Welcome, Guest. Please login or register.
Did you miss your activation email?
January 21, 2020, 12:18:49 am
News: 2020 Presidential Predictions (Primary) are now active.

  Atlas Forum
  Presidential Elections - Analysis and Discussion
  Presidential Election Process
  Polling (Moderator: muon2)
  2008 jewish exit polls are false (Jews aren't that libreal) (search mode)
Pages: [1] Print
Author Topic: 2008 jewish exit polls are false (Jews aren't that libreal)  (Read 16970 times)
Alcon
YaBB God
*****
Posts: 30,909
United States



« on: July 16, 2012, 02:03:03 am »

could you add up the total turnout in the orthodox Jewish EDs of New York so we can calculate what approximate percentage of NY Jewish voters live in these areas (using the exit poll for the statewide total)?

It's reasonable to argue that Orthodox Jews are underrepresented in exit polls, but unless they refuse/are missed by phone pollsters at unusual rates, you're going to struggle to dismiss the phone polls on the subject -- which are the superior evidence here.
Logged
Alcon
YaBB God
*****
Posts: 30,909
United States



« Reply #1 on: July 18, 2012, 08:11:55 pm »

Could you at least try for an estimate?  For instance, you could figure out the overall turnout in an area that's ~100% Orthodox Jewish, versus the 18+ population, and then extrapolate if there's an estimate for the number of Orthodox Jews in New York.

Jews were included in the NY exit poll, also being only 3% of the sample, the results were suppressed in the public release for MoE reasons.  The data are probably available somewhere.  I'm just trying to find some empirical evidence here, however rough.

I don't know why you didn't respond to the phone poll part of my post.
Logged
Alcon
YaBB God
*****
Posts: 30,909
United States



« Reply #2 on: July 27, 2012, 02:18:56 am »

You need to find the population for those areas too so we can estimate what percent of orthodox turn out.  Then, we can take a statewide projection and start empirically testing your claim here -- or doing a better job than we have so far this thread.  Get me?
Logged
Alcon
YaBB God
*****
Posts: 30,909
United States



« Reply #3 on: July 27, 2012, 09:59:49 pm »

You need to find the population for those areas too so we can estimate what percent of orthodox turn out.  Then, we can take a statewide projection and start empirically testing your claim here -- or doing a better job than we have so far this thread.  Get me?
the problem is that the many Orthodox communities are unique from one another and the communities could have drastically different turnout%.

what I'm trying to do now is estimate the actual numbers of the Orthodox vote in different areas

ok, then compile this for different orthodox precincts.  The Census has precinct population information, including 18+ statistics.

Right now you're dangerously close to cherry-picking.
Logged
Alcon
YaBB God
*****
Posts: 30,909
United States



« Reply #4 on: July 29, 2012, 03:05:24 pm »

You need to find the population for those areas too so we can estimate what percent of orthodox turn out.  Then, we can take a statewide projection and start empirically testing your claim here -- or doing a better job than we have so far this thread.  Get me?
the problem is that the many Orthodox communities are unique from one another and the communities could have drastically different turnout%.

what I'm trying to do now is estimate the actual numbers of the Orthodox vote in different areas

ok, then compile this for different orthodox precincts.  The Census has precinct population information, including 18+ statistics.

Right now you're dangerously close to cherry-picking.
understand the fact that the turnout rates between different Orthodox communities drastically differ from one another and need to be worked out differently.

I repeat there is a drastic difference between different Orthodox communities in many ways and each needs to be calculated individually. 

for example here are the breakdowns you want for the most McCain ED in different Jewish neighborhoods in NYC and you'll see why this can't be averaged out.

