Source: (2014) in, David P. Farrington and Joseph Murray, eds, Labeling theory: Empiricl tests. Advances in Criminological Literature. Vol. 18. New Brunswick, New Jersey: Transaction Publishers.

What effect does punishment have on the punished? The best answer to this question is "it depends." Sometimes punishment causes less future crime by the punished. Sometimes it causes more. Sometimes punishment makes no diffrence. That much is well-established. But why these effects vary so greatly across siilar tests with different effects will occure remains largely unpredictable. One reason we know so little is that the most prominent behavioural theories of criminal sanction effects account for only one possible effect of punishment: the one they predict. Deterrence doction explaines why punishment reduces crime by the punished, as the "specific deterrence" hypothesis predicts. Labeling theory explains why sanctions increase crime as the "labeling" hypothesis predicts. Neither deterrence nor labeling theory, however, can offer compelling explanaitons for observed failures of their own predicions. That also leaves them helpless to predict when sanctions will have different effects on individual criminality.