Date:         Mon, 14 Jan 2002 14:45:35 -0600
Reply-To:     Randy Richter <richterrr@SLU.EDU>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Randy Richter <richterrr@SLU.EDU>
Organization: Saint Louis University
Subject:      Stratified random sample question
Content-Type: text/plain; charset=us-ascii

Dear List

A colleague  and I are planning to survey physical therapists who belong
to our professional organization and practice in different settings.
The settings represent strata.  We are planning on drawing a simple
random sample from each of four strata (work settings).  Roughly equal
numbers will be drawn from each strata.  In the actual population, the
strata differ in size by about 5%.  Due to budget constraints and our
tolerance, we plan on margin of error of +/- 7%.

My question is:  Does the sampling method proposed require that I use
weighting when analyzing the data between strata?  For example, if the
final sample size for each strata is about equal and we use a a
chi-square analysis to examine responses between the strata should we
use a  a weighting procedure?  If so, how would the weights be
determined?  Any references, or books on sampling that provide a clear
explanation of the issues involved are appreciated.

Thanks in advance.

Randy Richter