I have 61.2 million records of people containing information on almost every
aspect of the citizens. I want to take a sample of records in order to check
the quality of data in them. If I want to be 95% confident that the selected
sample records will contain information that will be in error by not more than
5%, then how much records I should select in my sample?
In general, what should be the general rule for taking a reasonable sample
size if our aim is just to check the quality of data collected through
questionnaires or interview schedules?
I would be very thankful if anyone can help me in this regard.
Thanks
S.M.Asim
Statistician
Islamabad, Pakistan
____________________________________________________________________
Get free email and a permanent address at http://www.netaddress.com/?N=1
|