Pogarsky, G. & Piquero, A. R. (2003). Can punishment encourage offending? Investigating the "resetting" effect." Journal of Research in Crime and Delinquency, 40, 95-120.
It can be argued that punishment has become a widely utilized behaviour modification tool intended to deter an individual from repeating an action that is deemed inappropriate (Pogarsky & Piquero, 2003). According to the concept of specific deterrence, punished individuals should be discouraged from re-offending. However, several researchers have proposed that punishment may in fact reinforce behaviour rather than reduce it (Paternoster & Piquero 1995; Piquero & Paternoster 1998; Piquero & Pogarsky 2002; Sherman 1993). It has been suggested one of the foremost factors responsible for the inability of punishment to act as a deterrent is insufficient certainty of sanction. In other words, it appears that an individual is significantly less likely to alter his or her behaviour if he or she estimates that the associated punishment is unlikely to occur. This is important, given Pogarsky and Piqueros (2003) contention of the "positive punishment effect," the notion that offenders who have been punished in the past believe that the certainty of future punishment is lower than individuals who have not been punished. The authors investigate two competing theories to explain this intriguing finding: selection and resetting.
Selection theory revolves around the belief that punishment is most likely to be inflicted upon individuals who offend often (i.e. those at highest risk). For example, if five of thirty students throw rocks continuously at other students during recess, those five students are more likely to be identified and punished. Due to the repetitive antagonizing by the offending students, there is an increased likelihood they will be reported to, or noticed by, an authority figure and subsequently punished. Therefore, in response to the high volume of offences (i.e., the number of times the five students threw rocks) and comparatively low number of punishments (i.e., the number of times an authority figure has punished one of the five students) it is reasonable to expect that the five students may believe there is a decreased chance of being punished. If selection theory is the basis for an individual's sanction certainty estimate, individuals who are more likely to offend in the future are those who have been previously punished. In addition, punished individuals will be more likely to possess a greater number of individual and situational risk factors that influence offending than less or non-punished individuals. Thus, it is suggested that individuals who are of high risk-to-offend, should have lower sanction certainty estimates, irrespective of punishment history, than those in the low risk-to-offend group.
In contrast, the resetting theory suggests that, following punishment, an offender may change their sanction certainty estimate based on a decision-making heuristic called "gambler's fallacy." Gambler's fallacy suggests that individuals tend to believe seemingly rare occurrences are unlikely to be immediately repeated, irrespective of statistical reality. For example, following several black balls on a roulette wheel many gamblers will place their bet on red, believing the chance of another black ball to be less than 50%. Gamblers believe that, because the odds are 50/50, the wheel must even out the results and come up red. Thus, if resetting theory is the basis for an individual's sanction certainty estimates, the gambler's fallacy decision-making bias should be the underlying reason that punishment is encouraging participants to re-offend. In addition, it is suggested that risk status should not affect the sanction certainty estimate as the occurrence of punishment should be the true determinant of a lower perceived certainty of punishment
To further investigate the two aforementioned theories, the authors surveyed 256 participants from a public university. …