acknowledged is the intimate connection between one’s sense of justice in an
instant case and one’s general ideological predispositions.
An attitude is a predisposition, and a relativelyWxed one at that (but see Martin
and Quinn 2002 ). Behavior, on the other hand, is characterized by a distribution of
discrete decisions. Attitudes, if they are useful, should predict the central tendency
in such distributions, and they do. However, it would be unreasonable to expect
attitudes to predict every single individual decision. It may be, for instance, that the
votes of voters with Democratic attitudes are well predicted by the attitudes in
general, even if a speciWc voting decision may deviate from the attitudes
(e.g. ‘‘Reagan Democrats’’). We know that in general attitudes are moderate to
strong predictors of behavior (e.g. Kraus 1995 ), but also that attitudes never
provide a complete explanation of actual behaviors. What must be added to our
analyses is an understanding of the factors that might deXect a decision away from
the predisposition of the decision-maker. Thus, our models of decision-making
require an additional layer of complexity.
3 Normative Constraints on
Decision-making: What Judges Think
They Ought to Do
.........................................................................................................................................................................................
Attitudinalists are often challenged by those who assert that judges are frequently
strongly constrained by stare decisisand precedent. The argument has more
plausibility at the trial court level, where judges are often faced with routine
decisions requiring the application of existing law to instant disputes. At the
appellate court level it appears that law rarely dictates outcomes. Segal and Spaeth
( 1996 ), for instance, show that most justices of the US Supreme Court do not put
aside their policy preferences in order to defer to existing precedents. Others
disagree, however (e.g. Knight and Epstein 1996 ; Spriggs and Hansford 2001 ), so
this is an issue requiring further inquiry.
Undoubtedly, the answer is that laws (constitutions, statutes, and precedents)
vary in the degree to which discretion is aVorded to judges. And since fact patterns
rarely reproduce themselves precisely, some discretion is virtually always available
to judges to determine what body of law is relevant to an instant case. 3 Perhaps all
3 When a new policy is set, it changes the relative weights assigned to the fact-based aspects of the
cases. For example, Richards and Kritzer ( 2002 ) show that the factors related to decisions in free
518 james l. gibson