Slashdot Mirror

No. First, let us look at this little tidbit:

Having taken into account other factors such as alcohol or tobacco dependency or other drug use, as well the number of years spent in education...

So there was manipulation of the data to exclude the effect of these "other factors", which completely throws out any correlation that these could/would/should have. It would be akin to testing if teen pregnancy lowered IQ, but they threw out data belonging to private school girls.

This is a very common thing to do in medical studies. It's called correcting for confounding factors. If they hadn't done this then the results would have been less useful, as it could be claimed that the observed effects were the result of something else already known to negativly impact IQ.

they found that those who persistently used cannabis - smoking it at least four times a week year after year through their teens, 20s and, in some cases, their 30s - suffered a decline in their IQ.

This is plain bad science. These people they are studying are CHRONIC users. They are likely using right up to the morning of their "interview". It is like the kid who started smoking cigarettes at 8 years old vs. someone starting at, say 23. The former is most likely to smoke 2+ packs a day. The latter usually smokes less than one pack. Also, nothing has been done to show what happens when they would stop.

First off, why is it "bad science" to test the effect on chronic users? As for what would happen when they stop, from the abstract:

Further, cessation of cannabis use did not fully restore neuropsychological functioning among adolescent-onset cannabis users.

STATS.org has a nice, details, scientific-sounding article debunking a lot of anti-BPA reports out there, and appears to come from a legitimate source (George Mason Univ.).

I'm not a chemist/biologist/doctor, so I have a hard time judging whether the article is bunk or not.

Care to weigh in?

"...While the Lancet numbers are shocking, the study's methodology is not. The scientific community is in agreement over the statistical methods used to collect the data and the validity of the conclusions drawn by the researchers conducting the study. When the prequel to this study appeared two years ago by the same authors (at that time, 100,000 excess deaths were reported), the Chronicle of Higher Education published a long article explaining the support within the scientific community for the methods used.

"....And even as the Associated Press reported mixed reviews, all the scientists quoted in its piece on the "controversy" were solidly behind the methods used."

So yes, it was eviserated by some spin meiesters and morans who think they know statistics because they can count till twenty without taking their shoes off.

Various well know academics/statisticians have gone over the study. I have not heard a single voice say the sampling/logic/methodology of the cluster analysis was flawed. Quite to the contrary, every on has said the study is solid.

It should be noted that people under 25 are disproportionately conservative

It should be noted that Republicans disproportionately refuse to answer polls in general for a variety of reasons, and that polls taken over the weekends in particular favor Democrats.

My guess from looking at the trends and comparisons to past voting patterns and their relationship with polls at that time is that Bush has an overall 4-6% lead, and it's growing.

And this doesn't even consider the fact that using the 2000 party IDs for resampling is almost certainly too conservative! Democrats will be out in force this year, certainly much more than in 2000 when they were in full retreat after the Clinton scandal. They hate Bush with a passion and will vote vote vote.

I would not be surprised if Bush gets 55-57%, unless the course of the election changes dramatically in the next 6 weeks.

Holy crap! You know that You've shown ineptness in understanding polling trends three times now, so it's time to dismiss you.

from stats.org:

Gulf war illness
The "Gulf War Syndrome" allegation is that 40,000-200,000 veterans suffer from a single identifiable illness, Gulf War Syndrome; from the end of the war to November 1995, 2,900 vets had died, by May 1996, the number skyrocketed to 4,291, and by 1997, 10,000-12,000 vets had died
Derivation: The original sources for the number of veterans suffering from Gulf War illness is unknown, and the numbers cited often change. Activist Denise Nichols propogated the first two fatality figures, splicing two different numbers. The first number was deaths from all causes among veterans of Operation Desert Storm itself, about 700,000 people. The second number was deaths among anyone who had served in the Gulf region since 1990, over a million people. The third figure was fabricated and spread by activist and militia member Joyce Riley von Kleist.
Actually: Despite millions of dollars in federal research dollars, there is no evidence of an all-encompassing syndrome affecting Gulf War veterans. Actual illness and fatality rates among veterans are lower than those of the general public for natural causes, and higher for external accidents. As of May 1998, approximately 5,425 veterans (0.78%) out of 697,000 participants in the Gulf War had died. The Defense Manpower Data Center (DMDC) obtained this number by comparing lists of all Gulf War veterans with files of deaths recorded by the Social Security Administration. A similar comparison disclosed that, of the 2,372,327 members of the active duty force and the selective reserve who did not deploy to the Gulf, 19,475 (0.82%) had died. No information on cause of death was available from this data search.

