Don’t Judge the Supreme Court by Its Polls

As it gets ready to start a new term, the Supreme Court has a polling problem. Gallup reported in July that only 33 percent of Americans approve of the job it’s doing, a record low in the 26-year poll. Marquette Law School’s latest survey is less grim but still shows the Court underwater, with 47 percent approving and 53 percent disapproving.

For advocates of Supreme Court “reform,” the temptation is obvious: low approval means that the institution has lost legitimacy and needs structural repair.

But a closer look at Gallup’s polling numbers knocks down that argument. The decline since last year was driven almost entirely by Republicans, whose approval of the Court plunged from 79 percent in September 2025 to 58 percent now. Democratic approval remained at 12 percent and independent approval at 35 percent, both near their historic lows. Gallup says the partisan gap in Court approval averaged 18 points from 2000 through 2010 but has averaged 32 points since.

Marquette finds the same phenomenon from another angle: 74 percent of Republicans approve of the Court, compared with 35 percent of independents and 22 percent of Democrats. Among very conservative Republicans, approval reaches 84 percent; among very liberal Democrats, it’s 5 percent.

That’s an unhealthy picture, but it looks more like polarization than a broader crisis of legitimacy.

Moreover, Americans have lost faith in nearly everything. Gallup’s separate June survey puts average confidence in the 14 major institutions it has tracked since 1993 at 27 percent, just one point above the all-time low. The Supreme Court itself stands at 27 percent—the national average—compared with 17 percent for the criminal-justice system, 14 percent for television news, and 9 percent for Congress. Indeed, approval scores for 12 of these core 14 institutions are at or near their historical lows—and the only governmental institutions more trusted than the Court are the military and police.

There’s another reason to resist treating approval as a referendum on the Court’s work: Americans often like that work. Marquette asked about 13 decisions from last term. Majorities favored 11, and eight drew at least 60 percent support. Sixty-six percent backed the Court’s rejection of President Trump’s tariffs under the International Emergency Economic Powers Act; 63 percent supported its birthright-citizenship ruling; 62 percent approved allowing states to count mail ballots that arrive after Election Day; and 70 percent supported the ruling that prevented Trump from immediately removing Federal Reserve Governor Lisa Cook. The two decisions that drew majority opposition—presidential removal of independent-agency heads and political parties’ coordinated campaign spending—were toss-ups within the margin of error, 51–49 and 53–47.

In any case, polls are a bad test of judicial legitimacy, because judicial review has an inescapably counter-majoritarian function. For example, in 2024, Marquette found 79 percent support for Students for Fair Admissions v. Harvard, which ended the use of race in college admissions; by contrast, Dobbs v. Jackson Women’s Health Organization, which overturned Roe v. Wade, remains unpopular. The Court needs its independence to decide legal questions even when the political chips fall badly.

The actual decisions also put paid to the idea that this is simply a “Trump Court.” Last term, the justices expanded presidential removal power, allowed Trump to wind down Temporary Protected Status for Haitians and Syrians, set aside campaign-finance restrictions challenged by the Republican National Committee, and invalidated Hawaii’s gun restrictions. Yet the same Court rejected Trump’s tariffs, kept the National Guard from being deployed in Chicago, allowed late-arriving ballots to be counted, blocked Cook’s removal, and rejected Trump’s effort to narrow birthright citizenship.

There’s a jurisprudential pattern here. The Court’s majority commonly uses originalist and textualist methods, with a small-c conservative concern for separation of powers and federalism. That approach strengthens presidential control over the executive branch—the so-called unitary-executive theory—but it still leaves the president constrained by statutes, Congress, courts, and the Constitution. Trump v. Slaughter vindicated presidential supervision of executive officers, while the tariff case held that Congress hadn’t given the president sweeping economic power. The distinction is legal rather than partisan.

I drew the same distinction between political popularity and legal legitimacy in my book Supreme Disorder. The Court needs public confidence, but chasing it is a dangerous way to get it. As I wrote, “the only measure of the Court’s legitimacy that matters” is “the extent to which it gets the law right and applies it correctly.” Justices who bend toward the latest poll compromise the reason we have an independent judiciary in the first place.

Partisans naturally remember their losses. Progressive critics see Dobbs, affirmative action, and presidential immunity. MAGA critics see tariffs, birthright citizenship, and mail-in ballots. Each side turns disappointment with particular cases into an institutional indictment.

But court-packing, term limits, and other proposals should stand or fall on their own merits. A dip in approval—especially one produced by historically polarized voters during an era of collapsing trust generally—provides little reason to redesign the Constitution. The proof is in the Court’s actual work: it regularly frustrates both sides, reaches conclusions the public generally supports, and keeps returning to the same old-fashioned question: What does the law require?