Docket Report: Berk v. Choy

Update 11/1/25: The updated model (experience-3) continues to predict reverse (65 percent), though there is a reasonable amount of uncertainty (ninety percent band of 49-75 percent). Four clear reverse votes (Jackson, Kagan, Sotomayor, Gorusch), two clear affirm votes (Alito, Thomas), and three remaining close votes.

Update 10/22/25: The updated model (experience-2) continues to predict reversal, but now sees this as less of a close case (probability of reversal 82 percent after argument). The model continues to see a conservative bloc as the most likely to break off and vote to affirm.

Initial post: Berk v. Choy is a civil procedure case that centers on whether a state law that requires medical malpractice claims to be dismissed unless they contain an “affidavit of merit” from an expert can also apply in federal court. This requirement appears to conflict with the Federal Rules of Civil Procedure, which do not require this kind of statement. Often, this question would not come up because there would be no federal court involved—but here, the plaintiff and the defendant come from different states, so there is federal jurisdiction by virtue of diversity. Thus, can the state law be imported into the federal courts? This is a wonky case that turns, in part, on whether this affidavit requirement should be understood as “substantive” or “procedural.” If determined to be procedural, the normal federal procedural rules would apply under the long-standing Erie doctrine. If determined to be substantive, the state rules might be imported to federal court. The lower court allowed the state rule to pass to federal court, seeing it as substantive. As I tell my students, the substance-procedure barrier is highly permeable. And this case is no exception. The state rule carries out substantive values, seeking to control the number of medical malpractice suits by filtering claims through medical experts. But it is obviously also procedural. Both before and after argument, the model thought this would be a close case. The case-level prediction is reverse with a little over fifty percent probability. Argument actually shifted predicted votes very little. As seen below, most justices vote to reverse somewhere with roughly fifty percent probability. There is a modest ideology gradient: liberals appear more likely to reverse, thus allowing the plaintiff bar easier access to federal court in cases such as this.