Research 2000 for Daily Kos. 3/1/2010-3/4/2010. Registered Voters. MoE 2.8% (Last week's results in parentheses):
| FAVORABLE | UNFAVORABLE | NET CHANGE |
---|
PRESIDENT OBAMA | 53 (55) | 43 (40) | -5 |
| | | |
PELOSI: | 36 (38) | 56 (52) | -6 |
REID: | 27 (23) | 66 (67) | +5 |
McCONNELL: | 21 (20) | 63 (62) | 0 |
BOEHNER: | 19 (20) | 62 (62) | -1 |
| | | |
CONGRESSIONAL DEMS: | 37 (39) | 61 (59) | -4 |
CONGRESSIONAL GOPS: | 24 (19) | 67 (65) | +3 |
| | | |
DEMOCRATIC PARTY: | 39 (40) | 57 (55) | -3 |
REPUBLICAN PARTY: | 30 (29) | 66 (62) | -3 |
Full crosstabs here. This poll is updated every Friday morning, and you can see trendline graphs here.
If you are wondering what events in this week's news cycle might provoke this much volatility in the tracking poll numbers, the answer is: none, actually. The sharp adjustments to the numbers this week are not some snapshot of a wildly changed electorate. They are, rather, an adjustment to the tracking poll itself, a change that fundamentally alters the universe of people our pollster Research 2000 seeks opinions from as the poll is conducted.
Indeed, while the relative favorabilities of the public figures may not have changed much this week, the people that were asked about it did.
This week marks a new phase in the Daily Kos State of the Nation Tracking Poll, where the respondents have gone from being "adults" to the more restrictive universe of "registered voters".
To do so, the number of respondents has been knocked backwards from 2400 down to 1200 (although that only costs us about eight-tenths of point in statistical margin of error). Research 2000, has also made changes on the demographic weighing -- narrowing the universe of respondents from "all voters" to "registered voters" drags down the numbers of under-represented groups, like ethnic and racial minorities and younger voters. Those decisions are made by the pollster, not us.
The end result? It's hard to know. The proper thing to do, in all likelihood, is take these as the first numbers of the tracking poll in its new edition, rather than trying to make comparisons to a poll that was culled so differently.
There are two differences in the universe of registered voters that are markedly different from our previous universe of adults.
For one thing, this group is a great deal more willing to indicate that they are "unsure" about their voting plans in 2010. While the percentage of respondents before was in the single digits, it is now up to nearly 30% in most cases.
Also, some of the indecision about voter INTENT in the 2010 cycle has been removed. Whereas our previous incarnation of the generic ballot test had a huge number of voters undecided about which party they'd like to see elected in this cycle, the new universe of respondents is a bit more clear about their intent. Whereas more than a quarter of adults were undecided last week about their partisan preferences (with the Democrats holding a narrow 37-36 edge), that number drops to just 13% this week (with the Democrats staked to an incrementally wider 45-42 lead).
Next week, we will have the ability to start making some genuine comparisons on the favorabilities of the entities in question, as well as getting a clearer picture of things like voter intensity and general mood (there, the universes did not collide much. This bunch is as generally dissatisfied with the state of things as the adult class at-large was).
One final thought: it is practically inevitable that someone, somewhere, out among the right-wing noise machine will make the charge that this change in the polling universe was designed to "cook the books" in favor of Democrats everywhere. Unfortunately, a charge like that simply won't pass the idiot test (insert your own one-liner here).
A cursory look at the demographic weighting shifts (look at the "Full Crosstabs" link on the polling page) shows us that our new universe is:
-- More Republican (from 22% to 25%)
-- More Caucasian (from 71% to 75%)
-- Less African American (from 14% to 12%)
-- Less Latino (from 12% to 8%)
-- Less Young (18-29 year olds go from 18% to 16%)
If we did this to cook the books, we sure did a crappy job of it.