In November, 58 percent of Minnesotans polled told SurveyUSA they approved of Pawlenty’s performance. That fell to 54 percent in December, to 53 percent in January, and in the most recent results from late February, to 48 percent. (See Survey USA’s table here.) At that rate, a smaller proportion of Minnesota voters will approve of Pawlenty this month than voted for him in 2006 (46.7 percent). And by the time SurveyUSA conducts its poll in April, the governor’s approval will fall below even the 44.4 percent he won by in 2002.

Smart Politics observes that Pawlenty’s February showing is still fifth-best among the 14 governors they research (h/t Polinaut). KSTP-TV, which pays for the polls, reported that Pawlenty found the January results encouraging for his re-election in 2010. Going on radio and TV might help — or he could try taking action that voters see as being in the best interest of the state. If he does that, his approval rating by the end of the current legislative session might not tank below his 2002 showing at the polls.

Updated number-crunching after the jump.

Favorable and unfavorable opinions of Al Franken and Norm Coleman

in the Kos poll, 49 percent say they have a favorable opinion of Franken — 4 percent more than those with an unfavorable opinion. A favorable opinion of Norm Coleman is harder to find, being shared by only 41 percent of respondents. But what’s really stunning is Coleman’s negative number: 58 percent say they have unfavorable opinion of him.

Those results are markedly different from the other poll, in which both men’s approval percentages were below 40 percent but neither had unfavorable opinion rates above 45 percent. Variations in polling may explain some of the difference: the KSTP poll allows both “Neutral” and “No Opinion” responses, while Kos only allows “No Opinion.”  In the KSTP poll the two answer categories comprise a full 19 percent of the Coleman responses, but in the Kos poll only 1 percent have no opinion of Coleman.

Polls differ in respondents’ party affiliations

But the most significant difference between the two polls is in the distribution by party affiliation subgroups within their sample populations. In conducting the survey for the left-leaning Daily Kos, Research 2000 included far fewer (27 percent) voters who called themselves politically independent or “other,” and more who identify as Republican (33 percent) or Democrat (40 percent). SurveyUSA, whose polls generally and for KSTP-TV specifically have sometimes been accused of skewing right politically, included only 25 percent Republican and 33 percent Democratic voters — with a whopping 39 percent “independent.”

Perceived fairness of recount process

Sampling differences appear to play out in a question that both polls asked: Was the recount process fair, and if not, to which candidate was it unfair? The Research 2000-Daily Kos results show 63 percent think the recount was fair to both candidates, with only 17 percent finding the process “mostly unfair” to Coleman and 12 percent, unfair to Franken. In SurveyUSA’s poll, 31 percent saw the recount as unfair to Coleman, 3 percent unfair to Franken, and 56 percent fair to both.

Things really got squirrelly within the subgroups of each poll. The more heavily weighted SurveyUSA “independent” subgroup was twice as likely (28 percent) to see unfairness toward Coleman as their Recount 2000-Kos counterparts (14 percent). The dominant Democrats in the Recount 2000-Kos poll managed to find unfairness toward Franken in greater numbers (18 percent, compared to 5 percent in SurveyUSA’s). And SurveyUSA’s relatively scant pool of Republicans was far more inclined (68 percent) to see injury to the Republican candidate than those in equivalent subgroup of the Kos poll (32 percent). In fact, a majority of Kos’ Republicans (54 percent) saw the recount as fair to both candidates, a view shared by only 23 percent in SurveyUSA’s Republican pool.

Coleman court challenge

Both polls measured support for Coleman’s contesting of the recount results in court. The Daily Kos poll found 47 percent opposed and 34 percent in favor, with 19 percent not sure. SurveyUSA had 42 percent in favor and 49 percent opposed — a figure interpreted in news reports as meaning half of Minnesotans want Coleman to concede. SurveyUSA did ask flat out whether Coleman should concede, as the first of three supplied answers (all unlikely) to the question “What do you think should happen next?” Forty-four percent thought Coleman should quit; the other options were “Do the re-count again” (8 percent) and “Hold another Senate election altogether” (31 percent).

