We analyze thousands of trials from a substantial fraction of the nation's most populous counties as well as a smaller sample of less populous counties. Evidence from four major Civil Justice Survey data sets spanning more than a decade establishes that: (1) compensatory awards are strongly associated with punitive awards and (2) the punitive-compensatory relation has not materially changed over time. But (3) 2005 data suggest, for the first time, systematic differences between judges and juries in the punitive-compensatory relation. Despite claims that the Supreme Court's State Farm decision changed the punitive-compensatory relation, we present evidence that the 2005 shift is not attributable to the State Farm case or to other possibly relevant likely factors such as the relative flow of personal injury cases to judges and juries, inclusion of 110 small counties in the 2005 data, or changes in the 2005 data coding. The judge-jury difference more likely turns on unobserved factors driving the selection of cases for adjudication before judges and jurors.