At least Reuters/Ipsos was more successful than the WaPo/ABC poll yesterday.  The latter had a D+9 sample but could only muster a tie between Barack Obama and Mitt Romney, whereas the former takes a D+5 sample (D+4 when counting leaners) and transforms it into a six-point lead for Obama:

President Barack Obama expanded his lead over Republican challenger Mitt Romney to 6 percentage points in the White House race this month as voters became slightly more optimistic about the economy, a Reuters/Ipsos poll showed on Tuesday.

Four months before the November 6 election, Obama leads Romney among registered voters 49 percent to 43 percent. In June, Obama held a slim 1-point lead over the former Massachusetts governor.

Obama’s improved standing was fueled in part by a slight rise in optimism about the future, with the number of Americans who think the country is on the wrong track dropping 5 percentage points to 58 percent.

People became more optimistic about the economy?  Pray tell, when did that happen?  The NFIB shows small-business confidence dropping to eight-month lows, retail sales numbers have slid significantly in the past couple of months, and we’ve had three straight disappointments in monthly jobs reports. Just two weeks ago, Gallup found economic confidence eroding — and eroding significantly, just the opposite of what Reuters/Ipsos reports.

What might the source of this optimism be?  The sample maybe the issue here, but it’s not as bad as the WaPo/ABC poll, as noted above.  While a D+5 turnout would be a bit of a surprise in this cycle, it’s not entirely out of the question.  The ratios are odd, however; in the registered-voter subsample (where the Obama lead is found), the D/R/I is 31/26/43.  When adding in leaners, it’s 46/42/13 (rounding causes the 101%).  The 2008 wave election for Democrats produced an actual D/R/I of 39/32/29, while in 2010 it was 35/35/30.  In neither case did Republicans only comprise 26% of the electorate, but in neither case did Democrats only comprise 31%, too.

The Reuters results look like just bad sampling of another sort, creating the outlier effect into which most Reuters/Ipsos polls belong.  That’s further confirmed by a new Quinnipiac poll which sees the race as a virtual tie despite having an almost identical D/R/I of 31/27/36:

Driven by a yawning marriage gap, and a 2-1 lead among single women, President Barack Obama gets 46 percent of American voters to 43 percent for Gov. Mitt Romney, largely on the support of singles, according to a Quinnipiac University national poll released today.

Married voters back Romney 51 – 38 percent, while unmarried voters back President Obama 54 – 34 percent, the independent Quinnipiac (KWIN-uh-pe-ack) University poll finds. Among married voters, Romney leads 54 – 35 percent among men and 49 – 42 percent among women. Among single voters, Obama leads 47 – 38 percent among men and 60 – 31 percent among women.

Supreme Court Chief Justice John Roberts gets a higher job approval rating, 46 – 34 percent, than the president, who logs a negative 45 – 49 percent score.

In this case, respondents are hardly confident in the economy or Obama’s stewardship of it:

American voters disapprove 55 – 40 percent of the way Obama is handling the economy, but split 45 – 46 percent on whether Obama or Romney would do a better job on the economy. Romney would be better for their own economic future, 47 percent say, while 44 percent pick Obama. By a 50 – 44 percent margin, voters have an unfavorable opinion of Obama’s plans for the economy, compared to 46 – 40 percent unfavorable for Romney.

Just as in yesterday’s WaPo/ABC poll, Obama’s immigration pander did no good; he’s underwater on immigration by a wide margin, 39/52.  The ObamaCare legal victory has not changed its status as a political albatross, either, with Obama getting a 43/52 on health care.  Probably the most worrisome for Obama is that he is now being perceived negatively, with a 45/48 favorability rating — perhaps the result of his angry attacks on Romney while avoiding the risk of putting an agenda forward for a second term.