.

i want to know how anyone knows whether or not a person who fills out any anonymous survey is telling the truth about his/her answers?

You generally don't know for a particular individual. But you have comparative studies -- where using different survey techniques on the same sample, or randomly equivalent samples, for certain types of questions -- that compare the probability of the socially undesirable response for the different data collection methods. From a bunch of these studies, you can say (simplified) "if XX percent say that they did Z using data collection method A that means YY percent would using data collection method B." It won't be as precise as if you using method B in the first place, but it's better than a wild guess. If you compare back to a more accurate or intrustive or expensive measure, you can get a decent estimate using the less reliable method.

The same technique is used for when a variable is measured with multiple instruments, or you need to use a simple method for reason of time or monetary cost for some variable in your study (e.g. MMSE instead of a clinic diagnosis of dementia). There are probably other methods as well, but this is the main idea behind most.

dca,

i'm sorry, but i don't guess i understand your answer. suppose you use 10 different ways to collect data - basically asking 1000 people trying to get the answer to the question - when you were married, did you have any sex with anyone who was not your spouse. i know you ask the questions in a different way. suppose they all agree exactly - 400 people say yes, i did.

HOW do you know whether or not 1 person or half of the people answering lied since you can't check their answers? why would you think that a person would lie/misremember no matter HOW it was asked because it is the same real question

As a long time father of teenaged daughters I'd like to pass along a little helpful advice to those about to go through the experience of raising adolescent females.

When you bury a boy for the first time ... make sure you dig the grave deep--that way you can reuse the grave shaft leaving you enough room on your lawn should you wish to put in a pool or garden.

There's another benefit to this approach. The hardest thing isn't talking to your daughter about sex--it's explaining why your one dog has her boyfriend's rotting femur in his mouth while the other is rolling in the rest of the remains. If you dig deep, you can avoid this uncomfortable conversation.

Just remember the motto: "For a good night's sleep just bury 'em deep."

Do not feel squeamish about doing this if you're religious--you're just helping him make it to heaven before he does something that will land him in hell.

Blest Regards

John

Ha! My wife brought up the Clemens-McCready thing last night, using the exact same quote "how would *I* feel" And I replied precisely as DMN did - it won't stop at 18 or 30, but when I am dead.

Holy crap on a popsicle how did all you guys end up feeling so damned AWFUL and terrible about one of the greatest types of interactions two people on the face of the planet can have? The more comments I read like this the more depressed I get about humanity.

I get the feeling most men would sooner have their daughters punched in the face than ever have sex, apparently about which there is absolutely nothing good or positive at all. Sad, fundamentalist thoughts. And don't give me the crap about not having daughters. I have teenage cousins with whom I am very close. I don't want them abused, but lord knows I kinda want them to have joyful human experiences, of which sex is one of the best.

i'm sorry, but i don't guess i understand your answer. suppose you use 10 different ways to collect data - basically asking 1000 people trying to get the answer to the question - when you were married, did you have any sex with anyone who was not your spouse. i know you ask the questions in a different way. suppose they all agree exactly - 400 people say yes, i did.

Method 1 is ask them if they cheated; method 2 is to hire a PI to check them out. You do this one time, and you get an equivalency formula. The idea is that there is a way to get a more reliable response, but it's onerous and expensive and impractical. But once you figure out the degree of underreporting, you can do the easy way (the questionnaire) from then on. You can also generalize from similar questions, things that you might be able to verify independently.

Or, to use the example of dick size that was brought up early. You ask guys how big they are, then you measure them. They measure 2 inches less on average.* Next survey, you just ask them ... and then you subtract 2 inches.

* the difference isn't due just to lying. most people just aren't that great at estimating the sizes of things, and even an honest estimate may be biased. this takes care of all of that.

Like asking someone how much they smoke. They say a pack a day, adding 1/4 to 1/2 a pack to that would almost always be accurate. And they REALLY BELIEVE they smoke a pack a day, they aren't being dishonest at all, to their perception. A mind is a powerful thing, as far as belief goes.

baseball chick, think of it this way.

One survey asks questions about sexual behavior of, say, 5,000 people. These include whether they're churchgoers, and a bunch of other things like how educated they are, what race or ethnicity they are, how old, and so forth. This is done by having interviewers ask questions in their home and recording the answers.

