View Story | 116 comments
Comments: Expand Shrink Hide (Always) | Indented Flat (Always)
...but that's how most of these surveys are done, especially with Jews -- you take advantage of an already existing panel and contract with the market research organization that gathered them for their own purposes to screen for the desired population and ask them the relevant questions. Alternatively, sometimes if the N is large enough and the religion question is asked, you can piggyback off of large public use datasets -- but then you need those surveys to have asked about the variables that interest you.
It's done that way because if you have to use RDD (random digit dialing) to find Jews, you're going to spend millions of dollars creating a representative sample. In the market research panels, they already know who is Jewish, so you save all that money.
Seven score and five years ago... -7.62, -6.36
by wiscmass on Mon Dec 24, 2007 at 10:43:49 PM PDT
[ Parent ]
I wasn't thinking. I'm a heavy user of large public data sets, but they don't ask those types of questions. My data sets don't even have religion (race, Hispanic ethnicity, but no religion). Everything in them is free and it's of great use to me in my work, but it wouldn't do a thing for AJC.
If, in our efforts to win, we become as dishonest as our opponents on the right, we don't deserve to triumph.
by Tamar on Mon Dec 24, 2007 at 11:09:39 PM PDT
...but it's rare.
For instance, one of my friends tells me there is a research center that has been discussing buying questions about Jews on the GSS. I imagine that's very expensive, though.
by wiscmass on Mon Dec 24, 2007 at 11:56:46 PM PDT
I know CPS, SIPP, NLSY, YBS, NHIS, NHANES, SLAITS, ACS, and others (including the decennial census). But what is the GSS?
by Tamar on Mon Dec 24, 2007 at 11:59:24 PM PDT
by wiscmass on Tue Dec 25, 2007 at 12:53:45 AM PDT
and what are all those?
I'm a statistician, this sort of thing fascinates me
Now up: Daily Kos University
by plf515 on Tue Dec 25, 2007 at 04:31:54 AM PDT
for answers: CPS = Current Population Survey (census bureau) SIPP = Survey of Income and Program Participation (census bureau) NLSY = National Longitudinal Survey of Health (can't remember which agency conducts) YBS = Youth Behavior Survey (CDC, I think?) NHIS = National Health Interview Survey (National Center for Health Statistics) NHANES = National Health and Nutrition Examination Survey (also NCHS as above) SLAITS = State and Local Area Integrated Telephone Survey (also NCHS) ACS = American Community Survey (census bureau)
These are amazingly useful and rich surveys. All data is free to public. I work with the CPS and NHIS mostly but use things from all the others. And there are even more out there....
by Tamar on Tue Dec 25, 2007 at 04:23:23 PM PDT
do you know about my yahoo group stats_geeks_of_daily_kos?
by plf515 on Wed Dec 26, 2007 at 04:28:23 AM PDT
sounds like my kind of place, tho. I'll check it out -- guess just go on yahoo and look for the group?
by Tamar on Wed Dec 26, 2007 at 02:26:57 PM PDT
then click on groups on the left side, and from there, it's pretty straightforward
There hasn't been much action there....I gotta remember to use it to promote diaries, exchange data sets and so on
by plf515 on Thu Dec 27, 2007 at 04:43:34 AM PDT
I'm the senior health researcher for a children's advocacy organization.
by Tamar on Tue Dec 25, 2007 at 04:24:24 PM PDT
then, in addition to the list I just told you about (stats_geeks_of_daily_kos).... there's also the diary series A Little Bit Special, started by Frankenoid and me, which also has a Yahoo group
Children's advocacy is a vital issue
by plf515 on Wed Dec 26, 2007 at 04:31:59 AM PDT
Thanks!
by Tamar on Wed Dec 26, 2007 at 02:28:12 PM PDT
although no method is perfect.
One alternative is to sample heavily in areas where there are many Jews, and then attempt to adjust for that type of sampling. I've not seen this done for religious denominations, but it was done quite effectively by a guy who was trying to get a random sample of GLBT people
by plf515 on Tue Dec 25, 2007 at 04:30:44 AM PDT
...but that's still more expensive than using an already existing panel, which is why they do it that way.
by wiscmass on Tue Dec 25, 2007 at 09:12:19 AM PDT
wide narrow
View Story | 116 comments