Oh boy - another one! I need to start wearing a shirt that says "Kick Me - I Smoke!", as smokers seem to be the group du jour for kicking. Be that as it may be, let's look at this "study"...
I'm not going to pick on their numbers, nor am I going to claim they were ax grinding, although both could certainly be suggested. Rather, this study offers a perfect opportunity to demonstrate a principle of statistics that most folks blithely ignore.
I guarantee you that I could conduct a similar study in any major American city and reach the following "conclusions" from my numbers: Blacks perform worse at work than non-blacks. Blacks are also more likely to have a less than honourable discharge, to be demoted, to desert, and to earn less than their non-black colleagues. It would be a legitimate, random sample study, too, and I wouldn't fudge the data. My numbers would "prove" my conclusion.
But of course, the implications of the conclusion that blacks are lesser human beings would be laughably wrong. Why? Because I would've violated the prime directive of statistical science:
Correlation does not equal causation.
Do you think anybody'd publish my study? And if they did, would it be taken seriously? Or would I be attacked as a bigoted fool?
Why would they so easily pick up on the flaws of my "conclusive proof" and yet miss it so readily in this instance? Who's really grinding that political ax: the researchers or the media who report their results? Or both?
Inquiring minds want to know.
Smokers perform worse at work than non-smokers, finds a study of US navy female service members published in Tobacco Control. Smokers were also more likely to have a less than honourable discharge, to be demoted, to desert, and to earn less than their non-smoking colleagues, the study showed.
(link) [EurekAlert]/Politics | 1 writeback | permanent link
On 3/30/2007 07:12:05
Arwin wrote
Seems bogus
comment...