Running some statistical tests, they see that there is likely a significant difference between the churchgoers and the non-churchgoers regarding infidelity.

A second survey does the same thing to the same number of people, but it's a different group of people and instead of in-person interviews, when it comes to the questions about infidelity they give the person a computer and let them enter their answers without the interviewer see them.

Now, imagine this is done ten different times, with the questions asked differently and the way the information is collected being done differently. Maybe it's on the phone and not face to face. Maybe the person selected in the sample - because people aren't allowed to select themselves to eliminate a source of bias - is interviewed by someone not of their race or ethnicity or sex, because some research has shown that a white male will give different answers to a white male than he will a black woman, or a Latina will answer differently when interviewed by a black woman than by another Latina, and so forth.

And let's say that about 80% of the time, different studies using different methods and different questions show that the churchgoers tend to be more faithful than those that aren't churchgoers, and it's statistically significant. At that point, I think you have say maybe there's something to it, and maybe it's religion. Oh, one thing I forgot to add. The relationship between churchgoing and infidelity is true when you hold all other things equal - it's not race, it's not rural, it's not income that seem to be driving this relationship.

Kirby Kyle made a really good analogy over in the Lounge, where he said he's done hard and social science research. One point he made reiterated a point DCA made, that because there's more noise (bias, variation) in the data than in the hard sciences - like with atmospheric science per DCA - it's harder to be certain.

Kirby's brilliance was on display when he compared that to analyzing fielding stats versus batting stats in baseball. Just because it's harder to measure fielding than it is batting doesn't mean that it should be entirely disregarded. If you've got three different fielding measurements that approach the question differently saying that Derek Jeter or David Ortiz is a terrible fielder then at some point the weight of evidence suggests that, even though there is some uncertainty there, it may be well to make Ortiz your DH.

Lassus Posted: April 30, 2008 at 07:18 PM (#2764626)

I get the feeling most men would sooner have their daughters punched in the face than ever have sex

- grinning

you don't have a daughter. and i would bet you any father would tell you it is not the same.

it isn't your daddy don't never want you to have sex/get married/be loved

it is your daddy KNOW that teenage guys/older guys/middle age guys gonna disrespect you when you are 15 and he remember ALL about it and he don't want some guy hurting his Baby Grrrrrl.

- smile

it is kinda special, you know. because your daddy the ONE man in the world who loves you if your ass fat/your ass skinny/you not pretty/you not smart/you not cool/you not popular/you not whatever - AND he's the ONE man you don't have to worry if he's gonna love you and you KNOW he not gonna screw you and he gonna make sure other males do right by you. you got NO idea how bad a teenage grrl need that. we really DO.

most of the sex you do when you a teenage grrrl got exactly zero to do with "love" and most of the time you think it is something you HAVE to do to keep that boy and you depressed about it. and you worried about what you look like and what he thought you look like and how he probably thinking of some other female or not even about you. or how he disrespects you BECAUSE you let him. and trust me i have talked to a LOT of other grrrls and they sure as heck ain't gonna tell YOU that.

smile

then again i got married when i was a teenage grrrl and i knew he was right because i didn't feel disgusted/depressed/humilited/used.

bigger smile

smartest thing i ever done in my life

Man, the retarded posts are aflowin' like the river Nile tonight. Behold:

(Note that the issue here is the difference in fidelity, not absolute levels of fidelity, so unless religious people are more likely to lie about it, the fact that people will lie about fidelity is irrelevant.)

Nieporent has written some stupid #### in his day, and I mean really stupid ####, like Jose Canseco stupid ####, but he has really outdone himself here. Dave (sorry, you're back to Dave until you start to display the intelligence of an amoeba), how the #### are you going to know how much everybody is lying if you have no objective means of doing so? How? What, you assign values at the beginning how much each cohort is supposed to lie and then go from there? I'd really love to see the reaction to a review panel when you try to slide this passage past a review panel:

...the authors believe that enrollees of every cohort were not truthful in their answers to the study questionaire, or the extent to which the results of each cohort were distorted by this unthruthfulness. Nevertheless, despite this flaw, we believe the conclusions reached in this study to be valid and reliable.

