Re: Representative Samples and the 2004 Vote

[David Campbell]

I recently heard a presentation on TV from a representative of Gallup or 
another major poling org.  He said that after years of doing this, they 
are well aware of the bias to be expected from their methods.  So they 
have developed correction formulas to improve on the predictive accuracy 
of results (a parallel to our statistical corrections--for restriction 
of range, etc.).  But the news media typically fails to give such 
details.  If all that is reported is the raw data, we only know that 
there is a systematic bias there.  I would prefer to get the additional 
info on what the corrected values are and how they were obtained.
 --Dave

Christopher D. Green wrote:
Dennis Goff wrote:
Christopher,
This one does have teaching relevance for me. I will talk about 
sampling at least briefly in my stats course within the next two 
weeks. I often use political polls as examples there. Have you seen 
anything about the Gallup methodology that is producing the 
non-representative sample? Given the consistency it seems unlikely to 
be the result of sampling error.


--
___________________________________________________________________
David E. Campbell, Ph.D.        [EMAIL PROTECTED]
Department of Psychology        Phone: 707-826-3721
Humboldt State University       FAX:   707-826-4993
Arcata, CA  95521-8299          www.humboldt.edu/~campbell/psyc.htm
---
You are currently subscribed to tips as: [EMAIL PROTECTED]
To unsubscribe send a blank email to [EMAIL PROTECTED]