A new Harvard CAPS-Harris survey shows seven in ten registered voters side with Israel over Hamas, while majorities also back photo ID, citizenship checks, and a slower path on AI, findings that cut against media and activist narratives ahead of Election Day.

The online poll of 2,200 registered voters, conducted Sept. 26, 28 with a margin of error of plus or minus 2.1 points, put hard numbers on attitudes that often get filtered through activist framing and cable spin. Just the News reported the results on Oct. 2, weeks before the Nov. 3 midterms.

Support for Israel over Hamas stood at 70% to 30%. That lopsided split is the clearest signal in a survey that also tested views on Iran, the generic congressional ballot, 2028 nominees, election integrity, inflation, Ukraine, China, the United Nations, the press, and artificial intelligence.

Hamas rules Gaza and is a designated terrorist group. The poll did not treat the conflict as a moral wash. Most voters picked a side, and they picked Israel.

An Economist/YouGov measure cited alongside the Harvard numbers showed a much narrower split on broader Israeli-Palestinian sympathies, 26% more with Israelis, 25% more with Palestinians, and 21% equally sympathetic, the tightest gap in that series since 2011. The Harvard question on Israel versus Hamas produced a far clearer majority.

Most voters still distrust Tehran’s motives

On the roughly eight-month U.S. conflict with Iran, the picture was tougher for the administration. Sixty-three percent said the Trump administration’s war effort had not achieved its objectives; 37% said it had.

Skepticism of the Iranian regime ran deeper than scorekeeping on the campaign itself. Sixty-four percent said Tehran was using talks to buy time. Only 36% called the Iranian government sincere. Sixty-one percent believed Iran was becoming more defiant; 39% said the country was moving toward a deal to end the war.

Respondents also largely backed a U.S.-Lebanon plan to disarm Hezbollah by year’s end in return for Israel’s withdrawal and a halt to hostilities. Hezbollah’s military wing has long operated as a parallel armed structure inside Lebanon. Voters, on this evidence, were not eager to leave that status quo untouched.

Those Iran numbers will fuel debate in Washington. They do not erase the Israel-Hamas result, and they do not read as a blank check for Tehran’s stalling tactics.

Generic ballot locks at 50-50 while other averages favor Democrats

On the generic congressional ballot, the Harvard CAPS-Harris sample landed in a dead heat: 50% Democrats, 50% Republicans. Party loyalists broke 96% to 4% for their own side. Independents leaned Democratic, 53% to 47%.

That even split sits well apart from the most recent RealClearPolitics average, which showed Democrats ahead by 8.3 points. If the Harvard reading is closer to the electorate that actually turns out, Republican prospects look sturdier than the broader polling average implies heading into November.

House control often turns on late independents and turnout mechanics more than on cable chatter. A true tie nationally would put a premium on candidate quality, border messaging, and cost-of-living arguments in the districts that decide the majority, the same terrain where conservatives have pressed leverage fights over spending and enforcement priorities, including moves tied to the SAVE America Act in recent funding clashes.

Voters in this survey were not handing either party a blowout. They were drawing a line.

Vance and Harris still lead 2028 fields

Looking past the midterms, Vice President JD Vance remained the top Republican name for 2028 at 39%, down five points from the prior survey. Donald Trump Jr. rose to second at 24%, up six points, even though he has said openly he will not run and backs a Vance candidacy. Secretary of State Marco Rubio stood at 17%, down four points and in third.

On the Democratic side, former Vice President Kamala Harris led at 38%, down four points. California Gov. Gavin Newsom held 17%. New York Rep. Alexandria Ocasio-Cortez climbed to 12%, up three.

Name recognition still dominates these early trial heats. The movement toward Trump Jr., despite his own refusal to run, shows residual brand strength inside the Republican coalition. Harris’s continued lead on the left shows a party that has not yet moved on from its last national ticket.

Election rules draw broad backing for ID and citizenship proof

Election-integrity questions produced some of the survey’s strongest majorities. Fifty-nine percent said it is “difficult” for illegal immigrants to vote; 41% said it is “easy.” Twenty-eight percent said there is a lot of voter fraud; 35% said “some”; 21% said “not too much”; and 16% said there is none.

Photo identification won 82% support, with only 18% opposed. Proof-of-citizenship requirements drew 79% support; 21% were against. Those are not niche conservative preferences. They are supermajority positions in a national sample of registered voters.

