NBC / WSJ poll appears to oversample the lumpenproletariat, who don’t vote
Overview: today’s NBC/WSJ poll appears to substantially oversample people who do not vote. At least 40-45% of the sample are independents versus a quarter of the actual electorate. 37% of the sample are self-described “moderates” versus the 3% in the more highly screened Battleground poll of likely voters. We have no idea if the NBC/WSJ poll has any value whatsoever other than as propaganda.
We recently discussed a Battleground poll of likely voters that showed a 61% conservative, 34% liberal split in the electorate — self-described “moderates” were a mere 3%. This made sense to us in that we expect that most real voters know where they stand on important issues. Today we looked at a poll by NBC/WSJ of registered voters, and its self-described “moderate” group was 37% of the total, though the poll offered fewer choices. Here’s the liberal and conservative split from the NBC/WSJ poll we discussed below:
Here’s the same split from the Battleground Poll we discussed the other day:
We believe that a lot of these self-described “moderates” in the NBC/WSJ sample are not real voters, although they claim to be registered voters. To explore this thought a little more, let’s take a look at the party identification numbers from the NBC/WSJ poll, which make absolutely no sense compared to exit polls of actual voters:
This is nuts. Democrats in the WSJ/NBC sample were 28%; Republicans were 27%. Huh? These figures are out of whack — 10 points below — both parties’ actual participations in elections, which in 2004 were 37% each for each party. We’ll quote Michael Barone from a previous post of ours:
The 2000 exit poll showed an electorate that was 39 percent Democratic and 35 percent Republican. The 2004 exit poll, which was tilted toward Democrats, found a dead heat: 37 percent to 37 percent.
Thus we see that the WSJ/NBC poll included in its sample at least 40-45% independents, while in reality only a quarter of those who care enough to actually vote are independents. We are of the view that the NBC/WSJ poll appears to be targeting the lumpenproletariat, ir order to produce the same kind of propaganda that so many other polls and the MSM are pushing.
There appears to be emerging a kind of polling underclass in America, the lumpenproletariat (or economic hypochondriacs, as George Will put it and as we have previously discussed). This is a group that can be depended upon to whine and complain no matter what actual conditions are, but do not apparently vote in large numbers. Pollsters and the MSM possibly do themselves no good by pandering to their inertia and dissatisfaction.
We have no idea how the election will actually go when it happens in a few weeks. However, if the actual results are less favorable to Democrats than currently appear to be the case in the polls, the MSM will have only themselves to blame for their disappointment. Polls like this from NBC/WSJ will perhaps be seen in retrospect as exercises in self-delusion.
October 20th, 2006 at 6:15 am
We have no idea how the election will actually go when it happens in a few weeks. However, if the actual results are less favorable to Democrats than currently appear to be the case in the polls, the MSM will have only themselves to blame for their disappointment.
Oh no, no, no! They will have massive GOP voter fraud to blame! Guaranteed! That couldn’t care less about the damage these claims do to the very core of our democracy by eroding the public faith in the voting process – there are short term gains to be had, after all.
October 21st, 2006 at 10:11 am
I find your analysis seriously bizarre.
You seem to have the attitude that “moderates” are people who dont know what they believe in. That is ludicrous. Being moderate does not mean be indecisive – to the contrary, most moderates that I know are strongly of the view that the “pure” ideologies fail to capture the mix of positions that they strongly hold.
It is not proper to claim that the Battleground poll shows only 3% moderates. If I interpret the table properly, it seems that “moderate” was not an option for choice. If you force people to choose lib leaning or conservative leaning, then they might do so, but if given the choice of moderate, then many will choose that.
Lets parse out the percentages in the NBC poll. If you ignore the moderates, then their results are 37% conservative, 21% liberal. If you scale that up to 100%, that equals 64% of those responders being conservative, 36% liberal – very similar to the Battleground poll. In other words, the only difference between the polls is that one allows moderate conservatives and moderate liberls to choose the term “moderate” – and some do, whereas the other poll forces them to one side or the other. No real difference at all, in terms of results.
The party results in the NBC poll also strike me as far more accurate than your estimates. The Rasmussen organization (GOP-leaning) keeps regular monthly track of voter ID, and their latest results are (roughly, by memory) 37% Dem, 32% Republican, 30% independent (showing considerable deteriation in the GOP numbers over the past year).
The NBC poll has more categories, so that might be confusing you. To make the two polls comparable, lets do the following: I go along with your conflation of “strong” and “not so strong” into one category per party. Then lets divide the “Independent Democrat” repsonse in half, allocating half to “democrat” category, and half to “independent’ category. Do the same for “independent Republican”. New results: Dem-35%, Rep-32%, Independent – 30%. Almost exactly what Rasmussen finds (perhaps NBC is slightly to favorable to Repbulicans).
Your analysis seems entirely driven by a preconcieved desire to spin things a certain way. And you demonstrate near total illiteracy regarding polls, statistics, or any of the intellecutal tools necessary for an insightful analysis. Bad show.