A plurality of Americans trust Congressional Democrats more than either President-elect Donald Trump or their GOP counterparts on Capitol Hill to develop solutions to the major challenges facing the country, according to the newest installment of the quarterly Allstate/Atlantic Media Heartland Monitor poll.

Similarly, President Obama is enjoying his highest approval ratings in the survey since July 2009, with 56% of respondents rating his job performance positively, compared to 40% who give him a thumbs-down. And a plurality say his administration has increased opportunities for people like them to get ahead—Obama’s highest marks on that score since just after his 2012 reelection.

The results overall present a muddied picture of an electorate feeling ever more encouraged about their personal finances and yet ready, with qualifications, to welcome change in Washington. “One thing for sure we know is this is not a mandate for Trump,” says pollster Bill Cullo, who conducted the survey. “Nor does that mean that Democrats have a clear path either.”

The survey comes as each party marshals evidence from the election results to try to claim majority support for its agenda. Vice President-elect Mike Pence on Tuesday pointed to the Trump ticket’s victory in 30 states and more counties than in any election since Ronald Reagan’s, arguing the incoming administration has a historic mandate. Democrats, meanwhile, note that while they remain locked out of power on the Hill, they gained seats in each chamber; and Hillary Clinton won 2.7 million more votes than Trump nationally, and counting, the widest gap ever for a candidate who lost the electoral college.

The Heartland survey reflects some of the split thinking evident in those competing results. By a margin of 50%-36%, most Americans believe the country is on the wrong track, but that is the most hopeful view they’ve registered in roughly seven years. For the first time since the poll launched—in April 2009, amid the worst of the recession —a majority of respondents now describe their personal financial situation as either excellent or good. That result aligns with macro trends that show consumer confidence, for example, hit a nine-year high in November as the Dow Jones Industrial Average set a new record.

Yet apparently following Trump’s lead, most respondents now back growing the economy through new spending on education, training, infrastructure and research—even if that necessitates raising taxes and running up deficits. The ranks of those who’d rather focus on reducing the deficit have shrunk by half over the last four years, with the survey showing only 12% who want it prioritized over growth, down from 22% around the same time in 2012.

“There are a lot of pieces of dust in the air after an election like this,” Cullo says.

On individual issues, voters remain split over who they trust to lead. Trump claims solid pluralities on a suite of “strong man” concerns, with respondents preferring him over either party in Congress on stopping the offshoring of American jobs, renegotiating trade deals and protecting the homeland from terrorism. But they give Congressional Democrats the edge on issues that arguably require more nuance, including fostering compromise in Washington, increasing equality, reforming the tax, immigration and criminal justice systems, reducing gun violence, improving elementary education, easing college affordability, expanding clean and renewable energy, and revising Obamacare.

On the last score, while Congressional Republicans have pledged to make repeal of Obama’s signature healthcare overhaul one of their first moves in the new Congress, this survey found voters support tweaking the existing law over ripping it out and replacing it wholesale, 60%-36%. The results suggest Trump may have to earn Americans’ support “issue by issue,” Cullo says. “That doesn’t mean he can’t. But people will be saying, ‘Ok, give us your argument.’”

The survey of 1,000 American adults—half reached on landlines, half on cell phones—was conducted Nov. 16-21 and has a margin of error of 3.1%.