Here.

Your child has about 1 in a million chance of being abducted by a non-family member.

Ed.

I couldn't find the original study...

He did indeed source it: though he did not mention Blair in the UN speech itself, he did brief the British press about it afterwards. And he is rather embarrassed about it now.

How many would you consider `acceptable', especially in the absence of any studies substantiating Mr. Herold's claims? I have just answered this in detail on the other sub-thread where you asked it, though you probably haven't read this yet.

Or failed to answer it, as the case may be. In either case, my response to that answer is here.

In this post you directly alleged that the only missile Iraq fired on Kuwait was the single Seersucker which impacted outside a shopping mall in Kuwait City. The Iraqi Information Minister claimed the same thing, but every single news outlet referenced in that thread including the ones you posted links to claimed otherwise.

That post was by glrotate. There are many links in that thread to news reports of Scuds landing in Kuwait: all have the US Defense Dept as their source, none are verified. The military would not let anyone see all these alleged Scuds. I certainly can't blame a reporter for quoting US Defense Dept briefings: that's news. But I won't believe the reports until I read that the press went and actually saw 14 impact sites with fragments of Scuds.

In either case, in that thread I posted links to a number of reports on those firings, not all taken from DoD sources, including an eyewitness account from a CNN news crew. Interestingly enough, you did too .

Video of several of those intercepts was also taken by several news agencies, including al Jazeera.

I've yet to see a single actual statistician consider Mr. Herold's work as valid, but perhaps you could give us a counterexample? Even one?

A man is innocent until proven guilty. You have attacked Professor Herold's work, not me: prove his guilt, and I will accept it.

I've just shown you several statisticians and even a trade group of statisticians which consider Mr. Herold's work to be nonsense. Can you show me any that have confirmed his work? Any?

I can only recall one, that of STATS: please remind me of which other statisticians you quoted? Remember that the Iain Murray article is in fact a reprint of his original piece on the STATS site.

Can you show me a single statistician who has defended Mr. Herold's work? One?

There we have it ladies and gentlemen -- while many of you thought that Iraqi Information Minister Mohammed Sayeed ``No American Tanks within 100 Miles! What's that noise?'' Sahaf was inflating civilian casualty statistics, jdfox has uncovered the real truth: he was actually underreporting them! What a brilliant thesis...

Of course, there are plenty of concrete reasons to doubt the statistics provided by Womens' Studies (should that be `Wimmin's studies'?) professor Marc Herold:

• the fact that Mr. Herrod was so spectacularly (and famously) wrong in Afghanistan, reporting more than three times as many civilian casualties as the sources he claimed to have gotten his statistics from, and more than twice as many civilian casualties as the Taliban themselves did (see the section on civilian casualties at the end of that article)
• The fact that Herold's statistical methods have been thoroughly debunked, which is no surprise, as he has no training or background in statistics at all.
• The fact that he steadfastly refuses to explain where he gets his numbers, despite repeated demonstrations that he is double- and even triple- counting actual incidents, and accepting other incidents as genuine based only on the allegations of Taliban and Iraqi Information Ministry sources.
• The fact that the non-partisan Statistical Assssment Service, a group of professional and academic statisticians formed to combat the incorrect use of statistics in the media has examined Mr. Herold's methods and found them fatally flawed.