McCains best ed in Far Rockway 524 votes for McCain, 43 for Obama
18+ breakdown
758 whites
13 blacks
9 hispanics
2 asians
1 other
McCains best ed in Borough Park 322 votes for McCain, 8 for Obama
673 whites
8 hispanics
McCains best ed in Flatbush (Midwood-Gravesend) 364 votes for McCain, 37 for Obama
615 whites
4 hispanics
5 asians
McCains best ed in Williamsburg 211 votes for McCain, 22 for Obama, 1 for other
904 whites
2 blacks
36 hispanics
4 others

Yes, and that's why averages exist.  Al is right -- this is not proper electoral analysis.  Also, the Census gives 18+ populations by district too.  You can keep throwing numbers at us, but until you make an effort to contextualize it in a way that can actually be used to test a falsifiable hypothesis, this is all pretty silly.
Logged
Alcon
YaBB God
*****
Posts: 30,909
United States



« Reply #5 on: July 30, 2012, 02:32:28 am »

those were above 18 populations

Excellent, thanks.

this is how you do proper electoral analysis (and why my numbers are accurate and polls are almost always wrong when done in the Jewish community)  an intelligent person does electoral analysis by actual numbers not AVGS.

"Actual numbers, not averages"?  What do you think an average is, a color?

You're complaining that variance is high when it comes to turnout.  I'm saying that isn't a terminal problem to the analysis.

did any one of you nay sayers (who I might add were too foolish to take me seriously when I correctly said Turner was going to win because of marriage) actually look up the election results in the areas I said and check to see if you agreed at all with my estimations of the Jewish vote.

The Orthodox Jewish vote in 2008 was heavily Republican, as it is in most elections.  This is not new information to most of us.

avgs (even when 100% accurate which there is no way these numbers were) don't tell the whole story my numbers if accurate do.

No, they don't.  You're just showing statistics about a bunch of heavily Orthodox precincts, and claiming Jews are more Republican than exit polls and phone polls suggest.  You're skipping a few steps.

maybe I shouldn't have put this information on this website because obviously people here are not smart enough to know that 20 +1 + 1 + 1 + 2 practically does not mean the same thing as a avg of 5.

Is there a particular reason you're insulting my intelligence?  Also, yes, those numbers have an average of 5.  The fact that average may not be useful, for some reason, does not mean it doesn't exist.  Please tell me you're not one of those ridiculous people who says things like, "There's no average -- everything is different."  If we can reasonably approximate the average Orthodox turnout for New York, we can use that to test the hypothesis you're asserting.  There's no reason why the average shouldn't be used for that purpose, is there?

PS my electoral analysis led to won of the most shocking defeats the democratic party ever had in NY.  So I wouldn't bash it to much.

Did it?  I'm going to assume you mis-wrote this paragraph.  If your point was that you correctly predicted the unexpected outcome of an election, congratulations, that indicates you're at least as good as that chicken that chose the Super Bowl champion by pecking a photo of the winning quarterback.  Now, let's try to beat that chicken on soundness of methodology!
Logged
Alcon
YaBB God
*****
Posts: 30,909
United States



« Reply #6 on: July 30, 2012, 06:07:13 am »

actually it's all over the place and it depends drastically on the candidate and where their running.

yes, I know.  i was talking about the 2008 presidential race.

In the 2008 pres election it mathematically impossible for the exit polls in the NE to be accurate with so many McCain voting areas if there was such a low sample size.

Please, show us your math for this.  I was asking you to do this the entire time.  (See below if you need an elaboration.)

can you honestly say that Jewish vote (including Orthodox and Russians) in the NE were not much more for McCain then in it was in the West?

No, I imagine it was more Republican.  Can you stop ascribing me arguments I'm not making and start actually running the analysis required to soundly prove or reject your hypothesis?

that's not at all what I was doing I was trying to calculate the vote of all Orthodox jews in a certain neighborhood.
what I was doing by cherry picking EDs was just to show you how different each Jewish community was in turnout.  the other EDs in the same neighborhoods have similar %s of Orthodox Jews showing up

...how is that different from what I said you were doing?

in response to the first question because I was insulted first

Where did I insult your intelligence?  Moreover, unless you actually think I'm unintelligent or arguing unintelligently, why bother?

Also, why did you ignore the rest of that paragraph, which was my substantive criticism of your analysis?

everything I said (assuming you actually understood it which based on my "unique" writing style is very possible you didn't) would help you figure out what I'm saying.

