Examining General Deterrence Using Data from the National Football League

Abstract

To date, research studies have found only mild support for classic deterrence theory with the greatest support for increased certainty, little support for increased severity, and scant research on the effect of increased celerity. Much of this prior literature has used scenario-based data, relied heavily on student samples, and explored rule breaking behavior over relatively short time periods. Finally, the slow pace of punishment within the criminal justice system potentially reduces any existing deterrent effect of the certainty and severity of punishment. This dissertation seeks to address these limitations of prior deterrence studies by using 13 years of data (2000-2012) from the National Football League consisting of rule breakers who are punished with penalties and monetary fines almost immediately upon discovery of the infraction. The main question driving this research is whether there is evidence of general deterrence. Specifically, this dissertation seeks to determine whether prior punishment reduces current rule-breaking behavior. To address this question, this research explores the effect of on-field penalties and post-game fines on behavior within the National Football League at both the league and team levels. The dataset has several rare characteristics including: large variety and detail in the types of punishment administered, an opportunity to directly observe the effect of punishment, the near immediate imposition of punishment, and the transmission of almost perfect information about punishment.

The primary finding is that there is no evidence of general deterrence in the National Football League, independent of control variables. Specifically, penalties and fines do not appear to prevent future rule breaking behavior. In general, when controlling for particular seasons, opponents, or the record of a team, the effects of penalties and fines loose significance and approach zero. The different controls for seasons, opponents, or record are fairly consistent in their statistical significance for all penalties and violent penalties, although it appears that violent penalties vary less according to these outside factors than all types of penalties. In sum, this dissertation finds no evidence that punishment affects future rule-breaking behavior at either the team or league level and thus, does not provide support for general deterrence.