CONTACT: TOM SNEE
300 Plaza Centre One
Iowa City IA 52242
(319) 384-0010; fax (319) 384-0024
Release: Jan. 29, 2003
(Photo: UI law professor Stephanos Bibas)
UI law professor predicts Supreme Court cases for research project
Who can better predict the decisions of the U.S. Supreme
Court, man or machine? A project at Washington University in St. Louis is
trying to find the answer, and UI law professor Stephanos Bibas is taking
The Supreme Court Forecasting Project is the result of a friendly disagreement
between political scientists and legal scholars. The political scientists
have developed a computer model that tries to predict the outcomes of Supreme
Court cases by programming numerous factors based mostly on whether each
justice is conservative or liberal and using that information to determine
how they’re likely to vote on each case.
Legal scholars, however, question the idea that a computer can predict
the behavior of nine human beings based mostly on political ideology and
have assembled a team of 60 scholars from around the country to offer their
predictions of cases heard during the Court’s 2002-03 term. At the
end of the term, the legal scholars’ predictions will be compared to
the model’s predictions.
Bibas is one of those 60 scholars. He thinks predicting Supreme Court decisions
is a difficult art that can be done better by scholars than scientific models
because the justice’s decisions are based on nuanced judicial philosophies,
not just on politics or ideology. As a result, many of the factors that go
into Supreme Court decision-making can’t be taken into account by a
computer model that tries to define each justice just as simply liberal or
“The justices are complicated intellects who can’t be broken
into left or right politically to predict their decisions,” Bibas said. “For
instance, in some cases, Justice Stephen Breyer, a Democrat appointed by
Democratic President Bill Clinton, is more conservative than Justice Antonin
Scalia, a Republican appointed by Republican President Ronald Reagan.”
As a former Supreme Court clerk, Bibas said he also understands each justice’s
quirks. “I know what sets them off, and that’s something you
can’t program for,” he said.
Bibas, a former federal prosecutor who is an expert on criminal procedure,
has been asked to predict the outcome of three cases during the term. So
far, only one of those cases has had a decision handed down, Sattazahn v.
Pennsylvania. The case involved a convicted murderer from Pennsylvania who
received the death penalty and who was appealing his sentence. Argued in
November 2002, the Court ruled in Pennsylvania’s favor and upheld Sattazahn’s
death sentence on a 5-4 vote. Bibas predicted the Court would rule in Pennsylvania’s
favor and uphold the death sentence by a 6-3 vote. The computer model, however,
missed completely by predicting the Court would rule in Sattazahn’s
favor and throw out his death sentence.
“I beat the machine by a mile on that one,” Bibas said.
The Court will hand down its decisions on Bibas’ two other cases
later in the term. The overall results of the forecasting project will be
announced when the Court term ends in June.