Trust me, I'm doing my best to figure out what you're saying.  Communication problems happen, "unique" writing styles or not.

depends where and when understand the Orthodox Jewish community is really many different Jewish communities that are in many stats at opposite extremes.  The divergence in behaviors between the avg secular Jew in Connecticut and the avg secular Jew in Suffolk County is next to nothing compared to the divergence in behaviors between the avg Orthodox Jew in Far Rockway and the avg Orthodox Jew in Williamsburg.   I do believe that avgs are important informational tool but only if you know what your doing.  averaging out Orthodox trends with out knowing context is 100% useless. 

You're arguing that the means of finding the average might be wrong, not that averaging is inappropriate.

I did a lot more then just correctly predict the outcome in that race on these forum (not that I'm going to say what I did on the internet to a bunch of liberals)
besides I was referring to a different campaign that I worked on.

Though I do regret saying anything because I appreciate my anonymity.

Shrug, man.  I have no doubt that you know more about Orthodox Judaism in NYC than me.  I've spent a few days in Manhattan, and even my more devout Jewish friends will break kosher if they can use someone else's silverware.  That doesn't excuse you from using sloppy methodology in your analysis, though.  Mathematically: Scientifically-processed educated intuition > educated intuition > my intuition.  But just because my uneducated intuition is inferior to yours doesn't excuse sloppy methodology.

Here's what you need to do.  Find a decent representation of Orthodox communities, like a regression model that extrapolates to a theoretical 100% Orthodox community using as many data points (representative, varied Orthodox communities as you can.)  Use that to estimate what percent of Orthodox turn out, and how their votes break down.  You can then use statewide Jewish results from exit polls.  With those sets of information, you can estimate what percentage of non-Orthodox Jews would have to be Democratic voters for the exit polls to be accurate.  This is tricky and may involve small data sets, but dude, it is still better than throwing out numbers and saying "this seems wrong" like you're doing right now.

You seem to have a lot of nervous energy about this issue...dedicate that toward testing this in a sound way and you may prove yourself probabilistically right.  But short of that, this topic is kinda silly season.
Logged
Alcon
YaBB God
*****
Posts: 30,909
United States



« Reply #7 on: July 30, 2012, 08:10:41 pm »

but that was the whole proof that the numbers were false
I'll respond to the rest of your post later
then you didn't understand my point from the beginning
look at this
http://www.jewishdatabank.org/Archive/N-Jewish_American_Voting_Solomon_Project_2012_Main_Report.pdf
here is my basic argument
1. the sample size was only 952 people for the whole country (page 4) (so if the Orthodox vote was accurate there wasn't enough of a non Orthodox Jewish vote in the NE to counteract the NY strong McCain vote)
2. page 14 shows that the West voted more for McCain then the NE.

Your basic argument is that you think the Orthodox vote must mean that Northeastern Jews are more conservative overall, and because polls find otherwise, the polls must be undercounting the Orthodox.  That's a reasonable hypothesis.  The problem is that you're not testing it in a scientifically sound way -- or really testing it all, so much as searching for evidence to support your hypothesis.

what I was trying to do was to approximate the actual number of the Orthodox vote in the NE and I was just working community by community. 

That's what I'm asking you to do.  You have several steps to go before you have statistically useful information.

(the Jewish vote in NYS was to small a sample size for the exit polls (which also implies that it underestimated the NYS Jewish McCain vote))

What, why?  I can see how it would be potentially consistent with that, but how does it imply that?
Logged
Alcon
YaBB God
*****
Posts: 30,909
United States



« Reply #8 on: July 31, 2012, 01:36:36 am »

what I'm doing is going through different Orthodox (and soon also Russian communities) to estimate the actual vote (though I'm also trying to use low estimates for the most part) of conservative Jewish demographics in the NE based on the actual results so we can then figure out (assuming my number is correct) how much the liberal Jewish demographics would have to vote to have the NE numbers meet the actual number in the exit polls for the NE.