You'd get laughed off the planet for that.

Now this:

443. DCA Posted: April 30, 2008 at 03:56 PM (#2764490)

Of course they're unscientific. That's not a criticism.

What???? Of course it's a criticism. If data is generated in a non-scientific manner; that is, an objective manner, than the resulting conclusions are nothing but an opinion based on unsubstantiated and unreliable information.

Wow. And I mean WOW.

Let's move on:

445. Dave Posted: April 30, 2008 at 04:16 PM (#2764512)

JC in DC cited such a study, at which point kevin summarily dismissed it because it was a social science study.

More misreading. I criticised the methodologies!!!! Let me write that again, so you don't get confused: M-E-T-H-O-L-O-G-I-E-S. There. Can you read that now?

And along the same vein:

452. JC in DC Posted: April 30, 2008 at 04:49 PM (#2764550)

Social science is science.

Noooooo. Science is science. The scientific method can be applied to social studies and be called scienceas long as it adheres to fundamental scientific principles! One of those principles is: Devise experiments to test hypotheses and all valid scientific hypotheses must be testable. Did the authors follow up to test whether the interviewees were telling the truth about their sex lives? I think we both know the answer to that.

And JC staggers me with this one:

I had a nice exchange w/the author of the study Kevin so cavalierly and ignorantly dismissed.

What? You think the author of a paper is going to fess up to flawed methodological process? At the very least, you should have called someone else who didn't have a vested interest is supporting the data. That's called o-b-j-e-c-t-i-v-i-t-y. This statement highlights in bold relief how badly JC understands how science operates and what valid science is all about.

JC goes for the trifecta here:

Which is like a nuclear engineer saying of chemistry that it's just morons playing with beakers, that chemistry is the science failed scientists get into: dismissive and insulting.

Science is science, regardless of what subject matter one happens to be studying. All that statement reveals is the egotism of the phycisist he was talking to. It has nothing to do with what we are speaking of here, which is what is science and what isn't. Again, it underlines how badly JC misunderstands the nature of the discussion, and the nature of science itself.

DCA writes:

Several important limitations of this research underscore the need for
caution in interpreting these findings and the need for further research into
the links between religion and marital infidelity. First and most important,
the cross-sectional nature of the GSS data and the wording of the item tapping
infidelity make it impossible to establish the causal direction of these
empirical associations.

RDF.

And the walls come a tumblin' down.

That sure is a definitive study, JC.

DCA, I mean no offense whatever, but as I am sometimes guilty of when writing of my own field, the jargon in your post 459 is largely incomprehensible. If you'd care to rephrase, I have little doubt I'll enjoy engaging with your thoughts. Thanks.

So, flawed data is useful if it agrees with prior subjective observation (the "one good eyeball" test), but useless on its own?

vi, observable physical manifestations of drug abuse are forms of objective data. We have two types here: 1) Clemens physical changes in his late 30's and 2) his unique late career performance surge. FWIW, one of the hallmarks of a valid hypothesis is the ability to use it to predict observable phenomena. I predicted Clemens was using over 2 years ago.

What? You think the author of a paper is going to fess up to flawed methodological process? At the very least, you should have called someone else who didn't have a vested interest is supporting the data. That's called o-b-j-e-c-t-i-v-i-t-y. This statement highlights in bold relief how badly JC understands how science operates and what valid science is all about.

kevin, for the third time the researcher who wrote that paper had nothing to do the survey that collected the data she analyzed. The survey that she drew from is highly regarded by pretty much everyone in the universe who understands the nature of social science research, which you clearly don't. You're making wildly incoherent arguments that ignore the fact that uncertainty exists in almost every single bit of data collected, including in the hard sciences where error may creep in.

I'm done. You really don't get it.

Noooooo. Science is science. The scientific method can be applied to social studies and be called scienceas long as it adheres to fundamental scientific principles! One of those principles is: Devise experiments to test hypotheses and all valid scientific hypotheses must be testable.

Actually, they must be falsifiable, not testable. Not the same thing at all.

You're embarrassing yourself even further, Kevin. Nobody believes you're a scientist any more than you're a naval aviator.

More misreading. I criticised the methodologies!!!! Let me write that again, so you don't get confused: M-E-T-H-O-L-O-G-I-E-S. There. Can you read that now?

