A question for poll watchers
While reading post at Rantingprofs on The Military Vote ( a review of a column in the Washington Post by Peter D. Feaver) and, regarding a conclusion made in the article (registration required), this jumped out at me:
Now the question is whether the stream of state by state polling results reflect this subgroup as well as the other subgroups which are much more stationary to the area polled. No other subgroup is both as largely tilted in their preference AND more likely to absent from their residing state during polling (moreso with AD but only somewhat less so with NG, etc.)
Sure, this statistical and demographic situation will not be factor in Texas or, say, New York or California owing to the large difference in the percentages. But what about New Mexico, currently a Bush-Kerry tie, with a popular vote of 595k in 2000 and only 6,000 votes for a 1% swing. Total AD forces stationed in NM in 1999 was 11,642. (I could find no numbers for NG, et al.) And how about Nevada, NC, and Hawaii. (In Hawaii, active duty personnel stationing in 1999 is equivalent to 10% of the 2000 popular vote.)
I do not know how closely all active duty stationing relates to registered voter residency. I do not even know that polls, somehow, either take this into account or have shown it to be not significant. I am just wondering, mind you. Like I said before, I haven't seen any analysis to show that polls in states with, most importantly, disproportionate AD personnel stationed there reflect the large preferential difference in the military vote. Has anyone seen an article with more more supporting argument for such a conclusion than this one?
Update: Sorry, I used this for 2000 vote totals and this (pdf) via this site for Armed forces numbers.
"... Of course, the military vote is not large enough to decide the election except in the most extraordinary of circumstances, and then every other subgroup is decisive, too...."I wonder if this is an accurate conclusion. I really haven't seen any analysis to support this conclusion, so I ask those with much more demographic data at their fingertips and polling analysis experience, is this true? Here are some of the reasons I wonder about it. First, note the significant percentage difference (73% Bush to17% Kerry), and the large population of this subgroup in absolute terms (around 2,200,000 AD, RR & NG?). That is about 2% of the total vote in 2000.
Now the question is whether the stream of state by state polling results reflect this subgroup as well as the other subgroups which are much more stationary to the area polled. No other subgroup is both as largely tilted in their preference AND more likely to absent from their residing state during polling (moreso with AD but only somewhat less so with NG, etc.)
Sure, this statistical and demographic situation will not be factor in Texas or, say, New York or California owing to the large difference in the percentages. But what about New Mexico, currently a Bush-Kerry tie, with a popular vote of 595k in 2000 and only 6,000 votes for a 1% swing. Total AD forces stationed in NM in 1999 was 11,642. (I could find no numbers for NG, et al.) And how about Nevada, NC, and Hawaii. (In Hawaii, active duty personnel stationing in 1999 is equivalent to 10% of the 2000 popular vote.)
I do not know how closely all active duty stationing relates to registered voter residency. I do not even know that polls, somehow, either take this into account or have shown it to be not significant. I am just wondering, mind you. Like I said before, I haven't seen any analysis to show that polls in states with, most importantly, disproportionate AD personnel stationed there reflect the large preferential difference in the military vote. Has anyone seen an article with more more supporting argument for such a conclusion than this one?
Update: Sorry, I used this for 2000 vote totals and this (pdf) via this site for Armed forces numbers.
Posted by Dusty at 1:47 PM
0 Creaks:
Post a Comment