On 23 Oct 2003 07:34:44 -0700, [EMAIL PROTECTED] (Renee)
wrote:

> Hello.  I am trying to come up with a sampling strategy for the
> license renewal of substance abuse treatment facilities in Florida. 
> Part of the criteria for a provider's license renewal is to review a
> sample of their facilities' discharged records to see how many items
> out of a total all the records combined have, so I am dealing with a
> proportion.  

But the people you are dealing with are unlikely to have a
statistician on hand.  What they will have -- if this follows the
other examples that I have heard of -- is a state regulation
that spells out, in detail, what is required and what is allowed.

>            I have a software program that deals with sampling
> proportions already.  However, the methodology is further complicated
> by the fact that one facilities' records will be divided among a few
> or more counselors, which to me violates the assumption of
> independence of observations.  For instance, one facility may have 48
> records and 3 counselors, so there may be 16 records per counselor in
> this case or possibly an unequal amount of records per counselor also.
[ snip, rest]

If the usual assumption, or a common one, is that there
is only *one*  counselor  at a facility, then the usual Facility
*does*  meet that criterion (or the several criteria together, 
if there are several)  for each counselor.  I suggest that you 
should be safe, if you meet what they ask -- per   facility? --
for each counselor.

If they don't have details written out, for some lesser
standard for you to meet, with your three counselors --
then there is nothing  else that you can do that *should*  
be convincing, that I can think of.

-- 
Rich Ulrich, [EMAIL PROTECTED]
http://www.pitt.edu/~wpilib/index.html
"Taxes are the price we pay for civilization." 
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
.                  http://jse.stat.ncsu.edu/                    .
=================================================================

Reply via email to