Uhhhhh.... I'M confused. Those are two different words.

OK, normally I find snarking on typos to be rather pathetic, but this was too difficult to pass by uncommented upon.

kevin, for the third time the researcher who wrote that paper had nothing to do the survey that collected the data she analyzed.

So what? She wrote the ####### paper using the data. You think she's going to tell you how much the data sucks on a paper she chose to base her study on?

And anyway, she wrote in the passage DCA posted how unreliable the data is so, if even she (the author) is backing away from any firm conclusions, I don't know where JC gets off saying how definitive it is.

The ####### server keeps kicking in and out on me and prevents me from correcting typos, Lassus.

So what? She wrote the ####### paper using the data. You think she's going to tell you how much the data sucks on a paper she chose to base her study on?

Isn't it standard in academic papers to acknowledge potential weakness in the research and areas where further research is needed?

Or in your experience does one assert, attack, dissemble, obfuscate, and deny uncertainty and call it science?

From where I sit comfortably, it sure as hell looks like it.

And anyway, she wrote in the passage DCA posted how unreliable the data is so, if even she (the author) is backing away from any firm conclusions, I don't know where JC gets off saying how definitive it is.

God you're an idiot. She's actually exhibiting the scientific method you're talking about; iow, saying, "Here's the thesis, this is what the data shows and this is how I interpret the data. We must be cautious, these things are instruments of measurement, and we know how those work. They are not the things themselves, but instruments to measure them and as instruments, they have limitations. Here are those limitations."

You're so dense about this as scotto and others have pointed out you continue to confuse the distinction b/w the gathering of the evidence and the interpretive model. She didn't gather the evidence. I didn't contact her to ask about her evidence. I contacted her to discuss the method of interpretation and the problems in social scientific research. Unlike you, she's a good scientist, was utterly candid about the limitations of social science (much as she was in the passage you bolded and which in your bottomless ignorance you thought somehow counted against her), but stuck by her conclusions (which she also acknowledged were open to study and refutation by others).

And I love how you easily impugn her character, implying she'd falsify the soundness of her approach. You're a scientific scoundrel, a charlatan. And I say that in all seriousness. We're supposed to play nice about you and your scientific aplomb, but it's ironic you have no qualms about attacking her as too interested to be honest about her study's limitations. Guess what? You were wrong about that.

Isn't it standard in academic papers to acknowledge potential weakness in the research and areas where further research is needed?

Yes, it is. And since she is cautioning everyone to not draw any firm conclusions, I don't know where JC is coming from saying how definitive the study is.

Or in your experience does one assert, attack, dissemble, obfuscate, and deny uncertainty and call it science?

From where I sit comfortably, it sure as hell looks like it.

You need to change you seat then to someplace a little more uncomfortable because nowhere have I obfuscated or denied uncertainty. And there is nothing wrong with making assertions, attacking faulty methodologies or conclusions drawn from faulty data or dissembling a study to find fatal weaknesses. That's what the peer review process is all about.

AP:

Mindy McCready has “caught her breath” and gone back to work on a new album following her admission of a long-standing relationship with Roger Clemens, a representative for the country star said Wednesday.

“The first day was really difficult for her. She really has caught her breath,” her management consultant, John Dotson, told The Associated Press on Wednesday.
...
Dotson said McCready will be meeting with a New York-based public relations firm this week to “see what to do next.”

“We want to be sure we don’t inadvertently do something wrong,” he said.

So which tab/mag is going to get the exclusive, far-ranging interview w/ flattering cover shot for next week's issues?

God you're an idiot.

JC, I have to say that typically you are careful to characterize assertions made by people with which you disagree with as idiotic, but here you take it to the next level by calling kevin an idiot. I'm sure you did this after considering it.

Weight of evidence in this thread suggests (shakes magic 8 ball) Yes!

I've had enough of this, but this time I'm really done. Good night!

We must be cautious,

Then why aren't you?

God, you're a babbling fool. You haven't been cautious at all. In fact, I'm beginning to think you don't even know how to read a scientific paper, after reading the crap you have been spewing in this thread.

And I love how you easily impugn her character, implying she'd falsify the soundness of her approach.

