Please register to participate in our discussions with 2 million other members - it's free and quick! Some forums can only be seen by registered members. After you create your account, you'll be able to customize options and access all our 15,000 new posts/day with fewer ads.
I tend to give more preference to 'focused' surveys, those targeting specific groups, such as Hispanic voters, black voters, Jewish voters, Christian voters, low educated voters, higher educated voters, women only, men only, etc.
Very good point. It's not just race/ethnic/religion blocs though ... it also depends on those single issue voters: guns, abortion, the environment, animal rights-PETA, illegal immigration, etc.
Hip hop Obama inspired the young voters. Obama related to the pimp-with-a-limp crowd, too. But, Hillary will do well to get as many votes as white man Kerry did in 2004.
Gosh, I can't think of anyone I know, rich or poor, old or very young, who does not have a cell phone. Even if it is not total, it seems to me that sufficient numbers of people in all demographics do have them such that cell only surveys would be representative of the population. OTOH, I don't answer my cell phone either if I don't recognize the number. My "kids" - young adults - answer every call though.
There are homeless people living outside near my work, several of them. I drive by them everyday and see them talking on their cell phones.
We must recall that polling cellular phone users is much more costly. By Federal law, pollsters must manually dial a cell phone number. That means a person, being paid a wage, must sit and dial the numbers. Conversely, auto-dialing may still be used for landlines (if a person answers, then a person standing by will pick up the phone to conduct the survey)*.
*I, having a landline, sometimes get these robocalls. I pick up, say a cheery 'hello', only to be met with nothingness, save for some clicking sounds as the robo-calling machine is frantically trying to get a live person to pick up. Normally, I hang up before they can pick up. When I wait, it is usually for some insurance scam or such (for once you hit 60, such calls rise quickly). I did once, two months ago, wait and got an honest to goodness pollster, and so spent the 10 minutes or so answering the questions.
If you don't believe me, I'll be happy to take a picture for you next time I drive by there.
I was not being sarcastic to you. I was actually agreeing with you. Basically everyone has a cell phone these days.. that's why I said even my dog may have one!
When is Hillary getting off her butt and attempting to stop the bleeding? I hope they aren't thinking she can stay in hiding and nurse a 3-4 point lead until Election Day.
The article speaks of how, in 1936, Literary Digest conducted its own poll, by mailing out surveys to 10 million of its subscribers, asking whom they preferred: Alf Landon or FDR.
2.4 million responded. Based on these results, the Digest confidently predicted that Mr. Landon would win the election with 57 percent of the vote, versus 43 for FDR. They thought, with so many responding, how could they be wrong?
Of course, they were wrong. FDR won with 62 percent of the popular vote.
Why? Because during the Great Depression only the more well-to-do had subscriptions to magazines like Literary Digest. It was skewed. The article points out that a young man by the name of Gallop was much more successful in polling a much, much smaller sample, representative of the population as a whole.
As you have pointed out, the problem with the Literary Digest's was that the sample wasn't representative of the target population (general election voters), that is why current pollsters collect demographic information so that they can see how the people that responded compare to the population that they are interested in. If the numbers of the respondents are not in line with that population then weighting is done to bring it in line.
Quote:
Originally Posted by legalsea
It does serve as a warning about many of the polls out today. Recall, the Gallop organization is not doing 'presidential polling' (although they are doing other polling), since they have not, in the words of Donald John Trump, 'figured it out', in how to balance landlines and cell phones, etc.
Some of the polls still conduct only landline surveys. Of course, that means that they do not call those whom only have a cell phone, and no landline at all. Even getting people on the landline has become more difficult, with many having caller ID, and not accepting calls from unknown numbers. Many people have told me (joke) that the problem with calling only landlines is that you tend to get 1) older people; and 2) poorer people, or people on fixed incomes (i.e., they do not want the added expense of a cell phone). Many of the cell-phone-only people are younger.
I tend to give more preference to 'focused' surveys, those targeting specific groups, such as Hispanic voters, black voters, Jewish voters, Christian voters, low educated voters, higher educated voters, women only, men only, etc.
The reason that Gallup isn't doing horse race polling is that it is simply too expensive for them especially since there are other companies that do the same thing and do it well. I can say that they do it well, because we have something to compare their predictions to, actual election results. As long as these polls continue to correctly predict the outcomes there is no reason to dismiss them. Trump fans never seem to acknowledge that Nate Silver's models accurately predicted the 2008 and 2012 elections. Don't bother trying deflecting with Silver being wrong with his personal opinion about Trump's chances to win the nomination but if you insist on wanting to bring that up, it proves that an individual person can be wrong but the numbers are almost always right.
Also it is unlikely that the polls that you give preference to are solely surveying specific demographics, the time and cost of doing such research is just too expensive or if such research is done it usually done from targeted lists and the results have little to no generalizability. The results that you see are probably a subset of all of the respondent's of a larger poll. It is likely to be done from the larger sample polls where you have better chance of having larger subgroups so that estimates you get don't have huge margins of error.
Quote:
Originally Posted by legalsea
My point is this: many of these targeted polls consistently show Mr. Trump doing poorly, save for the lower educated voter. It is one reason I rather doubt that Hillary Rodham Clinton has such a slim, national lead, over Mr. Trump. To my mind, one cannot be losing by such significant numbers in so many target groups, yet 'make up' for such numbers on the national polls. I suppose it is possible, but I think it doubtful.
Here is a simple example of how it happens.
There is a state with 1000 people 520 are women and 480 are men
Based on the most recent poll 80% of women and 20% of men support candidate A
Candidate A has an overall support of 51.2% of the state and mere 2.4% lead over Candidate B. However even through Candidate B has huge support from men and some support from women and it appears close overall, Candidate B will lose.
When is Hillary getting off her butt and attempting to stop the bleeding? I hope they aren't thinking she can stay in hiding and nurse a 3-4 point lead until Election Day.
She's in a comfortable position right now (up by seven in Suffolk today). Traditional hard core campaigning begins after Labor Day. You will be seeing more of Clinton than I'm sure you care to. Biden and Obama are the two most popular politicians in the US right now by a large margin. Biden is working the union vote, reminding them that you dance with the one that brung ya. Look for some massive rallies with Obama coming up soon.
Please register to post and access all features of our very popular forum. It is free and quick. Over $68,000 in prizes has already been given out to active posters on our forum. Additional giveaways are planned.
Detailed information about all U.S. cities, counties, and zip codes on our site: City-data.com.