At 05:06 PM 12/3/2002, E. Jacquelin Dietz wrote:
I think that Gallup knows all the residential phone exchanges in the country. (As the article explains, a "phone exchange is a six-digit number: the area code and the first three additional numbers of the phone number.") They randomly sample exchanges, with probability proportional to the number of residential phone numbers in that exchange (which they evidently have some way to estimate). Then I think they just add a 4-digit random number to each exchange to generate a phone number. I don't think they know (or care) which numbers are listed or unlisted.though they do make a very special point of this ... being able to select UNlisted numbers too ... as part of their telephone sampling plan in the original link that jill sent
Jackie Dietz
i think they DO care and, if they knew that they had no way of making sure they include some, according to the approximate % in the population (ie, those with unlisted #s) ... they would worry a bit that their sample was not "representative"
again i query ... does anyone know if gallup has NO way of guaranteeing that SOME of the sample that they take ... will be from the UNlisted telephone # subgroup?
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
. http://jse.stat.ncsu.edu/ .
=================================================================
