DADT: Johhn McCain knows nothing about statistics

[RobinFinderson]

Originally, John McCain said he would oppose allowing gays to serve openly in the military until the generals supported that position.

When the majority of high brass came around to supporting gays openly serving in the military, John McCain then went on to oppose this policy on the ground that the troops hadn't been given the chance to offer their input. He demanded a study.

Now the study has been done, and it shows that accepting openly serving gay men and women would not affect combat readiness. So what is McCain's beef now?

In addition to my concerns about what questions were not asked by this survey and considered in this report, I'm troubled by the fact that this report only represents the input of 28 percent of the force who received the questionnaire, including completely leaving out a numerous members of the military in combat areas. That's only six percent of the force at large. I find it hard to view that as a fully representative sample set.

He doesn't like the survey, because it only polled SIX PERCENT of the force at large. In terms of a representative sample set, six percent of the total military force is a HUGE number. It is such a huge number that it easily predicts, with 99% confidence, the overall military attitude about gays serving openly to within a fraction of a standard deviation from the mean.

So... either John McCain doesn't understand statistics, or this one-time-almost president of the country is just a class-A bigot.

[Jeff Lebowski]

You beat me to it. I was going to start a thread about what a POS John McCain is for his role in fighting the elimination of DADT. He is the George Wallace of our time.

Here is a brilliant piece from the Daily Show on McCain and DADT. The bit at the end is best when they spoof the "It Gets Better" PSA's to show McCain that "It Gets Worse" due to his future spot in the history books as a bigot. Funny and spot on.

EDIT: Oops! I forgot the link. Here it is:

http://www.thedailyshow.com/watch/mo...gets-worse-psa

[ERCougar]

I'm not going to defend DADT, but doesn't excluding all combat personnel at least cast a little doubt on the study? Isn't combat the setting where we most care about "stability"? Wouldn't it be possible that the demographic in combat is different from the demographic at home? It's not a numbers thing--it's the question of representative sample, and not being in the military, I have no idea how valid the objection is.

Just playing devil's advocate here.

[Moliere]

Originally, John McCain said he would oppose allowing gays to serve openly in the military until the generals supported that position.

When the majority of high brass came around to supporting gays openly serving in the military, John McCain then went on to oppose this policy on the ground that the troops hadn't been given the chance to offer their input. He demanded a study.

Now the study has been done, and it shows that accepting openly serving gay men and women would not affect combat readiness. So what is McCain's beef now?



He doesn't like the survey, because it only polled SIX PERCENT of the force at large. In terms of a representative sample set, six percent of the total military force is a HUGE number. It is such a huge number that it easily predicts, with 99% confidence, the overall military attitude about gays serving openly to within a fraction of a standard deviation from the mean.

So... either John McCain doesn't understand statistics, or this one-time-almost president of the country is just a class-A bigot.

HOw can you make any claim without knowledge as to how the survey was taken? If they sent the survey to a random selection of people and those who didn't return the survey were replaced with randomly selected alternatives until they got a statistically significant sample of returned surveys, then I agree with you. HOwever it seems like that was not done and I would expect the participation to be higher for those who would like to overturn DADT. Positive confirmation is more difficult to get from people who want the status quo.

[Jeff Lebowski]

I'm not going to defend DADT, but doesn't excluding all combat personnel at least cast a little doubt on the study? Isn't combat the setting where we most care about "stability"? Wouldn't it be possible that the demographic in combat is different from the demographic at home? It's not a numbers thing--it's the question of representative sample, and not being in the military, I have no idea how valid the objection is.

Just playing devil's advocate here.

I am not sure where this "no combat units" thing came from.

http://www.cbsnews.com/8301-503544_1...09-503544.html

He pointed out that 67 percent of Marines and nearly 58 percent of Army soldiers in combat units believe that repeal of the law would have negative consequences on unit cohesion in a field environment or out at sea.

The survey found that personnel who reported having worked with a gay service member were much more comfortable with the repeal. Of those servicemembers, 92 percent said their unit's ability to work together was not negatively affected by their colleagues' homosexuality.

The arguments for keeping DADT are the same arguments that were used to segregate units by race. It's horseshit.

[Jeff Lebowski]

HOw can you make any claim without knowledge as to how the survey was taken? If they sent the survey to a random selection of people and those who didn't return the survey were replaced with randomly selected alternatives until they got a statistically significant sample of returned surveys, then I agree with you. HOwever it seems like that was not done and I would expect the participation to be higher for those who would like to overturn DADT. Positive confirmation is more difficult to get from people who want the status quo.