Yes...and unless every orthodox Jew fits neatly into an election district, the best way of doing that is to approximate the average Orthodox Jewish community turnout, and then use the estimate of Orthodox Jews 18+ in New York state times that number to estimate the Orthodox Jewish vote.  Your current method is going to miss a lot, isn't it?

according to this site (this site is easier to read then the one it quoted from)

(this underestimates NYC, this is because it uses 2002 Federation numbers for NYC and not the recently released numbers)
http://www.jewishvirtuallibrary.org/jsource/US-Israel/usjewpop.html

total Jewish population in the country 6,588,065 in 2011 (these numbers can also be nitpicked but lets assume their accurate)
this is not of sample size
total Jewish population in NYS 1,635,020
these are of sample size
NE 3,157,670
West 1,613,225
South 1,107,140
MW 710,030

(considering over half of all Orthodox Jews in America are living in NYS I think my point is very clear)

...no
Logged
Alcon
YaBB God
*****
Posts: 30,909
United States



« Reply #9 on: August 09, 2012, 05:54:04 pm »

NYS 1,635,020 (to small a sample size to report on)

NE 3,157,670 (good sample size)
West 1,613,225 (good sample size)
South 1,107,140 (good sample size)
MW 710,030 (good sample size)

wouldn't there be a very good chance you'll underestimate the NYS vote

What?  That's not how random sampling works
Logged
Alcon
YaBB God
*****
Posts: 30,909
United States



« Reply #10 on: August 10, 2012, 06:08:22 am »

What?  That's not how random sampling works

Random sampling of what? This polls wasn't designed to give a good estimate of the Jewish vote - it's prime objective is to figure out who won. A "Jewish" poll would have to sample New York a lot heavier than Texas. A national poll would need to oversample NY to give reliable data on NY Jewry - and the fact that a separate number on NY is not reported suggests this is not done (sensibly enough). NY ultra-Orthodox are pretty negligible as far as the national exit poll is concerned, but much more important to figure out the Jewish vote.  Given how concentrated the ultra-Orthodox vote is, it is unlikely to have been sampled much. Perhaps some statistical adjustment has been made to take account of this: but it is not obvious. And, in any case, it is very unlikely that substantial enough numbers from those insular communities have been sampled in the first place to make such adjustment doable. Not at all a problem for the national exit poll. But a big problem if you are after figuring out how "the Jews" voted.

No, his point is totally valid with exit polls, like you point out.  My exchange with him has been going on for several pages and has moved onto the standard fare about how exit polls aren't representative.  Mainly, I'm still unclear what sort of empirical use he's going for when he adds up these Orthodox communities...as far as I can tell, his point is, "look how many Orthodox Jews, and they're very Republican, so Northeastern Jews must be much more Republican than believed!"  Actually, I'm still not quite sure of his central claim here...sometimes he seems to be saying that national polls are useless for telling how Orthodox Jews vote (duh?); sometimes he seems to be complaining about the exit polls (fair); sometimes he seems to be suggesting Northeastern Jews are much more conservative than reported (his methodology is inadequate for that); and often it's unclear.

Random sampling of what? This polls wasn't designed to give a good estimate of the Jewish vote - it's prime objective is to figure out who won. A "Jewish" poll would have to sample New York a lot heavier than Texas. A national poll would need to oversample NY to give reliable data on NY Jewry - and the fact that a separate number on NY is not reported suggests this is not done (sensibly enough). NY ultra-Orthodox are pretty negligible as far as the national exit poll is concerned, but much more important to figure out the Jewish vote.  Given how concentrated the ultra-Orthodox vote is, it is unlikely to have been sampled much. Perhaps some statistical adjustment has been made to take account of this: but it is not obvious. And, in any case, it is very unlikely that substantial enough numbers from those insular communities have been sampled in the first place to make such adjustment doable. Not at all a problem for the national exit poll. But a big problem if you are after figuring out how "the Jews" voted.

?  He listed a bunch of populations, and then claims New York's vote share will be "underrepresented" in a random poll because New York is a smaller population than "the Northeast."  At least, I think that's what he meant; again, it's unclear.  Even if he was talking about non-random sampling (e.g., an exit poll), I still have no idea what he's saying.  You're right about the flaws of the exit poll, but what does that have to do with New York being a relatively small population/subsample?

You seem to find his argument a lot more cogent than I do.  Maybe I'm just being thick.
Logged
Alcon
YaBB God
*****
Posts: 30,909
United States



« Reply #11 on: August 10, 2012, 06:05:03 pm »

Yes, the argument is reasonable. Think of it that way. We do know that ultra-Orthodox live in some compact isolated communities in NY state. These ultra-Orthodox form a substantial part of the Jewish population not merely of the state, but of the entire US. Any poll that would have been designed to find out how the "average Jew" voted would take care of actually sampling from these communities.

