New poll: Democrats' real problem isn't being too liberal — it's being seen as too weak

Americans, including swing voters, see the Republican Party as 20 points more extreme than Democrats — and the Democrats as weak and ineffective. So why would the *Democrats* moderate?

This article is the last in a four-part series presenting findings from the latest Strength In Numbers/Verasight poll, conducted February 18-20 among 1,566 U.S. adults. The margin of error is +/- 2.5 percentage points. See the full poll release and methodology for details.

---

Americans are unhappy with the way things are going in the country, and don’t feel particularly well represented by either major political party. In our new February Strength In Numbers/Verasight poll, 53% of U.S. adults say the Democratic Party is out of touch with the concerns of most Americans. An identical percentage — 53% — say the same about Republicans.

The conventional reading of numbers like these — especially after Kamala Harris’s loss in 2024 — is that when voters say a party is “out of touch,” they mean so in terms of ideology. For the Democrats, for example, “out of touch” gets mapped onto “too progressive” — with the implication that to become “in touch,” the party needs to tack to the ideological center.

Our February poll tested this assumption directly, and the assumption is simply wrong. When Americans say Democrats are “out of touch” they don’t only — or even primarily — mean “too progressive.” This type of thinking is another example of people committing the Strategist’s Fallacy instead of thinking about what is really being measured by the poll question being asked.

Whether a party is “in touch” or “out of touch,” we found, is a product of more than just ideological perceptions. In our survey, U.S. adults call Democrats weak (48%), ineffective (47%), and out of touch — but also empathetic (54%) and principled (49%). They call Republicans extreme (60%), elitist (57%), and cruel (51%). Both parties have brand problems. But the kind of problem is fundamentally different from what most people are assuming — and that difference matters enormously for 2026.

Today’s Chart of the Week: Americans — including swing voters — say Democrats should fight harder, not moderate. Republicans, in contrast, have a major extremism and cruelty problem. I sit down and actually do the math on how many votes Democrats could flip by messaging on ideology versus fighting for what members believe in.

---

Democrats are weak, Republicans are extreme

In our February poll, we asked voters whether each of 10 adjectives describes the Democratic and Republican parties. Each person was asked to rank how well each word — such as “extreme”, “elitist”, “tough”, and “weak” — described both parties on a scale from 1 to 10, with 10 indicating the word described the party very well.

The Republican Party’s defining traits in voters’ minds are extreme (60% agree), elitist (57%), tough, (56%) and cruel (51%). The percentage of Americans agreeing with descriptions of positive traits is comparatively smaller: just 41% say the party is competent, 41% say principled, and only 31% — less than a third — say the GOP can be considered empathetic.

The average American sees Democrats in a much different light. The top descriptors of the party are empathetic (54%) and principled (49%). Comparatively few people think of it as “tough” (31%), and nearly half the country calls the Democrats weak (48%) and ineffective (47%). Democrats’ competence rating is 46% — five points higher than the GOP’s — but it’s the weakness and ineffectiveness labels that dominate voters’ impressions and national discourse about the party.

The chart above is ordered by the gap between the percentage of Americans who say each descriptor applies to each party — and these gaps are revealing. Democrats lead on empathetic by 23 points (54% versus 31%) and on principled by 8 (49% versus 41%). The GOP leads on exactly one positive trait: toughness (56% versus 31%). That’s it.

Being seen as tough is an advantage in a politics where voters want parties to deliver for them no matter what, but it’s likely not worth being called cruel and elitist. In our poll, Democrats lead the Republicans on the U.S. House generic ballot by 10 percentage points among registered voters. At least in the short term, that’s a worthwhile trade.

But the Democrats’ weakness problem stands out as a particularly strong signal of intra-party dissatisfaction. When we look at how each party’s own identifiers rate their own party, the weakness gap for the Democrats really jumps out. Just 53% of Democrats call their party tough, compared to 80% of Republicans. And 31% of Democrats say their own party is weak — almost three times the 13% of Republicans who say the same about theirs.

On most traits, partisans rate their own party similarly. Democrats and Republicans are within a few points of each other on being competent, principled, and cynical. But on toughness and weakness, Democrats are far more self-critical. That matters electorally: a party whose own base doubts its strength will struggle to turn out its base. And in an era of close elections, that is not a loss a party can afford.

---

Swing voters think Republicans are extreme, too

National numbers are useful, but elections are decided by persuadable swing voters, not partisans. What do these same patterns look like if we restrict the analysis to pure political independents?

