The first poll is the NBC-Wall Street Journal Poll released today which shows Obama up +8. Here's a few puzzling things about this poll. The first question about the poll is there sampling method. For some strange reason, they only accepted responses from individuals who only had a cell phone and no land line. Any respondents who answered that they had both were terminated from the poll.
1b. Do you have a landline in addition to your cell phone, or is your cell phone the only telephone you have? +
Likely
Voters
Have a landline ................................. - TERMINATE [139]
Cell phone is only telephone............. 100 CONTINUE
Not sure .......................................... - TERMINATE
+ Results shown reflect responses among likely voters.
As you can see, anyone who answered that they had a landline or unsure were terminated from the poll. Now this makes me wonder what type of population target would only have a cell phone. I understand it is becoming more and more popular, however, the pollsters do not share why this is such an important qualifier to participate in their poll.
The second point of confusion is the question dealing with the potential voters level of interest in this election. Basically, the only two criteria that allow you to vote in this poll is either if you rate it a 10 out of 10 or you have already voted.
2. Using a ten-point scale, please tell me how likely you are to vote in the November fourth elections for
president and Congress. If you are certain that you will vote, pick a number closer to "eight," "nine," or
"ten." If it is less likely that you will vote, use a number closer to "one," "two," or "three." You may choose
any number from one to ten. +
Likely
Voters
10, certain to vote ............................. 91 CONTINUE [140-141]
9 ....................................................... -
8 ....................................................... -
7 ....................................................... -
6 ....................................................... -
5 ....................................................... - TERMINATE
4 ....................................................... -
3 ....................................................... -
2 ....................................................... -
1, less likely to vote........................... -
Already voted (VOL) ....................... 9 CONTINUE
Not sure .......................................... - TERMINATE
+ Results shown reflect responses among likely voters.
Now for some reason they are using this response as their only way to determine whether or not someone is a likely voter. Now, I'm far from an expert, but why would someone who only viewed this election as a 9 out of 10 in importance be considered an unlikely voter?
Next on the hit parade is the Marist Poll which shows Obama +9 over McCain. Now this poll really makes me wonder about its validity all together. To me, if you were a pollster who really wanted to claim the validity of your poll, would you describe your respondents as the following:
Nature of the Sample: 830 Americans
This survey was conducted on November 2, 2008. 830 adults 18 years of age or older within the
continental United States were interviewed by telephone. Telephone numbers were selected based
upon a list of telephone exchanges from throughout the nation. The exchanges were selected to
ensure that each region was represented in proportion to its population. The results of the entire
survey are statistically significant at ±3.5%. There are 747 registered voters and 635 likely voters.
The results for these sub-samples are each statistically significant at ±4%. The margin for error
increases for cross-tabulations.
First of all, I'm glad that they sampled 830 Americans. I mean, if the results from Zimbabwe would have showed Obama +9 I would definitely be thinking that Obama could be posed for a landslide victory. But here is the problem with this description. There is none. Who is to say they didn't call San Francisco, New York City, Detroit, and Atlanta to come up with their sample? I mean, that would cover each region of the country right? Second, I don't know how in the hell you can take responses from 830 Americans and determine what nearly 130,000,000 plus potential voters will do. That is a sample of .0000061% of the population. That's roughly equivalent to to asking one person in California how the voted and then calling a race on it as an exit poll.
Next up is the Rasmussen Reports Poll. To see the full sampling method of this poll, you have to buy a subscription the the site, so you have to be a little leery to begin with. Second, Rasmussen admits to setting a goal to weight its polls by seeking out 9.7% more Democrats to responds than Republicans, which is rather unprecedented. There has never been that great of disparity in turnout before and I doubt that this election will be that much different. But hey, it sells subscriptions right?
Gallup is its own animal. They've hedged their bets using three different sampling methods. There final prediction is Obama 55% to McCain 44%. What doesn't make sense here is that third party candidates Bob Barr, Ralph Nader others have combined been attracting anywhere from 3-6% of the vote. So unless Gallup is suggesting that 105% of the electorate is going to vote this year, which is totally possible with ACORN registering more voters in Indianapolis than there are legal voting aged people, there is one red flag against them.
Then Gallup comes up with this gem of doublespeak:
Survey Methods
Results are based on telephone interviews with 3,050 national adults, aged 18 and older, conducted Oct. 31-Nov. 2, 2008. For results based on the total sample of national adults, one can say with 95% confidence that the maximum margin of sampling error is ±2 percentage points.
What the hell does that mean? I'm not exactly sure how you can have a percentage of a percentage. It's kind of like saying there will be a 90% chance of having 20% chance of showers this evening. How am I supposed to read that. Am I supposed to add the 5% uncertainty in their own polls to the 2% admitted margin of error? Is this their way of trying to say, hey, we could be off by 7% but it sounds a lot better on us.--OR--
Should I multiply the two? Would that mean that they are trying to say that there is .0001 chance they are wrong? Do I carry they one? What about the numerator?
--OR--
Was in Col. Mustard in the Kitchen with the Candlestick?
Okay, then there is the Diageo/Hotline poll. Now if you visit the link provided by Real Clear Politics you get a page that looks like a third graders attempt at building a web page.
Once again, third party candidates aren't even factored into this poll:
The Early Line: Diageo/Hotline Tracking Poll
Obama/Biden 50%McCain/Palin 45%
Undec 5%
So, as you can see, you either supported Obama, McCain or were Undec (which I'm assuming means undecided but could also be an abbreviation for a new UN agency Obama is proposing, totally funded by the US of course). There is no factoring for the 3-6% of third party votes.
Diageo/Hotline's description of their sample consisted of one louse sentence:
Today's Diageo/Hotline tracking poll, conducted 10/31-11/2 by FD, surveyed 887 LVs and has a margin of error of +/- 3.3%. Party ID Breakdown for the sample is 41%D, 36%R, 18%I.
Now I ask you folks. If you were a student and used that as your sample description in a junior high class, would you get away with it? Probably not. But, apparently its good enough for professional pollsters.
One last thing about this poll which is peculiar is this statement:
Based on an analysis of responses to questions on candidates' images, party ID and demographics, "Refused" responses in the head-to-head question were assigned to Obama, McCain or undec.
Um, so what exactly does that mean? Does that mean if someone responded that they were a registered Democrat and then refused to answer who they were voting for that the vote automatically went to Obama? The poll apparently only asked 3 questions: What party are you, did you vote yet and who will/did you vote for. I mean, if electing a president came down to who had more registered voters in their party, this poll and statement would make a lot more sense. Of course, ACORN has also been using the same philosophy, "stock-watering" the populace with fraudulent voters so pollsters skew and weigh their numbers to unrealistic proportions. I mean, they have to account for the fact that Mickey Mouse will be voting 15 times this year in 8 states and 3 counties in Ohio.
So folks, I'm not a professional by any stretch and I'm sure the Obama-trons will say that 50,000,000 Frenchmen can't be wrong, but there are a lot of fundamental questions about these polls that are accepted for gospel truth. The only thing we can do is make sure we get out and vote tomorrow if you haven't already and make these pollsters eat crow along with the rest of the main-stream media. That is of course if MSNBC doesn't call the race for Obama at 08:01 AM EST.
Posts
No comments:
Post a Comment