I need to do a sample size calculation to determine the number of audits
required.
The results of an audit will be:
1) OK
2) Minor variances
3) Multiple variances
4) Severe variances
We want to know the sample size required to determine the proportions in
each category within say 5% with a confidence of say 80% (specific
values to be determined).
I have used the test of proportions before but that has been in
reference to only two possible responses (pass fail). I searched the
R-archive as well as my stats books and other stats web sites etc for a
procedure that allows multiple dependent proportions (as described
above). I am unsure if such a procedure even exists, if not procedure
exists so does the standard calculations for the test of proportions
apply or is an adjustment recommended (it seems to me the standard tests
would give you a conservative sample size if you framed the question
towards the least common out come, for example use the standard formulas
to estimate sample size based on proportion of severe variances vs
everything else). Based on my experience, if something exists anywhere
it is in R, but I have my doubts. Information re R-packages, code,
references etc gladly accepted and I apologize in advance for asking a
question that is not really R specific.
Michael J. Bock, PhD.
ARCADIS
24 Preble St. Suite 100
Portland, ME 04101
207.828.0046
fax 207.828.0062
[[alternative HTML version deleted]]