The results are shown below. Among independents, 57% call Republicans extreme, 54% say cruel, 54% say elitist. Democrats, in contrast, are disproportionately called empathetic, principled, and weak — though fewer than 50% said so.

Democrats’ brand among independents differs from their brand among all adults in one hugely important way: the “extreme” label fades significantly. Just 35% of independents call Democrats extreme, versus 42% of all adults. That 7-point drop suggests the people telling pollsters Democrats are “too extreme” are disproportionately Republicans. Among actual swing voters, “extreme” isn’t really a Democratic problem. The GOP, on the other hand, has a big problem with perceived extremism.

What is a Democratic problem is that 45% of independents say Democrats are weak, and 44% say ineffective — nearly identical to the national numbers. Contra the extremism drop-off, these traits aren’t sourced to partisan talking points. Just 23% of independents call Democrats tough, the second-lowest rating either party gets on any trait — surpassed only by the GOP’s 22% on empathy among independents.

But Democrats lead among independents on being empathetic (43% versus. 22%), principled (38% versus. 30%), and competent (36% versus. 32%). And independents are almost as likely to call Republicans weak (39%) as Democrats (45%) — a 6-point gap, compared to 16 points among all adults. Independents don’t see the GOP as an unstoppable force. They see two flawed parties — one that’s full of cruel, elitist extremists, and one that’s weak and ineffective.

---

What is the value of moderation versus looking tough?

The post-2024 conventional wisdom says Democrats need to moderate ideologically to win in 2026 and 2028. Some recent advice to the party has been to stop talking about immigration, climate, and LGBT issues, and to move to the right on spending. The idea here is fairly intuitive: if voters think you’re too extreme, become less extreme.

But the survey data among independents above suggests this gets the problem exactly backwards. Independent voters do not think of the Democrats as more extreme than the Republicans. So let me show you some rough math on the value of moderating versus getting tough.

The moderation scenario

Let’s start by establishing partisan baselines for two groups of voters: (1) those who say the Republicans are less extreme than the Democrats, and (2) those who say the Republicans are more extreme. The first group — voters who view the Republicans as less extreme than the Democrats — say they are going to vote 87% for Republican and 13% for Democratic candidates in the 2026 U.S. House midterm elections.

The second group, in contrast, would vote Democratic by 68 points in November — 84% for Democrats to 16% for Republicans.

Now what we need to know is what proportion of voters sit in each group. Looking at each voters’ individual scores on the extremism trait, and removing people who rated the Democrats and Republicans equally, we get the following composition of voters:

• The GOP-more-extreme group: 62% of voters with a preference on the 2026 generic ballot

• The Dem-more-extreme group: 38%

You can use these composition percentages with the above vote loyalty margins to calculate the impact on Democrats’ vote margin like so:

• GOP-more-extreme group: 0.62 * 68 = D +42

• Dem-more-extreme group: 0.38 * -74 = D -28

• sum: D+14

And now we can play around with some simple composition math. If we increase the percentage of voters who view the GOP as more extreme to 67%, like this…

• GOP-more-extreme group: 0.67 * 68 = D +46

• Dem-more-extreme group: 0.33 * -74 = D -24

• sum: D+22

... the predicted Democratic margin among voters rises 7 points. We can thus speculate that making the Democrats 5 points less extreme (in relative terms) would increase their overall vote margin by about 7 points.

The “get tough” scenario

Let’s calculate the partisan loyalty of our two groups again, but this time on a different variable: how tough each party is. The first group — voters who rate the Democrats as stronger than the Republicans on a composite index of toughness, principle, and (lack of) weakness — say they are going to vote 91% for Democratic and 9% for Republican candidates in the 2026 U.S. House midterm elections.

The second group, in contrast, would vote Republican by 49 points in November — 74% for Republicans to 26% for Democrats.