Lawful elections require lawful voters. Large majorities in this poll treated basic safeguards as common sense, not as voter suppression. That gap between public opinion and the loudest activist objections remains one of the clearest divides in American politics, and it sits alongside other administration fights over who may remain in the country, including cases where a judge cleared a national-security deportation path.

Economy edges to Republicans as inflation memory splits the public

On the economy, 51% trusted Trump and Republicans more; 49% preferred Democrats. The survey also tested a factual statement that inflation reached 8% a year under President Biden while standing at 3% under Trump. Respondents split 50-50 on whether that statement was true.

The article noted that inflation cleared 9% under Biden and stood at 3.4% in August. Even with prices off their peak, half the sample would not accept the comparison as stated. Memory of the spike still shapes trust, and the party brands remain nearly even on pocketbook leadership.

Ukraine produced another near split. Fifty-two percent said Ukraine would need to concede territory to secure peace; 48% believed Kyiv could win back all of its land. Forty-eight percent backed Trump’s approach to the war; 52% opposed it. Americans are weary of open-ended conflict and unconvinced that maximal aims are realistic.

Xi invitation wins majority approval

President Trump’s invitation of Chinese President Xi Jinping to Washington drew majority support. Sixty-two percent called it the right decision; 37% disagreed. Fifty-nine percent said Trump showed an appropriate level of respect to Xi; 41% thought he was too generous. Fifty-five percent expected the trip to be a success; 45% did not.

Voters can favor tough China policy and still back direct talks. This sample did both, within the limits of the questions asked.

Personnel moves around technology and intelligence often travel with that same mix of hard lines and deal-making, including plans such as the reported choice to name a White House AI czar from the intelligence world.

United Nations earns a bare thumbs-down

The United Nations barely lost the confidence test. Fifty-one percent said the UN was no longer helpful; 49% said it still was. The same 51-49 split appeared on whether Iran should sit on a United Nations human rights commission, with a slim majority opposed.

Fifty-seven percent said the UN focused equally on scrutinizing Iran and Israel; 43% said it focused only on Israel. Even with that plurality giving the institution some credit for balance, overall confidence stayed underwater. A body that seats rights abusers on rights panels will keep forfeiting American patience.

Press questions cut both ways

Media findings were mixed and politically charged. Fifty-six percent said the press covered Trump fairly; 44% said the coverage was unfair. At the same time, a majority said Trump’s efforts undermined the free press rather than respected it. Sixty percent called the decision to ban CNN, MSNOW, and Politico from the White House the wrong move; 40% called it the right one.

Voters can believe coverage is reasonably fair in the aggregate and still bristle at access limits, or the reverse. The poll captured that tension without resolving it. White House credential fights have already drawn loud pushback from prominent anchors, including cases where Jim Acosta attacked press limits as outside the American tradition.

Accountability for reporters and access for the public are not opposites. The numbers show a country still arguing over where one ends and the other begins.

AI use is common; appetite for a slowdown is stronger

Artificial intelligence cut through daily life and long-term fear at the same time. Fifty-eight percent said they use AI personally. Fifty-five percent said AI mostly helped society; 45% said it was mostly damaging. Similar margins expected AI to increase economic growth, yet 60% said it would lead to unemployment for tens of millions.

Sixty-six percent supported a slowdown on AI development; 34% favored a rapid approach. Trump has opposed a slowdown. On this question, the public leaned toward caution even as it reported personal use and expected growth.

That is a governance problem as much as a technology problem. Voters want the gains and fear the displacement. Officials who treat either half of that statement as optional will hear about it.

Confirmation fights and nominee battles will keep testing how far the administration can move its wider personnel agenda, as when a Senate panel advanced a surgeon general pick after sharp hearing exchanges.

Taken together, the Harvard CAPS-Harris results sketch an electorate that is not where cable panels often claim it is. Israel over Hamas was not close. Photo ID and citizenship proof were not close. The generic ballot, in this sample, was a tie rather than a Democratic march. Iran policy drew skepticism about results and about Tehran’s honesty at the same time. The UN barely held half the public’s confidence. AI was already in people’s hands, and most still wanted the brakes tapped.

Majorities still know the difference between a terrorist rulers’ project in Gaza and the democratic partner under fire, and they are done pretending otherwise.