> Dr Iain E. Buchan wrote:
> I keep a
> database of common statistical misconceptions and I want to construct
> teaching materials that address these effectively in the fullness of
> time. Also, I would be grateful for any examples of statistical
> errors from peer-reviewed, published literature.
>
>
Iain,
Try the following article if you don't already have it:
Liu, K., Moon, M., Silvetta, M., and Chawla, J.; "International infant
mortality rankings: A look at the numbers." ; Health Care Financing
Review. Summer 1992 13(4):105-118.
The article addresses some of the assumptions and fallacies in data
collection and manipulation that lead to the WHO IMR stats. I have used
this article in a couple health policy courses and it generally provokes
animated discussion, especially amongst those fond of quoting
international IMR stats as a measure of the adequacy of a country's
health system...
An example I particularly like is the variation in the definition of
'live birth' from country to country - a definition critical to
calculating the IMR. In the US, it is "one postpartum respiratory effort
or one postpartum heartbeat."
In France, a 'live birth' occurs when the event has been recorded at the
appropriate municipal registry - during regular office hours. French
office hours. An infant born Friday night that fails to survive till
Monday morning is a 'delayed stillbirth'. The WHO doesn't have such a
category so these are dumped in under 'stillbirth' instead of in the
numerator of the IMR calculation.
Phil
--
Philip G. Dunlap, D.O., M.P.H., Ph.D.
4 Bailey Hill Road
Natick, MA, 01760, USA
[log in to unmask]
(508) 650-9097 - voice
(508) 650-9152 - fax
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|