Now what we need to know is what proportion of voters sit in each group. Looking at each voters’ individual composite strength scores, and removing people who rated the Democrats and Republicans equally, we get the following composition of voters:

• The Dem-stronger group: 45% of voters

• The Rep-stronger group: 55%

You can use these composition percentages with the above vote loyalty margins to calculate the impact on Democrats’ vote margin like so:

• Dem-stronger group: 0.45 * 82 = D +37

• Rep-stronger group: 0.55 * -49 = D -27

• sum: D+10

And now we can play around with some simple composition math. If we increase the percentage of voters who view the Democrats as the stronger party to 50%, we get these results:

• Dem-stronger group: 0.50 * 82 = D +41

• Rep-stronger group: 0.50 * -49 = D -24

• sum: D+17

Here the predicted Democratic margin among voters with a feeling about either party rises 7 points. We can speculate that making the Democrats 5 points stronger (in relative terms) on this composite index of toughness, principle, and strength would increase their overall vote margin by about 7 points.

However, these shifts won’t happen in isolation. Moderation carries the obvious downside that looking more conciliatory reinforces the perception that Democrats are weak. Their “weak” rating among independents is already at 45% — the trait they lose on most. So the party would be trading marginal gains on a winning metric (already looking less extreme than the GOP) for further losses on a losing one (the party is perceived as very weak!). If you take that into account, then the risk of appearing weak via moderating is clearly not worth it.

An even more robust method of doing this

The back-of-the-envelope math above is useful for illustration, but it relies on splitting voters into binary groups based on trait scores (7 out of 10 for Dem. extremism and 4 out of 10 for strength) and assuming that traits shift in isolation (becoming less extreme has no impact on perceived strength).

A more rigorous approach would be to use a statistical model to estimate the continuous relationship between trait perceptions and vote choice across the full sample, and then simulate what would happen if perceptions shifted. That would also allow me to control for the predictive value of past voting behavior on congressional vote; above we assumed that Harris and Trump voters were equally present in all groups regardless of their partisan trait scores, which is very unlikely to be hold in reality.

So let’s do just that.

---

Warning: the rest of this section is pretty wonky! But I felt it was necessary to show my work. If that’s not your thing, skim the graphs or skip to the next section — the data presented here only reinforces the findings from the section above.

PS: This is the kind of analysis I’d normally reserve for paying subscribers! If you find it useful, consider signing up for more premium analysis:

---

I’m not going to bog you down in the definition of a statistical model or the exact specification I used for this analysis. If you’re curious about more details, you can email me; I’m always happy to talk.

What’s important to know is that I ran a model to predict the likelihood that someone said they would vote for a Democratic candidate in their House district in November based on their 2024 vote history (Harris/Trump/Other/Non-voter) and how they scored each party on the extremism and strength questions.

The chart below shows the relationship between how a voter feels about the Democratic and Republican party on these two variables — strength and extremism — and their probability of voting for the Democratic candidate in November. The left-hand panel, for example, tells us that someone who gives the Democrats a 9-point higher rating than Republicans on whether they think each party is strong is about 95% likely to vote for a Democratic House candidate. On the flip side, someone who gives the Republicans a 10/10 and Democrats a 1/10 for strength is about 90% likely to vote for the GOP candidate.

The right-hand panel of the chart, on the other hand, shows a weaker relationship between midterm vote and whether someone says the Democratic and Republican parties are extreme. Someone at 10/10 Dem and 1/10 Rep only has about an 80% chance of voting D in November.

This tells us that, adjusting for past vote, perceiving a party as strong is a stronger predictor of voting for that party than perceiving the other party as extreme.

To compare the electoral value of getting tough versus moderating, I run two pretend counterfactual elections. In each, I shift every respondent’s perception of each party on either trait by 5 points on the 1-to-10 trait scale — one scenario makes Democrats look 5 points stronger (tougher, more principled, less weak), the other makes them look 5 points more moderate (less extreme relative to Republicans), relative to the GOP. I then re-predict each person’s vote probability under the shifted perception and take the weighted average across the sample.

The results confirm the composition-based estimates above. Being seen as 5 points stronger would boost Democrats’ predicted vote share by roughly 11 percentage points (with a 95% uncertainty interval of +9 to +14). Being seen as 5 points more moderate, on the other hand, would boost it by about 6 percentage points (+5 to +8). The difference — about 5 percentage points — is itself statistically significant, with a 95% uncertainty interval that excludes zero (+2 to +8).

However, we aren’t done yet. We also need to adjust for the fact that Democrats are starting from very different positions on each dimension. In our poll, 52% of voters already see Democrats as the weaker party, while only 34% see them as the more extreme one. Democrats are already winning on moderation relative to Republicans — there just aren’t that many voters left to persuade on that front. On strength, the deficit is much larger.