Yes, and I've never disputed that.

To the best of our knowledge, no such poll has ever been conducted. Rather, national exit polls were used. As far as the behavior of an "average American" these isolated communities are not particularly important and could well have been missed. Neither, it seems, there has been an attempt to match the weights so that the ultra-Orthodox are properly represented in the sample (in fact, the data on the number of the ultra-Orthodox is not reported and, likely, wasn't even collected). Ultra-Orthodox are entirely inconsequential for the purposes of the national exit poll. But they ARE important for the answer to the question at hand. This MAY be a problem - finding out whether it is, would require looking at the data closely.

That's not true.  The Gallup Poll, among others, aggregates Jewish data -- that's why I earlier mentioned that it matters whether Orthodox Jews were disproportionately likely to be underrepresented in phone polls.  The problems of exit polls are widely known and, again, I haven't disputed them here.

This is, I  believe, a well-known statistical problem of dealing w/ very small subsamples. It rears its head, for instance, whenever they try to estimate things about the ultra-rich from the general census surveys - there are just not enough ultra-rich in those samples, and the ones there are may very well be not very representative.

But that argues that New York's Margin of Error is large, not that New York is somehow "underestimated" -- there is no reason to believe that it is any more likely to be underestimated because it is a small subpopulation; any given subpopulation, regardless of its size, is equally likely to be underestimated/overestimated, regardless of the size of the "supra-population."  You're making a fair point that doesn't match the language he used.

Also, none of this really explains why we're (seemingly arbitrarily) adding up the populations of select Orthodox neighborhoods, and the (valid) arguments you're making don't seem possibly related to that project.
Logged
Alcon
YaBB God
*****
Posts: 30,909
United States



« Reply #12 on: August 10, 2012, 06:13:43 pm »
« Edited: August 10, 2012, 06:17:20 pm by Alcon »

no what I'm saying is that since Orthodox Jews are tremendously underrepresented in the polls.  The polls are useless to giving an accurate figure of the total Jewish vote.

that's the main point

OK, so the point was to add up enough Orthodox Jews in New York to show that, if missed (as we can expect them to be by exit polls), they'd affect the numbers in exit polls?  That's a pretty heavy-handed way of doing that, but I agree you're probably right and they're missed.  I don't know why we're looking at exit polls when we have superior information from national phone polls.

I'm saying that Northeastern Jews are much more conservative than Jews out side of the North East because of the numbers of non stereotypical Jews.

Ah, yes, the Jerry Seinfeld Effect.

Assuming the numbers to the poll are accurate only for the stereotypical Jews.  the statement I quoted from you is accurate.  and even if the polls are inaccurate for stereotypical Jews the statement will still be accurate if the numbers for the avg stereotypical Jews are similar in all 4 regions.

This has never been a claim I've criticized.  I'm criticizing your methodology, and also suggesting you look at phone polling instead, which is generally much better than exit polls -- in this case especially, but also in pretty much all other cases.

It's also been consistently unclear throughout this thread what you're doing, why, and to what end...ag and I, for instance, came to completely different inferences of a statement that you were previously asked to clarify several times.  Sorry; I might be being impatient, but yeah.
Logged
Alcon
YaBB God
*****
Posts: 30,909
United States



« Reply #13 on: August 10, 2012, 07:02:32 pm »

I don't know what Gallup does about Jewish polling. Have they actually polled Jews specifically?

No, but their daily poll adds up to a statistically significant number of respondents from even small sub-populations over time, and they periodically report it.