Who won, and who would you vote for today?

“Who do you think won the U.S. Senate race, Al Franken or Norm Coleman?” The response to that Research 2000-Kos question is: Franken, 46 percent; Coleman, 41 percent; and “Not Sure,” 13 percent.

It’s intriguing to compare that with the results on another Kos question: “If you could vote again for U.S. Senate would you vote for Al Franken the Democrat, Norm Coleman the Republican, or Dean Barkley, an Independent?” (Barkley should actually be termed not as “an Independent” but as “the candidate of the Independence Party” — a third party with major-party status in Minnesota.) Overall, 43 percent would vote for Franken, 40 percent for Coleman, and 15 percent for Barkley. An element of softness in Democratic voter support for Franken widely observed on Election Day is still perceptible: 83 percent of Democrats polled think Franken won, but only 77 percent would vote for him today.

It’s also intriguing to compare the Kos poll’s election re-run with another SurveyUSA-KSTP poll taken Dec. 7, which also asked, if the election were held again today, “who would you vote for?” The results then were much closer to the Election Day results: Coleman, 41 percent; Franken, 40 percent; Barkley 15 percent.

Margin of error

The Research 2000-Daily Kos poll surveyed 600 voters, with a margin of error of plus or minus 4 percentage points. The SurveyUSA-KSTP poll surveyed 500 voters, with a margin of error of plus or minus 4.5 percentage points. The margin of error increases at the level of subgroups such as party identification.

Does any of this matter?

Minnesota’s Senate election remains unresolved. So are the polls taken since Nov. 4 really election polls? The crosstabs by gender, age and party affiliation and the questions the polls ask combine to conjure up the feeling of a late-campaign pulse-reading of the electorate, a tantalizing forecast of the coming result. Except, of course, those polled have already voted in this race and aren’t likely to vote again. The shifting opinions among the subgroups can matter only as much as they affect the outlooks of three judges, yet to be named, who will decide Coleman’s court challenge. And perhaps 98 or 99 U.S. senators who have the last word on who is seated and who as politicians are used to putting stock in polls.

It’s true that 49 percent of respondents said they disagree with Coleman’s decision to challenge the recount results. But when asked directly what they thought should happen next, only 44 percent said Coleman should concede. And that’s significantly better (for Coleman) than the 55 percent who told the same pollsters a month ago that the loser should not file a legal challenge in court if he thought the process was unfair. (UPDATE: The Daily Kos released a similar poll the day after this one.)

If you read about the Minnesota recount in a book, you’d think it was fiction. Here are some more nuggets from a reading of the SurveyUSA poll results, organized by (mock) book titles.

Men Are from Franken, Women Are from Coleman

KSTP and others emphasize that smaller percentages of Minnesotans in the new poll approve of Franken and Coleman (37 and 38 percent, respectively) than voted for either man on Election Day (42 percent each). But hidden within the new results is a growing gender gap — not that it matters, electorally speaking. Men disapproving of Franken’s handling of the recount comprised 52 percent of the sample pool a month ago; now only 42 percent of guys disapprove of Franken generally. But women approving of Franken in the recount has fallen from 45 percent in December to a shockingly low 30 percent generally favorable today.

Meanwhile, Coleman’s esteem sank in the eyes of both sexes. Women approving of Coleman’s handling of the recount was 48 percent on Dec. 4, with only 38 percent generally approving of him on Jan. 7. Men dropped even more dramatically, from 54 percent to 37 percent.

Zero: The Biography of a Dangerous Recount

The only statistically significant “0 percent” that appears on the pollster’s cross-tabs is for the number of Republicans who said they feel the recount was unfair to Franken. (Interesting: There was no “unfair to both candidates” option in either poll taken after the election.)

