How Conservative Is the Supreme Court?
The 6–3 Court, measured justice by justice: every vote of the 2020 through 2024 terms, the blocs the votes reveal, and a deep ideology profile of each member of the bench.
Conservative — six justices appointed by Republican presidents, three by Democrats, and the votes divide almost exactly where the appointments predict. But “6–3” hides as much as it says. The six span a dozen points of voting record; one of them leaves a united bloc in nearly a third of his chances; one of the three liberals votes with the Chief Justice more often than against him. The profiles below measure each justice the same way — the record, the bloc, and every case where they left it.
The bench, left to right
Share of each justice’s votes coded liberal by the Supreme Court Database, over the same cases, the 2020 through 2024 terms:
The gap between the benches’ ends is 37.6 points, and the space between the blocs — between Gorsuch at the conservative edge of the middle and Kagan at the liberal edge — is wider than the space inside either. That gap is the 6–3 Court, drawn by its own votes.
Who actually breaks ranks
Each profile counts the cases where a justice’s own bloc was united — at least four of the other five conservatives together, or both liberal seatmates together — and the justice voted the other way. The rates, for the 9 justices profiled so far:
The spectrum runs from Sotomayor, who left a united liberal front in one case in sixteen, to Gorsuch, who left a united conservative bloc in nearly one in three. And the manner of leaving differs as much as the rate: Alito’s departures are all dissents, Thomas goes alone, Gorsuch goes left, Kagan and the Chief Justice go into the majority — Roberts fourteen times out of nineteen, which is how a 6–3 Court keeps producing 5–4 decisions with three liberals in the five. Each profile names every case.
What nine profiles taught us
Each profile below answers a question about one justice. Set side by side they answer a different question, about the Court itself — one no single profile can reach. The figures in this section are court-level aggregates over the 2020 through 2024 terms — the same cases the profiles measure, read from the same file.
1. The Court divides less than its label suggests
Over these five terms the database records votes in 309 cases, and 177 of them drew a dissent at all. Only 46 — 14.9% of the docket — split with one bloc entirely on each side, and in 43 of those the line was the full six against the full three (the other 3 had a justice recused or a seat unfilled). In 74.0% of divided cases at least one justice crossed. The 6–3 label describes who was appointed by whom; it describes about one case in seven. One thing it does describe exactly: when the blocs do split cleanly, the conservatives win — all 46 times, without exception.
The close cases make the point sharper still. There were 38 decisions by a single vote. Of those, 12 were the expected shape — five conservatives, with the three liberals and one conservative in dissent. But 19 were the other shape: all three liberals plus two conservatives. (The remaining 7 were mixed further still — four with two liberals and three conservatives in the majority, three with one liberal and four.) On this Court, the one-vote margin is more often assembled across the blocs than within one.
2. The coalition that decides
Those 19 cross-bloc majorities did not draw their two conservative votes at random. Roberts and Kavanaugh supplied both in 10 of the 19, more than half. The next most frequent supplier is Roberts with Gorsuch, at 4. Widen the lens and the pattern holds: across the 177 divided cases in which both sat, they voted the same way 163 times and parted only 14. That 92.1% agreement is the highest of any pair on the bench, by nearly six points. And when the two of them are on the same side, that side wins: 154 of those 163 cases, 94.5%.
Some of that is arithmetic — two members of a six-justice bloc voting together are most of the way to five. What is not arithmetic is the direction of travel. When the Chief Justice leaves his bloc he lands in the majority fourteen times out of nineteen, and the two of them are tied for the lowest dissent rate on the bench: each was in the minority in exactly 16 of the 177 divided cases, 9.0%, where the next-lowest is 15.2%. The pair does not drift away from the Court’s center; it is the Court’s center, and a case that persuades the two of them has found five votes from either direction. That is the arithmetic behind the phrase advocates use about writing for the middle.
3. Method, not just ideology — but the data is stubborn about it
The most attractive explanation for the breaks is that they are methodological rather than political: a textualist follows the words of a statute wherever they lead, an originalist follows the history, and the resulting vote gets coded liberal or conservative by accident. The case notes on the profiles support that reading again and again — Gorsuch on Indian treaties, Jackson on statutory text, Barrett on the scope of a jurisdictional rule.
But the tempting version of that claim does not survive a count. If breaks were mostly method, they should cluster in the dry cases — economic activity, judicial power, federalism, taxation, unions and the smaller technical categories — and thin out in the cases about criminal procedure, civil rights and the First Amendment. They do not. Those dry areas are 44.1% of all divided cases and 48.1% of all breaks. That difference is noise. Justices leave their blocs in rights cases at very nearly the rate they leave them anywhere else.
What the count does show is direction. When a conservative leaves a united bloc, the vote is coded liberal 70 times out of 116 (60.3%); when a liberal leaves, it is coded conservative 19 times out of 35 (54.3%). Read that with care: leaving a united bloc means, almost by construction, voting against the side that bloc usually takes, so the asymmetry is partly built into the definition. What is not built in is who does it and how often — and that varies nearly fivefold across the nine, from Sotomayor to Gorsuch.
4. Standing alone is a second axis, not a further step right or left
A solo dissent is the purest measure of a justice’s willingness to be outvoted eight to one and write anyway. Over the same five terms, counting only the cases each justice actually sat, and ordered by rate rather than by count — Jackson has sat for three of the five terms, so her two are a higher rate than Alito’s three. Bear in mind how rarely the chance arises at all: only 31 of the 177 divided cases were decided eight to one.
