Applied Statistics & hypertension

[huy.duong@ctsu.ox.ac.uk]

Dear anh Tuan,

>The real issue is the magnitude of bias, rather than bias per se.
>I guess you can compensate this problem by increasing the sample size. 

Increasing the sample size reduces the random errors. The bias in
your sample selection or measurement or analysis is not reduced
by increasing the sample size. In fact sometimes having too big
a sample can make you more open to bias:

Suppose your calculation says that you need 5000 people to detect with
some given certainties that a difference d does or does not exist.
It is better to devote your resources to recruiting ~5000 people (randomly)
and getting a high response rate from them than to try to recruit
50000 people and not be able to ensure a high response rate. If there
is a bias in the response rate then the result from 50000 people can be
less reliable. 

>	This is called "white coat" effect. Hypertension is 
>normally increased when it is measured with the presence of 
>a doctor. On the other hand, when the measurement is 
>performed by a nurse, then it is OK!

I don't know. I think my hypertension is increased if the nurse
is young and sexy. And I think anh Tuan Pham's hypertension is
increased in the nurse wears an ao dai and he looks at the slits
at the sides :-).

Huy