In a message dated 14/11/96 01:52:11, you write:
>Technically to be correct and have a statistcal representation it should be
>35% of the population,
Surely the number needing to be surveyed depends on the sample size and the
size of the sub group you are seeking. The smaller the % of percentage of
people connected, the larger the percentage surveyed has to be to get an
accurate result. As the prevelance has no aproximate value, the survey may
need to be as large as a third or could be as small as 15%.
We are hoping to find the prevelance of internet connection among all GP
practices and all GP's (not the same thing as clustering occurs within
practices and larger practices are more likely to have a connection). There
are overt 30,000 GP's in about 10,000 practices. Any friendly statisticians
out there?
cc GP-UK - sorry for the cross posting
Trefor Roscoe Email [log in to unmask]
Beighton Health Centre Tel 0114 - 269 5061
Queens Road Fax 0114 269 7186
Sheffield S19 6BJ
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|