How to Get Richie Without Even Trying

Secretary of State Mark Ritchie’s approval rating remains high, falling only from 61 percent a month ago to 56 percent today. His negatives snuck up a bit more, however, with the proportion of voters polled who disapprove of the job he’s done on the recount rising from 26 percent to 34 percent.

KSTP’s report on the survey:

The new poll has a margin of error of plus or minus 4.5 percent, or in the case of the question about Ritchie, 4.4 percent.

But a rerun of the Nov. 4 election, if held today, wouldn’t be any more conclusive than the original was, according to a new poll that Survey USA conducted for KSTP-TV. If Minnesotans could vote again, the results would be just about the same as on Election Day, with Coleman at 42 percent and Franken at 41.

There’s apparently no “give” yet among supporters of either candidate — or indeed, among those who backed the third-place finisher, Independence Party candidate Dean Barkley. Given the chance in the hypothetical do-over to tip the historically close election to either Franken or Coleman, 15 percent of voters would stick by Barkley, who briefly held the same Senate seat in 2002 as Gov. Jesse Ventura’s appointee after Paul Wellstone’s death.

Responses to other questions that KSTP reported suggest that Minnesotans find the recount process fair, approve of Secretary of State Mark Ritchie’s handling of it, don’t want the loser to sue, and do want rejected absentee ballots reviewed.

Here are the results of the KSTP-TV/Survey USA survey questions presented by KSTP on Dec. 7:

Do you think Minnesota’s U.S. Senate recount process has been fair to both candidates? Unfair to Norm Coleman? Or Unfair to Al Franken?
Fair to both candidates: 58 percent
Unfair to Coleman: 20 percent
Unfair to Franken: 13 percent
Not sure: 9 percent

Do you approve or disapprove of the job Minnesota Secretary of State Mark Ritchie has done with the recount?
Approve: 61 percent
Disapprove: 26 percent
Not sure: 13 percent

Do you think the losing candidate should file a legal challenge in court if they think the process was unfair?
Yes: 40 percent
No: 55 percent
Not sure: 4 percent

Do you think absentee ballots that were previously rejected should be reviewed?
Yes: 58 percent
No: 39 percent
Not sure: 4 percent

If another election for U.S. Senate was held again today who would you vote for? Dean Barkley? Norm Coleman? Or Al Franken?
Coleman: 41 percent
Franken: 40 percent
Barkley: 15 percent

In the latest official recount tally of results from the Nov. 4 election (not including 6,655 ballots that either the Franken or Coleman campaigns have challenged), Franken and Coleman are essentially tied at 41 percent, with 17 percent of the vote going to all others (including Barkley).

The poll’s sampling margin of error is ±4.2 percent. It was conducted on Dec. 4, just before the recount finished (most of) its work and just as news was breaking about 133 ballots being lost in Minneapolis. The Survey USA crosstabs include three questions KSTP didn’t include in its report:

Do you approve or disapprove of how the Coleman campaign has handled the recount process?
Approve: 51 percent
Disapprove: 40 percent
Not sure: 9 percent

Do you approve or disapprove of how the Franken campaign has handled the recount process?
Approve: 44 percent
Disapprove: 48 percent
Not sure: 7 percent

In light of the Senate recount process, do you now have more faith in Minnesota’s electoral system? Less faith? Or about the same amount?
More faith: 11 percent
Less faith: 32 percent
Same amount: 53 percent
Not sure: 4 percent

Survey USA’s crosstabs indicate that the results are based on the opinions of 556 registered Minnesota voters. The breakdowns by political party affiliation are surprising, given the sometime right-heavy pools Survey USA has been accused of using: 41 percent Democrat, 27 percent Republican, and 28 percent “Independent” (that’s small “i”-Independent, not necessarily affiliated with the Independence Party).

As Georgia’s U.S. Senate recent run-off experience showed, turnout for special election isn’t the same as for general elections. Did Survey USA target Minnesotans who would be likely to vote if indeed another election is held in this race? “We did not screen for likely voters at any point,” writes Ken Alper of Survey USA in response to an e-mail this morning from the Minnesota Independent.

