USING THE CORRECT STATISTICAL TEST FOR THE EQUALITY OF REGRESSION COEFFICIENTS

Criminologists are often interested in examining interactive effects within a regression context. For example, “holding other relevant factors constant, is the effect of delinquent peers on one's own delinquent conduct the same for males and females?” or “is the effect of a given treatment program comparable between first-time and repeat offenders?” A frequent strategy in examining such interactive effects is to test for the difference between two regression coefficients across independent samples. That is, does b₁= b₂? Traditionally, criminologists have employed a t or z test for the difference between slopes in making these coefficient comparisons. While there is considerable consensus as to the appropriateness of this strategy, there has been some confusion in the criminological literature as to the correct estimator of the standard error of the difference, the standard deviation of the sampling distribution of coefficient differences, in the t or z formula. Criminologists have employed two different estimators of this standard deviation in their empirical work. In this note, we point out that one of these estimators is correct while the other is incorrect. The incorrect estimator biases one's hypothesis test in favor of rejecting the null hypothesis that b₁= b₂. Unfortunately, the use of this incorrect estimator of the standard error of the difference has been fairly widespread in criminology. We provide the formula for the correct statistical test and illustrate with two examples from the literature how the biased estimator can lead to incorrect conclusions.     

GENDER BIAS IN THE DISPOSITION OF JUVENILE COURT REFERRALS: THE EFFECTS OF TIME AND LOCATION*

Theory based on sex role traditionalism predicts a more punishing decision for female than for male offenders, while theory based on chivalry (paternalism) predicts greater leniency by the courts for female offenders. This paper tests these two models using a large sample (36,680) of juvenile court referrals in metropolitan, urban, and rural locations spanning a nine-year period. Nonparametric analysis of covariance is used to control for differences in offense, previous contact with the court system, and other background variables. Evidence of gender bias in dispositions was found. The patterns of bias across time, location, offense committed, and previous referral to the court system supports the persistence of chivalry and a decline in sex role traditionalism in court decisions. Greater punishment for girls than boys was found only for repeat offenders committing more serious offenses. Even in those cases, girls were more likely to be taken out of the home environment by the courts through a custody transfer while boys were more likely to be sent to a lock-up facility.     

On the Concept and the Nature of Law

Abstract. The central argument of this article turns on the dual‐nature thesis. This thesis sets out the claim that law necessarily comprises both a real or factual dimension and an ideal or critical dimension. The dual‐nature thesis is incompatible with both exclusive legal positivism and inclusive legal positivism. It is also incompatible with variants of non‐positivism according to which legal validity is lost in all cases of moral defect or demerit (exclusive legal non‐positivism) or, alternatively, is affected in no way at all by moral defects or demerits (super‐inclusive legal non‐positivism). The dual nature of law is expressed, on the one hand, by the Radbruch formula, which says that extreme injustice is not law, and, on the other, by the correctness argument, which says that law's claim to correctness necessarily includes a claim to moral correctness. Thus, what the law is depends not only on social facts, but also on what the law ought to be.     

Congressional Performance,Incumbent Behavior,and Voting in Senate Elections

Conventional wisdom suggests that individual members of Congress have no real incentive to act in ways that might improve public evaluations of their collective body. In particular, the literature provides no clear evidence that public evaluations of Congress affect individual races for Congress, and little reason to expect that voters would hold specific individuals responsible for the institution's performance. We suggest that this conventional wisdom is incorrect. Using multiple state‐level exit polls of Senate voting conducted by Voter News Service in 1996 and 1998, we arrive at two key findings. First, we find that evaluations of Congress do have a significant effect on voting within individual U.S. Senate races across a wide variety of electoral contexts. Second, we find that punishments or rewards for congressional performance are not distributed equally across all members, or even across members of a particular party. Instead, we find that the degree to which citizens hold a senator accountable for congressional performance is significantly influenced by that senator's actual level of support for the majority party in Congress, as demonstrated on party votes.