WTF?? It's human nature to pimp your own stuff. You don't have to drink the kool-aid, you know. Especially after she's already warned you not to take it too seriously.

Criminy. This thread is gettng more surreal than the one where Treder said Champ Summers had a similar career surge to Bonds.

That's what the peer review process is all about.

I may know what I'm talking about, unlike you, but I'm an idiot for replying again.

So, you assert her study wasn't peer-reviewed?

Which is like a nuclear engineer saying of chemistry that it's just morons playing with beakers, that chemistry is the science failed scientists get into: dismissive and insulting.

My wife, a nuclear engineer, says this about me, a chemist, plenty.

Weight of evidence in this thread suggests (shakes magic 8 ball) Yes!

The tried and true French approach to warfare. Get you ass handed to you, declare yourself the victor, pack up your broken equipment and go home.

Social Science and Science are both limited.

Social Science has some variables that are difficult to measure with any sort of accuracy. However, due to the focus on impact and the tools of the fields, there are a lot of people who are talented at practically applying the findings.

Science often has a great degree of accuracy, however the impact of the science in a societal context is often neglected. That's why we get the "Guided missiles, misguided men" dynamic that Dr. King referred to and has only gotten more extreme since his untimely death.

The fielding analysis/cheating analysis analogy doesn't work for me. With fielding, as with the black-hating voters and the mysterious shrinking genitalia, the data points are all known and verifiable. For baseball, you can read the official counts of errors and chances and such, and if that's not good enough for you, you can get the PBP data, or even go into the tape room and watch every play and graph 'em all yourself. If you want to check the election results, you can re-count the responses and re-count the votes. If you want to measure the all-important penis question, you have a bunch of photos of people's meat next to a ruler.

With an infidelity study, all you have is a series of educated guesses about how often people lie. You don't have the same kind of documentary evidence as in the other cases. For the two to be truly analogous, you'd need to have supremely talented investigators follow each and every one of your subjects for their entire lives, meticulously documenting each and every one of their couplings. You'd still have sources of error (from determining whether a particular act did or did not qualify as cheating, for example), but at least that's an error of interpretation, rather than an error of collection.

We haven't even addressed the really hard question yet, either: How big is the gap between the number of people who say they're religious, and the number who actually are religious? At least cheating is a quantifiable physical act. How on earth are you going to figure out what people think about God, when most don't even know themselves, and the answer can experience dramatic shifts within a person's life as they re-interpret their own experiences (using a thousand different sets of standards to do so, I might add)?

Looks like The Daily News is at it again:

Several sources told the Daily News Wednesday that Clemens had a relationship with Paulette Dean Daly, a former wife of champion golfer John Daly.

The sources said Clemens, a married father of four, arranged trips to Anaheim Stadium for Daly - the latest woman to emerge as an alleged Rocket flame - to watch him pitch forthe Yankees against the Angels.

I apologize for calling Kevin an idiot. That was unnecessary. And with that, I'm done with the discussion.

There's audio here of McCready appearing on a radio show. It's a few minutes long. They ask her about the Clemens story and again she oddly says that she "cannot refute" "the story," yet she won't discuss any details and won't confirm or deny it.

They ask her about the Clemens story and again she oddly says that she "cannot refute" "the story," yet she won't discuss any details and won't confirm or deny it.
Anything to keep her name in the news whilst she's in the recording studio.
If she laid out details and got Clemens to confess, then she's yesterday's news by tomorrow.
If she keeps the kettle a-roiling, then her name remains out in public and her upcoming release will get lots of free press.
Just one theory.

The latest comments from Hardin:

NEW YORK (AP)—Roger Clemens’ lawyer says he will talk with his client about whether to press ahead with a defamation suit following a wave of unpleasant publicity in the wake of reports linking the pitcher to numerous women.

“He’s getting pummeled,” attorney Rusty Hardin said Friday.

There hasn’t been any change in plans,” Hardin said. “Everybody keeps asking these questions. We’ll sit down and see what his views are.”

The decision on whether to drop the suit rests with Clemens.

“That’s always a decision the client has to make. That’s not the lawyer’s decision,” Hardin said. “I’ve never seen somebody get beat up like this. In some ways, I think we’re on uncharted ground.”