Loughran, T. A., Paternoster, R., Chalfin, A., & Wilson, T. (2016). CAN RATIONAL CHOICE BE CONSIDERED a GENERAL THEORY OF CRIME? EVIDENCE FROM INDIVIDUAL‐LEVEL PANEL DATA. Criminology, 54(1), 86– 112. https://doi.org/10.1111/1745-9125.12097

The authors point out that the Rational Choice Theory has become something of a central theory in social sciences, although criminologists still tend to mainly overlook it. Loughran et al. claim that this has been the case for the reason that the mere understanding of it is simply incorrect and too direct, since the theory is frequently thought of as being able to take into account only financial costs and gains (and sometimes other formal incentives). The other direction of the critique is said to have attacked the rational choice in violent crimes that are said to be committed under circumstances in which an individual does not weigh their alternatives, and the authors hypothesize that this is critique is not fully consistent. The authors also state that the existing literature on the theory focuses primarily on just one factor, which is deterrence, both in forms on the severity of hypothetical sanctions and the probability of them coming true, while their work encompasses all of the variables that determine the individual’s expected utility function.

The authors propose a standard and simple theoretical framework and proceed by testing it empirically via the usage of panel survey data collected from 1354 adolescent respondents over the period of 7 years, namely data on self-reported criminal activity, data on the perceived costs and benefits of crime, and survey data describing each individual’s experiences and circumstances. The sample mainly consists of underprivileged non-White males from urban areas, the average age is 14.9 years and prior to the first interview, on average, they had 3 petitions.

The model being used is Poisson regression, as the dependent variable (number of crimes) being described has a count distribution. The authors also account for various controls, including individual and time period fixed effects and perform several robustness checks. The main results are going to be discussed in the further section, yet here I would like to share some interesting findings that are not worthy of referring to as the “main results”. Interestingly enough, although there is little variation across racial groups, there is substantial variation in sexes. Women are more sensitive to the probability of apprehension, less sensitive to personal rewards of criminal activity and more sensitive to social rewards.

1. The results are homogenous across racial groups, meaning they are valid!
2. Severity of the sanction (in contrast with other determinants of the expected utility function) seems to be statistically insignificant across all types of crime and racial groups, whereas the probability of apprehension is crucial.
3. The results are more or less consistent across crime types.

Подтверждает

Очень забавно читать, как авторы на протяжении всей статьи периодически троллят невидимых социологов, которые не могут понять RCT.

As for the critique, there is a couple of things worth mentioning. First, the most obvious threat to the validity is probably the self-reported nature of the dependent variable. Second, the authors do not control for possibly observed covariates including geography-related ones. Third, there is certain underrepresentation of specific demographics, and the sample size is not very large. Finally, the crucial attribute of working with panel data is its consistency. Here, obviously, people’s perception of risks and rewards could change with time and thus, the measurement errors could be picked up