Anyway, my point wasn't that NY Jew was right - it was that he was raising a valid issue. His choice of language was very unfortunate: so unfortunate, in fact, that I only started to read his posts in this thread to be able to claim he is talking nonsense. He isn't Smiley))

Fair enough!  Although I'm not convinced that his posts are nonsense-free, even if there's a valid thesis or two in them somewhere...
Logged
Alcon
YaBB God
*****
Posts: 30,909
United States



« Reply #14 on: August 11, 2012, 09:40:33 pm »

when it comes to the Jewish vote in various I think I remember a few times that Gallop had some extreme outliers.   this is more true in polls of NY state.  I've seen a few phone polls for the Jewish vote in NY state that had completely different results a few weeks a part (the media of course took both set of polls as accurate)

for example when I saw different opinion polls of the Jewish opinion on gay marriage in NY state Iv'e seen different  polls with more then a 40 point spread.  factually on the ground Orthodox Jews could be as much as 1/3 of the NY state Jewish electorate (the few times I ever saw polls of "Orthodox" jews it was never more then 10% who supported it and even in the most liberal orthodox Jewish communities it never approaches 20% the only "question" is should we become 1 issue voters)  (just to make clear my point is just to show how polling numbers are all over the place)

And why is that, sample size?

PS I doubt almost anyone in the most insular (this doesn't mean most religious) Orthodox communities will ever talk to pollsters (these are the type of Jews who would vote in American elections but not Israeli ones).  In other non modern Orthodox communities also I think the response rate would be less then in the avg public by a significant margin.

I also doubt Russian Jewish voters are polled to much do to the language problem.

Both of these are reasonable hypotheses, but I'm pushing you to estimate what we can expect the impact to be, using some sort of empirical measure.  Right now you seem to be throwing out hypotheses and numbers and not doing much with either -- just kind of hoping something sticks that is compelling enough to prove your thesis.
Logged
Alcon
YaBB God
*****
Posts: 30,909
United States



« Reply #15 on: August 12, 2012, 09:24:48 pm »

Phone polls are not done by ZIP code.

You haven't "proven" anything.  You've presented one obviously compelling thesis (the exit poll issue) and then one potentially true thesis (the phone poll issue) which you've "proven" in a way without any clear, consistent, prearranged methodology.  You may well be right

The MoE on those polls -- even if they got a perfect sample -- is nearly +/-10%.  Why would you think it's not sample size?

Again, your thesis is reasonable, but you seem like you're actively setting out to prove your thesis right...which is a problem.
Logged
Alcon
YaBB God
*****
Posts: 30,909
United States



« Reply #16 on: August 15, 2012, 02:08:09 am »

My point about the zip codes was for exit polls. (the last time I reffered to zip codes I meant block)
the other point was referring to phone polls (not sure which ones do this) that take random blocks of phone numbers (which as far as understand would be based on geography)

Right, but I already said exit polls are problematic.

I see no reason why choosing random phone numbers based on registered blocks would undersample Orthodox households...

One thing that might effect phone polls (this would only be a minor effect though) is that since it it done by each individual phone line it might overestimate single person households.

In general Orthodox Jews have some of the highest being married rates.

not sure if they adjust for this factor that certain demographics are more likely to share 1 phone line.

(not sure how cell phone will play into this now.)

Again, you're throwing out hypotheses to bolster your intuitions and hypotheses, and not creating methodologies to actually prove them.  I just don't know where you're going with this.  You can create hypotheses that run against your intuitions, too -- so this exercise doesn't seem to be leading to "proving" anything, if that's what you want to do.
Logged
Alcon
YaBB God
*****
Posts: 30,909
United States



« Reply #17 on: August 16, 2012, 07:25:47 am »

the question I have is
is random sampling 100% random?

No.

is there a system that overplays having many different area codes and first 3 digits in a phone number.

I'm not sure what you mean -- pollsters may skip dialing phones in ranges they know to not be in use.  Besides the valid household issue, I'm really not sure where you're going here, or what area codes have to do with it.

do we agree that polls work by household an single people (with out weights) will be over estimated in the polls.

if this is true then Orthodox Jews will be underestimated in the polls.

I don't know how/if pollsters account for this, but this topic is starting to be very frustrating for me for reasons I've stated several times.  It doesn't seem like you're interested in approaching this methodically.
Logged
Alcon
YaBB God
*****
Posts: 30,909
United States



« Reply #18 on: August 19, 2012, 03:35:05 am »
« Edited: August 19, 2012, 03:37:33 am by Alcon »

As far as I can tell I already proved that either the exit polls were wrong
either the NE was more for McCain or the rest of the country went more for Obama.

no, you found a highly intuitive hypothesis for why they might be wrong.  "Proving" it would require substantiating your hypothesis and proving that variables that might make the non-Orthodox NE Jewry substantially more liberal than it is in the West are untrue.