On top of that, McCain and other senators said they were concerned that only 28 percent of the military personnel who were sent the "don't ask" questionnaire returned it, representing just 6 percent of the armed forces at large. Army Gen. Carter F. Ham, one of the leaders of the Pentagon study, told the senators that the results were "well within the historical range of DOD surveys of department personnel" and that each category of questions had a statistically significant number of responses.

28% response sound pretty high to me, based on surveys I have been involved with.

Either way, this is a fundamental issue of fairness and civil rights. This type of polling should have no bearing on the decision.

[Moliere]

The survey found that personnel who reported having worked with a gay service member were much more comfortable with the repeal. Of those servicemembers, 92 percent said their unit's ability to work together was not negatively affected by their colleagues' homosexuality.

This is why it should be repealed regardless of the outcome of the survey. There is a natural stigma against gays and that stigma goes away once someone knows/works with some of them. Take away DADT and I would imagine in 5 years there will be a report that says the repeal of DADT had no effect whatsoever on combat readiness. Just my thoughts.

[Moliere]

28% response sound pretty high to me, based on surveys I have been involved with.

Either way, this is a fundamental issue of fairness and civil rights. This type of polling should have no bearing on the decision.

I agree (see my post above) but I was laughing at RF's 99% confidence level assertion and what not. I'm no statistician (cue Indy) but I used statistics in my former life and the method above (which I guess is not fully known) seems more intent on getting a general feel more than statisically asserting some level of confidence in the responses.

[mUUser]

In the next generation, our kids will shake their collective heads as to how we handled gay and lesbian issues during this time, just as we shake our heads at how our parents handled minority issues, and how they shook their heads with how their parents handled gender issues.

DADT is dead policy. Repeal it already.

[RobinFinderson]

I agree (see my post above) but I was laughing at RF's 99% confidence level assertion and what not. I'm no statistician (cue Indy) but I used statistics in my former life and the method above (which I guess is not fully known) seems more intent on getting a general feel more than statisically asserting some level of confidence in the responses.

If you were a statistician you might remember that a 99% confidence interval could be constructed from just about any N-sized sample of a population. If the N is very small, the 99% confidence interval would become very wide in order to compensate for all of the unknown information and capture the population mean. Polling 6% of the armed forces represent a HUGE sized sample, which would result in a very narrow 99% confidence interval. 6% of the total population represents the kind of statistical dream that could only come true on a Pentagon sized budget.

What you and Senator McCain seem to be challenging is basic concern with statistical sampling -- randomization. McCain, specifically, takes issue with the fact that the poll was not given to people IN combat areas. The poll was, however, given to people who serve in combat positions, just not when they were IN a combat area. I guess it didn't make sense to have them put down their weapons and take a poll while under the threat of taking enemy fire. I also guess that McCain thinks that the troops who took the poll while stateside would change their mind as soon as they were in a combat situation.

Challenging the survey on the grounds of randomization is tantamount to asking a brick-layer, after he built a wall, whether he used mortar in the joints. McCain's challenge assumes the DOD is COMPLETELY PROFESSIONALLY INCOMPETENT.

[statman]

I'm not going to defend DADT, but doesn't excluding all combat personnel at least cast a little doubt on the study? Isn't combat the setting where we most care about "stability"? Wouldn't it be possible that the demographic in combat is different from the demographic at home? It's not a numbers thing--it's the question of representative sample, and not being in the military, I have no idea how valid the objection is.

Just playing devil's advocate here.

Combat troops are the ones where it matters the most. Those are the ones they're really worried about. Other military functions are more or less like any other type of job - where you just tell people to get over it and do there job. That's harder to do with combat troops...

[RobinFinderson]

Combat troops are the ones where it matters the most. Those are the ones they're really worried about. Other military functions are more or less like any other type of job - where you just tell people to get over it and do there job. That's harder to do with combat troops...

The survey did NOT exclude combat troops. It simply wasn't administered in combat areas. Combat troops were indeed polled.

[statman]

The biggest issue was that it was a voluntary response survey. With a hot-topic like DADT, it's going to be FULL of self-selection bias. Opinions of responsers are NOT going to be indicative of the opinions of the general population.

[Indy Coug]

Give it up. Robin Finderson is well-versed in survey methodologies.

[Moliere]

Give it up. Robin Finderson is well-versed in survey methodologies.

Did he learn that in architectural school?