lies, damned lies, and epidemiology


[ Follow Ups ] [ Post Followup ] [ Cluster Headaches Messages ]

Posted by lc bob on April 05, 1999 at 16:29:11:

In Reply to: just wondering posted by alex raffini on April 05, 1999 at 02:37:35:

... apologies to Prime Minister Benjamin Disraeli

"study", "investigation" and "analysis" have become 4-letter words for me today, thanks to commercial statisticians who sell their "expertise" to the highest bidder.

my disgust with this "profession" is contained in the anecdotal story of a recent scrutiny of the effect of various lighting on productivity. first they recorded current productivity. then they turned the lights up. productivity increased. then they turned the lights up again. predictably (heheh), productivity rose again. THEN they turned the lights DOWN. productivity ... INCREASED ... again. not to be fooled, they turned the lights down AGAIN. AGAIN, productivity INCREASED!!! omigod!! (turned out that productivity was driven by the employees' paranoia at seeing a bunch of white-frocked, clipboard-bearing "studiers" walking about them throughout the "study"!)

careful what you study - you just might fall into a vat of manure and decide the stink caused your result, when what stinks is your result.

i'm inclined to believe that the "smoking, drinking, divot-faced hazel-eyed 20-something muscular male" prototype CH victim is more effect than cause. in other words, whatever causes CH may also make one more likely to smoke and drink and party (and have zits?) than are the partying, drinking and smoking likely to cause CH. subtle but important difference. i believe we rely on statisticians to supply the distinction for us. they do not - to their discredit.

damn, i'm in a bad mood today, huh?
oil well ...
-lcb



Follow Ups:



Post a Followup

Name:
E-Mail:

Subject:

Comments:

Optional Link URL:
Link Title:
Optional Image URL:


[ Follow Ups ] [ Post Followup ] [ Cluster Headaches Messages ]