I'm really not trying to be a pedant...but you don't "prove" anything by throwing out hypotheses blindly and then throwing out some data blindly.  The fact that I think your hypothesis is almost certainly true doesn't mean your methodology hasn't been cray-cray.

the area codes comment basically means this
if only trying to track the Jewish vote certain area codes and first 3 digit codes have more Jews then others.  thus lets say I would need to take a very high sample size from 718, but if I was doing the whole country (including non jews) as a whole I would take much less.  To prevent this from happening my question is do randomized numbers include a cap on how many from a specific area?

The polls I'm referencing don't specifically seek out Jews.  They are national polls that, over time, happen to have enough respondents (thousands and thousands) that even proportionally small sub-samples (like Jews) gain hundreds of respondents, and can be reported.  Their sampling is normal for a randomized national poll.

Your area code point is still confusing me.  Are you saying that, if you were specifically looking for Jewish responses, you might limit your poll to Jew-heavy area codes?  None of the polls I'm doing do this.  They also wouldn't, because a representative sample of Jews must be equally likely to sample a Jew in New York as a Jew in Utah.  If the two aren't equally likely to be sampled, the poll would not be of a representative sample.  However, I still feel like I'm missing your point, because capping respondents by area codes would be a pretty nonsensical solution to this problem.


Prove what, your hypothesis?  And which one?

in regards to phone polls there are plenty of reasons that would guarantee a underestimate in the polls by Orthodox Jews unless there are some weights in place to counteract it.

Yes, but again, just stating this is not how science works -- at least not good science.

there is some proof that Orthodox Jews respond much less than their numbers to the pollsters
see page 25 (and they underestimated the % of Orthodox Jews in the country anyway)
http://www.jewishdatabank.org/Archive/N-Jewish_Values_Survey_2012_Frequencies.pdf.

There's a good start!  You can even assume that Orthodox Jewish response in other surveys is comparably lower (although an MoE would apply) and re-weight accordingly.  Something.  Just something with numbers.  That's all I'm saying -- you can do this objectively, and that's better than just throwing around hypotheses.

Also (I'd check but I have to sleep) not entirely clear on why the parameter sample is assumed to be the true Orthodox percent (although 8% does seem more realistic.)

Pretty impressive that, even in the weighted sample, the same-sex marriage approval numbers come out 81%-18%.
Logged
Alcon
YaBB God
*****
Posts: 30,909
United States



« Reply #19 on: August 20, 2012, 03:08:36 pm »

My basic point is that the only way to poll all jews accurately in a national poll would be to have a high sample from NY state.  Which would not happen in a National poll when trying to also poll the whole country accurately.  Thus the likelihood of underestimating NY's Jewish vote in these polls.

That's incorrect.  A national poll randomly samples people from all across the country.  It would take a while to get a statistically useful sample size of Jews, and that sample size would disproportionately be from New York, but that would not require oversampling New York.  That would, as you point out, make your sample of Jews non-representative.  It also makes no sense for a national pollster who isn't trying to find Jews, but rather finds a healthy subsample because they conduct so many interviews.  Got it?

There is no estimate in recent years that places the Orthodox % at 8%  so assuming there is any basis to this they may be trying to guess the turnout rate.

Right, which makes me wonder where the 8% comes from, if not a phone poll.

Also I'm almost 100% sure Orthodox Jews who are more to the right would answer the polls less.

Again, hypotheses like these are fine and seem intuitive enough, but there's no reason not to do legwork in trying to prove this in some way.  You're kind of making me work to test your hypothesis when you should be doing it.
Logged
Alcon
YaBB God
*****
Posts: 30,909
United States



« Reply #20 on: August 25, 2012, 10:15:18 pm »

exactly what sort of rigorous study did you do to arrive at that conclusion?
Logged
Pages: [1] Print 
Jump to:  


Login with username, password and session length
Logout

Terms of Service - DMCA Agent and Policy - Privacy Policy and Cookies

Powered by SMF 1.1.21 | SMF © 2015, Simple Machines

© Dave Leip's Atlas of U.S. Elections, LLC