Here is KSTP’s report:

This is the second e-mail about this poll release situation and I have no idea what it’s about.  We release poll results to the campaigns only if they ask for them.  And then only with the understanding they are not to shared with anyone else in anyway (news releases, etc) until after we air them.  As for the last poll, I had one phone message from the Paulsen campaign Sunday morning after they heard on “At Issue” that we would have results Sunday night.  I called them back about five hours later and left them a message to call back for the results.  They never called. I’m not certain if they got them from someone else, but generally I coordinate all polling and release of the information.

I also asked Hauser about whether it was true, as David Dillon’s campaign manager told me today, that complaints about the Independence Party candidate being left off Survey USA’s candidate preference question in the firm’s first poll last week led to a do-over survey in its second 3rd District poll, which was released yesterday. Hauser:

Yes, that is true.  SurveyUSA had earlier done a poll in that race for ”Roll Call” out of Washington, D.C. back in September.  “Roll Call” only had them survey Madia and Paulsen, not understanding the importance and tradition of third-party candidates in Minnesota.  We later commissione[d] our own poll (as a separate SurveyUSA client) and included Dillon.  When we ordered our final poll in the race, SurveyUSA mistakenly duplicated the original “Roll Call” poll question that only included two names, rather than ours, which included all three.  When I pointed this out they agreed to do it again free of charge.  The first poll [done last week] had Paulsen up by one, the second (with Dillon included) had Paulsen up by 5.

The new results arrive six weeks after SUSA conducted a similar poll for Roll Call newspaper that showed Paulsen with a 3-point lead. And only a week ago, 400 likely voters told pollsters working for the Democratic Congressional Campaign Committee they favored Madia by a 5-point margin (a spread that equaled that poll’s margin of error).

Independence Party candidate David Dillon was the choice of 8 percent in the new SUSA poll, with only 3 percent undecided. In August, SUSA didn’t ask about Dillon by name but reported that 10 percent preferred “Other” to the two leading candidates, with 6 percent undecided.

While the new poll shows a 6-point swing toward Madia in overall results, other numbers buried deep within the survey’s crosstabs show widening splits as compared with the SUSA poll of six weeks ago:

- Madia widened his lead among moderates, who now say they prefer him to Paulsen by a margin of 19 points, 54-35 percent; in August it was 46-32.

- The barrage of negative ads from the DCCC appears to have taken a toll on people’s opinion of Paulsen, whose “Unfavorable” rating shot to 31 percent, from 12 percent in August.

- The Paulsen campaign’s negative attacks haven’t pulled Madia down. The 39 percent who hold a “Favorable” impression of Madia in the new poll represent a 10-point increase from August’s results.

- As Congress fell further into disfavor amid weeks of legislative disarray and financial chaos, 3rd District support for Sen. John McCain and Paulsen remained flat. But the proportion of poll respondents who say they’d vote today for Sen. Barack Obama and Madia rose since August by 10 and 11 points, respectively, among the 78 percent of those surveyed who now say they disapprove of the job Congress is doing.

And here are observations gleaned elsewhere:

SUSA still seems to have a questionably GOP-tilted sample of young voters (they break for Paulsen by a 51-40 margin in this poll), but that might be offset by a possibly-skewed 53-38 Madia lead among 50- to-64 year-olds. - Swing State Project

Ashwin has a greater lead over Paulsen in this poll than Obama has over McCain … [and] he has more crossover support than Paulsen, gaining the support of 11 percent of conservatives, while Paulsen only got the support of 6 percent of liberals. – Minnesota Campaign Report

The Swing State Project also observes in the poll results that the indicator that drew Democrats’ attention in the first place — Sen. John Kerry’s narrow 2004 loss in the 3rd district — has swung Democrat: Obama leads McCain by 2 percentage points.

Meanwhile, the 3rd district moves up a few notches on Real Clear Politics’ rankings of the nation’s 50 most competitive House races, from No. 40 on Sept. 19 to No. 35 today.