![]() |
|
Posted by Brian in Columbus, GA on February 15, 2002 at 17:10:54:In Reply to: Re: A new study on Sleep Requirements? posted by Kevin (in Walla Walla) on February 15, 2002 at 10:32:29:
Bad study design. Bad data analysis. Bad article...BAD JOURNAL?
Since my earlier post, I found the article in the newspaper, and I was appalled that this study was published in a scientific journal.
The study violated at least two major rules of statistics. (1) A study should be designed to test a specific hypothesis. This study was based on a survey about cancer risks. Apparently, they collected a lot of data peripheral to cancer risks and went fishing with the extra data. When you fish long enough and with enough different baits, you are bound to catch a fish. The "fish" here is the link between long sleep and mortality. (2) The survey subjects were not randomly selected - they were mainly frieds and relatives of volunteers for the American Cancer Society. I wonder how many of them actually had or were recovering from cancer? Who knows? In any case, nonrandom sampling alone renders the study nonscientific and meaningless. How could a supposedly scientific journal publish such crap?
So, what's the big deal? The big deal is that there is dangerous misinformation out there. My local paper led with the headline "Too much sleep may be deadly". It began: "Don't fret if you don't get eight hours of sleep a night - new research suggests adults live longer if they get six or seven." It is not until seven paragraphs into the article than any flaws in the study are mentioned. And the fact that the study is flawed to the point of being meaningless is never made clear. Also not mentioned in the article is the fact that most Americans are sleep-deprived, contributing to reduced productivity, illness, and accidents. Bad science is dangerous.
Archived Apnea Forum85 viewing only. To post a new topic go to the Apnea Forum Homepage
|
Copyright ©1995-2005 Sleepnet.com., All rights reserved