The order has almost nothing to do with the order on the first chart of this page. The two justices at zero, Elena Kagan and Amy Coney Barrett, sit in opposite blocs. The three most willing to stand alone — Thomas, Gorsuch, Sotomayor — are two conservatives and the Court’s most liberal member. Correlate the liberal share against the solo rate across the nine and you get −0.14 — and +0.22 if you drop Thomas. A relationship whose sign depends on one justice is not a relationship. Temperament and ideology are separate axes, and the profiles read differently once you stop expecting them to be the same one.
5. What the votes cannot show
Everything on these pages is downstream of a number in a column. A vote is the residue of a process that leaves no public record: the circulated draft, the clerk’s bench memo, the fifth vote offered on condition that a paragraph comes out, the concurrence withdrawn when the majority narrows. A justice who never dissents alone may be winning those arguments in chambers or may simply dislike writing separately, and the vote record cannot tell the two apart. Everything historians know about that process comes from papers and memoirs opened decades later — which is why the accounts of the Burger and Rehnquist Courts read so differently from anything we can honestly say about this one yet.
So take the aggregates for what they are: a floor of things that are true, on which the harder story will one day be built. They are enough to retire a few confident claims — that the six vote as one, that the breaks are confined to technicalities, that a willingness to dissent alone marks a justice as extreme. Nine profiles were needed to see it. Every number here can be traced to a named case in one of them.
The deep dives
One profile per justice, all built the same way: the career record ranked against every justice since 1946, agreement with each sitting colleague, every break from a united bloc with a verified note on its ground, and a documented method. All nine of the sitting justices now have one.
Only Rehnquist’s record is more conservative — and when the bloc holds without him, he is usually standing to its right, alone.
The lowest liberal share on the current bench, the tightest bond to the bloc, and fourteen breaks in five terms — every one a dissent.
The most liberal justice by nearly every ruler — and the most loyal to her bloc, leaving a united front just nine times in five terms.
The only justice who votes with three opposite-bloc colleagues in most divided cases — and, in the data’s fifteen terms, never once dissented alone.
The highest break rate measured on any of these pages — 38 departures in five terms, three of every four landing beside the liberals.
The Court’s most frequent dissenter breaks both ways: twice alone to the left of her own bloc on labor, five times into the majority on statutory text.
The closest justice to the Court’s center and the least likely to leave her bloc — and in seven of her nine departures she wrote the opinion herself.
Almost never in dissent, and half the Court’s closest pairing — when he leaves the six he usually takes the Chief Justice and the liberals and makes a majority.
Fourteen of his nineteen departures from his own bloc are majorities — the Court’s most frequent builder of cross-bloc coalitions, and the only justice with no cliff on his chart.
Questions
Is the Supreme Court conservative?
Yes, by composition and by votes. Six of the nine justices are Republican appointees, and over the 2020 through 2024 terms — the life of the current alignment — every one of the six cast liberal votes (as the Supreme Court Database codes them) less than half the time, from Gorsuch's 42.4% down to Alito's 30.4%, while the three Democratic appointees sit between 65.3% and 68.0%.
Who is the most conservative justice on the Supreme Court?
It depends on the ruler, and the two candidates are Thomas and Alito. Over the current Court's own five terms Alito's 30.4% liberal share is the lowest on the bench; across whole careers Thomas's record (32.2% liberal since 1991) sits below Alito's (34.3% since 2006), and only Rehnquist's is lower among the 33 justices rankable since 1946. The scholarly Martin–Quinn score has the two trading first place by term.
Who is the most liberal justice on the Supreme Court?
Sotomayor, by nearly every measure: the highest liberal share over the current window (68.0%), the highest career rank of any sitting justice (7th most liberal of the 33 rankable since 1946), and the furthest from the Court's center on the Martin–Quinn score — though Jackson and Kagan sit within about two points of her, and Jackson dissents more often than anyone.
Which justice is the least predictable?
Gorsuch, and it is not close. He left a near-united conservative bloc 38 times in 125 chances (30.4%) — the highest rate of any justice measured — with three of every four departures landing beside the liberals, most of all in Indian-law and criminal-text cases. His highest agreement with any colleague is lower than anyone else's.
Do the justices always vote in blocs?
Mostly, but measurably not always. In divided cases the six Republican appointees agree with each other far more than with the three Democratic appointees, and vice versa — the agreement charts on every deep dive show the same cliff. But each justice leaves a united bloc at a measured rate, from Sotomayor's 6.3% of chances up to Gorsuch's 30.4%, and the deep dives list every one of those cases by name.
How this page is made
Every figure derives from the same payload the deep dives read, computed from the Supreme Court Database (Spaeth, Epstein, Nelson, Martin, et al.), 2025 Release 01, over the 2020 through 2024 terms (how the voting figures are made). “Liberal” and “conservative” are the database’s terms of art; each deep dive flags the places its coding strains. The break definitions — four-of-five for the six-member bloc, both seatmates for the three — are stated on every profile. Career rankings and the same-case baselines live on The Court from Left to Right; pairwise agreement for every Court since 1946 on Who Voted With Whom.