Nearly four out of every five federal court of appeals opinions are unpubli
shed. For more than twenty-five years, judges and scholars have debated the
wisdom and fairness of this body of "secret" law. The debate over unpublis
hed opinions recently intensified when the Eighth Circuit held that the Con
stitution requires courts to give these opinions precedential value. Despit
e continuing controversy over the role of unpublished opinions in the feder
al system, limited empirical evidence exists on the nature of those opinion
s. Working with an especially complete dataset of labor law opinions and us
ing multivariate statistical methods, Professors Merritt and Brudney were a
ble to identify numerous factors associated with publication. Some of those
factors, such as a decision to reverse the agency, track formal publicatio
n rules. Others, such as the number of judges on the panel who graduated fr
om elite law schools or the number with expertise in the disputed subject,
are more surprising. Merritt and Brudney also discovered substantial eviden
ce of partisan disagreement within unpublished opinions, suggesting that th
ose cases are not as routine as publication rules seem to assume. These emp
irical findings should guide constitutional and policy deliberations about
the future of unpublished opinions.