A new poll shows that Democrats and Republicans do not see eye to eye on the US economy or the state of its government—nor do young adults and seniors.

According to the results, which come from 1,000 responses to a 24-question survey conducted online on October 10 and 11, 77% of self-identified Republicans in the US believe the economy is getting better, while just 14% of Democrats believe the same.

And while 51% of adults age 65 or older were optimistic about the country’s economic future, only 31% of adults ages 18 to 29 thought things were improving.

Asked about the current state of the economy, just 28% of Democrats characterized it as “good,” compared to 85% of Republicans. And while nearly two-thirds of seniors thought the economy was doing well, only 41% of adults younger than 30 agreed.

“We saw evidence of deep partisan and age divides in our country in the 2016 and 2018 elections, and the results of this poll confirm those differences in opinion haven’t eased,” says Susan Moffitt, director of the Taubman Center for American Politics and Policy at Brown University. “If younger generations were to turn out to vote at higher rates than usual in the 2020 election, we could see changes in leadership.”

Disagreement on Everything but Distrust in Government

Poll respondents were also divided along partisan lines on what they perceive as the biggest problem facing the country today, Moffitt says. Nearly half of Democrats polled said they believe the problem is US President Donald Trump, ahead of the economy, health care, education, and eight other hot-button issues; Republicans, by contrast, said they believe the biggest problems are immigration (28%) and government corruption (21%).

There was at least some disagreement between Republicans and Democrats—and between younger and older adults—on almost every topic covered in the poll, from safety in schools and cities to personal finances. But the majority of the population did agree on one thing: the government is not trustworthy.

In response to the statement, “When I think about our government, I think that it can be trusted to do what is right,” 56% of all respondents says they somewhat or strongly disagreed. Seniors were the least likely to find the government trustworthy, with just 15% registering agreement with the statement; Republicans were most likely, with 27% agreeing.

More than half of all respondents also disagreed with the statement “I think that our government listens to people like me” and agreed with the statement “I think that [the government] benefits other people more than it benefits me”—revealing, Moffitt says, that many millions of Americans don’t feel heard or supported by national leaders.

“These views reflect a core challenge facing American democracy today—the challenge of representation,” Moffitt says.

The poll also shows that Americans between the ages of 50 and 64 feel the least secure about their personal finances. Just 39% of those in that age group said they felt good about their current savings, compared to 63% of seniors, and 41% of adults younger than 30. And 48% of 50- to 64-year-olds said they did not feel confident about their ability to pay for an unforeseen expense, compared to 20% of seniors, 36% of 30- to 49-year-olds, and 39% of under-30s.

The poll also reveals that most Americans believe safety in their own communities has neither improved nor worsened in the past decade—but half of respondents said they believe schools and cities are less safe today. Forty-two percent said they believed that the nation as a whole had become less safe.

How the Poll Worked

The poll results were based on an online panel of respondents that the nonpartisan polling organization YouGov selected using a technique called matched random sampling.

“YouGov starts by drawing a random sample from the target population—US adults—using the 2016 American Community Survey (ACS) one-year sample,” explains Paul Testa, an assistant professor of political science.

“Next, YouGov uses an algorithm to select respondents from its online panel of 1.8 million US residents who are a close match to the individuals sampled from the ACS on gender, age, race, and education. The final results are then weighted by respondents’ 2016 presidential vote choice and a four-way stratification of gender, age, race, and education.”

Overall, the approach yields a national sample of comparable utility to other common methods like telephone surveys using random digit dialing, Testa adds.

The raw data are available on the Taubman Center website, along with an interactive tool that allows users to quickly create visual representations of the data.

The full sample has a model-based margin of error of +/- 3.4%, with a larger margin for subgroup results. This margin is based on a specific set of statistical assumptions rather than conventional random sampling—only if these assumptions hold is the margin accurate.