So is Relativity, or Darwinian evolution. This does not help your argument, you imbecile.

His point did a nice job of helping his argument. He mentioned that human caused global warming is far from being proven. Therefore, we shouldn't jump on the stop-global-warming-bandwagon blindly. Instead, we should instead give it careful study, and weigh out future consequences of our actions.

As does almost any area of study in science, from cold fusion to ESP. But the great majority of evidence is behind it.

What the hell? Where did you get this "great majority of evidence" supporting human caused global warming? The Sierra Club of San Fransisco? This issue is FAR from reaching any conclusions, as this site explains quite well.

An analogy: you are trapped in a bank vault with ten other people...the vault is airtight...blah blah blah...so you light up a great big cigar.

I'll let that analogy speak for itself.

the vast majority of the world's scientists accept that global warming is a serious problem and to a significant extent caused by human activity

Have you ever spent any time reading scientific journals on global warming? Or do you get your information from quick, eight paragraph articles from the New York Times?

Your post shows that you have a good understanding of political debates, but a terrible understanding of the current state of science. One good way to fix this problem is by impartially and properly researching complicated issues like this in the future. Then you can back up your ideas without sounding like a complete idiot.

It was referred to as DHMO. Dihydrogen monoxide. First mentioned in the Washington Post--here's a link to an article about it, from the same site I posted earlier in the article about misleading statistics.

In an attempt to figure out how statistically significant the article's 6-in-a-million chance of the asteroid hitting earth is, exactly, I ran a search on the most popular statistic--the odds of being hit by lightning. Turns out there's even controversy about that. The odds cited range from 1 in 10,000 to 1 in 5 million. So was this asteroid statistically significant? Turn to Mark Twain for that one.

The whole "on collision course" phrasing thing was, in my opinion, poor choice of a headline, but news is a product just like any other media item, and sensationalism sells.

(Thank you Samuel Clemens.) Well, fortunately, it seems as though a lot of people who use those doctored statistics often wind up hoist on their own petard. Looks like WorldCom's getting there, as was Child Find in an article in the Denver Post that won a Pulitzer Prize for reporters Griego and Kilzer.

There are some spectacularly bad examples in the posting above... I'm not sure anyone ever said there were a million homeless people. However, the widely-criticised (as to methodology) US census survey cited almost half a million, which you can add for yourself here. Also, as to the "statistics" quoted by the poster on sexual orientation, I know that as early as 1972 the University of Guelph's Veterinary and Agricultural Colleges were using the 10% figure in training films (one of which, my friend, a student in another department at the time, narrated) on animal breeding, and in Animal Days, the British naturalist Desmond Morris mentions something similar based on his work with ten-spined sticklebacks (1958). Similar figures seem to hold through all animal species.

The problem seems to be that too many of the general public fall for that same old Ad Verecundiam Fallacy. I think it's a lack of critical thinking skills.

And in this day and age, if a CEO doesn't qualify as an "improper authority"... --smirk--

There's a nice article about the conflicting statistics relating to guns here.

If "childhood" is under 21, then the number is 12 a day.

But who cares? What possible relevance could it have to the argument. Even if it is one a day, the point is still true: Guns kill more people every day than Dimitry's code ever will. So why is the gov't so obsessed with code?

Actually, I was suggesting that your chance of buying a lotery ticket and winning on that ticket was very low. As for numbers, I'll give you what I have... Here in Georgia, the chance of winning our "Big Game" lottery with the maximum prize is 1:76,275,360. If I assume that you buy 1 ticket each week of your life, and I assume you live for 80 years, that's ~4160 tickets... All things considered, 4160:76,275,360 odds are still not good. As for the odds of getting hit by lightning in your lifetime, according to Stats Spotlight, "the odds of being struck by lightning are 1 in 709,260". The probability of being struck twice is simply 1:1418520. Unless I miscount, that is STILL much smaller than the chance of being struck twice in your lifetime by lightning.