The uniform 5-point shift doesn’t account for this asymmetry. It moves everyone by the same amount, including voters who already see Democrats favorably on a given trait — voters whose minds don’t need changing. A more policy-relevant question is: what would happen if Democrats simply closed theirdeficit on each dimension? That is, what if every voter who currently sees Democrats as weaker than Republicans were moved to parity — and separately, what if every voter who sees Democrats as more extreme were moved to parity?

This adjusted counterfactual tells an even stronger story for the crowd that wants to strengthen the party’s image. Closing the strength gap — shifting the 52% of voters who see Democrats as weaker to seeing the parties as equal — would boost Democratic vote share by about 6 percentage points (95% CI: +5 to +8). Closing the moderation gap — shifting the smaller 34% who see Democrats as more extreme — would boost it by less than 2 points (+1 to +2). The strength advantage is now nearly four-to-one, because there are simply more minds to change on strength, and those voters are more persuadable to begin with.

We can be confident, in other words, that being seen as stronger is more valuable to Democrats right now than being seen as even less extreme. And remember, Democrats currently have the advantage on perceived extremism, whereas they lag the Republicans on strength/toughness.

In other words, the model-based evidence reinforces the simpler math: an equal improvement in strength perceptions is worth nearly twice as much to Democrats as an equal improvement in moderation. And because these estimates come from the same regression — controlling for partisanship and past vote — they can’t be dismissed as artifacts of partisan sorting. Among voters with the same party ID and the same 2024 vote, those who see Democrats as tougher, more principled, and less weak are substantially more likely to vote Democratic in 2026 than those who see the GOP as extreme.

---

What this means for 2026

The advice that Democrats should focus on increasing their party’s perception as a strong party, instead of as a moderate party, is easier said than done. It’s not like a party can just wave a magic wand to improve how voters see it. Perceptions are earned through sustained, visible action. Democrats will have to earn a reputation as a tough party of strong, principled fighters through repeated impressions with voters.

Equally, while the analysis suggests moderation could earn Democrats extra votes, doing so without looking weak is a real problem for the party. Recent Democratic proposals to moderate on immigration, climate change, and LGBT issues, for example, include language like “don’t raise the salience of immigration” and “don’t talk about climate change”, and that Democrats should “reform and retrain” immigration enforcement — rather, eg, than the more popular message, to decrease ICE funding and refund health care, a position that gets read as strength.

The counterargument to this analysis from the centrist wing of the party is likely to be that moderation here shown to have a positive effect on vote margin, so Democrats should just do that. But this ignores two big problems: First, that voters already see the Dems as the more moderate party, whereas they don’t see it as the stronger one. And second, that messages to moderate are often coded as being weak. That’s not my opinion, that’s just evidenced by current evaluations of the party.

So what are some things they could do? First, they should treat any newsworthy confrontation with the Trump administration as a campaign opportunity, rather than a risk. Democrats may need to rethink their strategy in and out of Congress to orient around toughness. Maybe that means holding up nominees, forcing procedural votes, shutting down the government, or even showing up at protests. Democratic leaders should adopt the mindset of doing whatever it takes to shift perceptions. The strategy of issuing carefully worded statements or rolling over on funding challenges is a big reason the party finds itself where it does today.

---

I started off this article by arguing that the conventional wisdom that moving to the right will solve Democrats’ brand problem is a naive view of their brand problem.

After looking at more detailed data on the parties’ brand perceptions and how that relates to voting, we can conclude that at the very least, focusing on ideological perceptions is an incomplete strategy that leaves lots of voters on the table. And at worst, a moderation-first approach could also backfire badly by making the party appear weak, already its biggest problem compared to the Republicans.

The simple truth of the matter is that Democrats today are not perceived as an ideologically extreme party, despite what certain strategists in Washington, D.C. would have their donors and friends in the media believe. The path to 2026 doesn’t run through the center. It runs through the fight.

For more, see the full topline release and methodology for this month’s poll.

---

Related Articles

A lot of powerful people just don’t realize how unpopular Trump is

Americans voted for Trump, but never supported Trumpism

---

If you’re not a paying subscriber to Strength In Numbers, you’re missing out on premium analysis of American politics and my commentary on what the data really says about the 2026 midterms. Subscribers also help financially support polls like this one, and make interactive data journalism possible. You can subscribe now or make a one-time donation.

If you already subscribe, thank you; and if a subscription is not in the cards for you, the best thing you can do is share this article to help Strength In Numbers grow.