Problems of reliability and validity in criminal justice research

Reliability and validity are critical standards of scientific inquiry. This analysis directs attention to problems of reliability in data collection and to problems of validity in measuring two key criminal justice variables: the offense and the sentence. The discussion and empirical analysis indicate that reliability of official information sources in criminal justice cannot be assumed. Furthermore, there is evidence to suggest that the criterion of validity is even more elusive. This elusiveness is demonstrated in the operational definition of criminal offense, an important predictor variable, and in attemps to operationally define the criminal sanction, i.e., the disposition of charges against convicted defendants. These considerations suggest that research in the sentencing area may not be directed to the subject at hand and that research results should be interpreted with “scientific” caution.