Sample Size Determination
Message posted by Dana Thomas on November 18, 2000 at 12:00 AM (ET)

Hello,

I need help with this problem.

It's five years after graduation and you are becoming well known in your field.
An association of which you are a member has asked for your help. They want to survey
members about the number of topics, but two questions are most important. The first is whether
or not the membership would vote to approve $50 per year increase in annual class. The second is to estimate the
number of weekends members have free each year for participation in continuing education sessions.

Here is some information you need to estimate sample size for a simple random sample:

N=20,000
A=accuracy=95%=3.84
B=booumd on your estimate=3% for the yes/no question and +-6 for the number of weekends

Using the formula n=Npq / (N-1) D + pq, where D=Bsqaured/A

and the formula n=N Sigma squared / (N-1) D + Sigma squared, where again D = Bsquared/A

Estimate how large a sample size you would need to collect via a simple random sample of members in
order to estimate approval of an increase in dues and number of free weekends with the desired accuracy and
sampling bound.

If you come up with more than one sample size, what do you recommend and why?

READERS RESPOND:
(In chronological order. Most recent at the bottom.)

Re: Sample Size Determination
Message posted by JG on November 19, 2000 at 12:00 AM (ET)

Re: Sample Size Determination
Message posted by DT on November 19, 2000 at 12:00 AM (ET)