Could've, Should've, Would've

I was reading over many of the polls posted on The Roper Center's website, and found a few things quite surprising.

I looked mostly at the polls in "The Presidency" section and noticed that most, if not all, of the polls were conducted by telephone. This is interesting to note because, as Gawiser and Witt point out in "20 Questions Journalists Should Ask About Poll Results," some people either don't have phones, refuse to answer or weren't home at the time that the call was made. This can skew results in 2 main ways: First, this automatically eliminates from the poll anyone who can't afford a home telephone (or who chooses not to have one and just have a cell phone), and, as we've discussed in class, these people usually come from the lower income brackets and are most of the time Democratic voters (which will be important a little further down in this post).

I want to focus specifically on the following question and its data for a little bit of analysis:

National Security Survey [September, 2006]
(Now I'd like to read some proposals President (George W.) Bush and the Republicans in congress may offer to improve anti-terrorism efforts. Please tell me if you support or oppose each proposal.)...The following is a statement a Republican candidate could say in support of this proposal. Democrats think this is a law-enforcement exercise where you have to go to a judge to get approval. The Republicans understand that we are at war with terrorists and sometimes a court order takes too long. The President needs new powers to fight that war. Does this make you much more favorable, somewhat more favorable, somewhat less favorable, or much less favorable to a Republican candidate?

Much more favorable 27
Somewhat more favorable 24
Somewhat less favorable 14
Much less favorable 26
Don't know 9

Source: Survey by Third Way.
Methodology: Conducted by Benenson Strategy Group, September 13- September 14, 2006 and based on telephone interviews with a national likely voters sample of 600. Likely voters are registered voters who said they are absolutely certain/very likely to vote in the 2006 election.
Data provided by The Roper Center for Public Opinion Research, University of Connecticut.

There are a few points (and critiques) I would like to raise about this poll:
1) This poll was conducted by the Benenson Strategy Group, which boasts on its website about having "Presidential candidate Barack Obama, Governors, U.S. Senators, Congressmen, international labor unions, Fortune 100 companies and major non-profits" as clients. I find this interesting for many reasons, one of them being that this company has a wide variety of clientele ranging (presumably) across the entire political spectrum. This is also interesting because in this poll we see that they chose to do the polling by phone, and eliminate a very large, and arguably important, demographic of people who may not own phones (see above)

2) Although I think that the question was asked to the right group of people ("Likely voters are registered voters who said they are absolutely certain/very likely to vote in the 2006 election"), the size of the voting sample is relatively small. Although there is no "confidence level" listed, if we assume that it is 95%, then a 600 person sample size yields a 4 point margin of error (according to "Best Estimates") and while this is not a large number, a smaller number, around 2-3% is preferable.

3) The wording of the question is what I find most fascinating. Part of the question reads "The following is a statement a Republican candidate could say in support of this proposal." The pollster is trying to be politically, and morally, correct by saying the word COULD, but this word is clearly being swallowed by the rest of the question. (Read the question out loud starting from "Now," which I also think is an interesting and friendly word to use, and I think you will get my point) The pollster is trying to convince the person on the other end of the telephone line that this IS what Republicans will say and therefore they should answer accordingly.

4) I think that the results are very interesting, but not surprising at all knowing what we know about home telephones and about people who tend to answer polling questions. I would like to point out that 51% percent of the people answered were at least "somewhat more favorable" of something that was being stated by a potential Republican member of Congress and was in support of a Republican President.

5) The timing, September 13-14 2006, is fascinating as well. According to the USA Today on September 19th 2006, "President Bush's approval rating has risen to 44% in a new USA TODAY/Gallup Poll. That's his highest rating in a year." Furthermore, the article goes on to say that "The new poll found likely voters more prone to vote for candidates who support Bush on terrorism, 45%-28%, and evenly divided on those who support and oppose Bush on Iraq. More than a quarter said Iraq is their top concern this fall. For the first time since December 2005, a majority of people did not say the war there was a mistake; the split was 49%-49%." Although we would assume that at a time somewhere in late 2006 the President's approval ratings would have been much lower, this is not true and we must take into account what Gawiser and Witt point out (question 8):

"Events have a dramatic impact on poll results. Your interpretation of a poll should depend on when it was conducted relative to key events. Even the freshest poll results can be overtaken by subsequent events. The President may have given a stirring speech to the nation, the stock market may have crashed or an oil tanker may have sunk, spilling millions of gallons of crude on beautiful beaches."