Re: [ossig] (Fwd) STI News: Firm discovers good IT help hard to find

[Dr Molly Cheah <drcheah@pc.jaring.my>]

> your statement above does beg the following question ...
> 
> Does your statement apply to the usefulness/'meaningfulness' of the
> large number of "statistical" reports/surveys we read daily where the
> variables are probably just as wide ?
> 
That's why statistics are sometimes related to "damn lies" :) It depends
on how you interprete the results.

> eg:  i remember reading one report which linked "high variability in an
> infants weight in the first year" to obesity in later adult life ?
> (maybe this is not a good example ... but i believe you get drift here)
> 
Well, if you think that there is a correlation between infant weight and
the "number of fat cells" i.e. that the weight is a good surrogate
measurement of fat cells, I guess I would accept the correlation if
there is statistical significance of say P <0.005 or 0.001

> Are the results in report of this kind made more meaningful by specific
> survey techniques (specific sampling, controlled demographics etc...) ?
> or should we just ignore them ?
> 

Sure, but what is your purpose? Are you going to use the interpretation
to tell parents or school children that if you score very high grades,
you're not going to do as well as those who didn't? :)


Rgds,

Molly
-- 
Dr Molly Cheah
Primary Care Doctors' Organisation Malaysia (PCDOM)
eMail: drcheah@pc.jaring.my
Web-site: http://pcdom.org.my
DAGS Project: http://pcdom.org.my/dags/
DAGS Pilot: http://pilot.pcdom.org.my

------------------------------------------------------------
To unsubscribe: send mail to ossig-request@mncc.com.my
with "unsubscribe ossig" in the body of the message