In general terms, âdesistanceâ is used consistently to define the cessation, suspension, pause, or remittal of a certain action; in the field of criminology, however, definitions of desistance are as varied as their authors â and they are rarely clear, precise, or straightforward. Even in those articles, which do manage to provide some clarity, it is almost always preceded by an authorâs statement about the difficulty of arriving at their definition. It has been noted that desistance definitions are so idiosyncratic that, in some studies, it is possible to find ones that are made only to match the research design, the available data, and other practical concerns (Laub & Sampson, 2001; Stouthamer-Loeber et al., 2004).
Desistance and the criminal career approach
The criminal career approach divides criminal trajectories into three stages: emergence (onset), pattering (duration), and abandonment (desistance) (Soothill, Fitzpatrick, & Francis, 2009; Sullivan & Piquero, 2016). Although a key component in criminal career research, desistance has received less attention than persistence and onset (Bushway, Thornberry, & Kohn, 2003; Laub & Sampson, 2001; Piquero, 2004; Shover, 2004).
This has been partially justified by the fact that, from a public safety approach, it is more worthwhile to focus efforts on those prolific and chronic criminals who generate great costs for society than it is on those who have already stopped committing crimes. Furthermore, the Juvenile Justice Systems in many countries have been continually ineffective in attempts to reduce crime, thereby enforcing the idea that prevention is better than cure. Thus, efforts to prevent crime have been more oriented to understanding the factors that provoke its onset (Van Der Put & Dekovic, 2011; Farrington, 2007; Farrington & Welsh, 2007; LĂŽsel & Beelmann, 2006).
Yet putting the focus merely on the causes of crime may limit criminal rehabilitation and indeed may accentuate the problematic issues that rehabilitation aims to diminish (McNeil, 2004). By identifying the process of crime abandonment, on the other hand, desistance research can open the way to thinking more prospectively about how and why people change â thus paving the way for a more successful method of crime prevention in the long term (Maruna, 2001).
Despite the limited attention that the criminal career approach has given to desistance, it has made great contributions to the field by providing relevant information about the patterns of crime along the life course (see Basto-Pereira & Farrington, 2020; Farrington et al., 2014; Piquero, Farrington, & Blusmstein, 2003, 2007). For instance, with data from official criminal records and self-reported crime, we now know that offending most typically starts between the ages of 8 and 14 and stops between the ages of 20 and 29 (Piquero, 2011) â though other studies have shown that desistance peaks earlier, around 18 years of age (see Bowles & Pradiptyo, 2005; Stouthamer-Loeber et al., 2004).
Nevertheless, it is worth asking whether the first and the last self-reported or recorded crime (including arrest or conviction) really represents the starting and the ending of an individualâs criminal career. Does a criminal career stop after committing the last offence, or does it finish when one decides to stop? How long must an individual be away from crime to qualify as a desister? These questions have raised a debate regarding the understanding of desistance: whether the abandoning of a criminal career should be viewed as a fixed end point in offendersâ lives, or whether it should rather be seen as a process that supports the termination of offending (Sampson & Laub, 2003).
Beyond the termination point: desistance as the process of moving away from crime
In almost all the studies carried out before the nineties (with the notable exception of Meisenhelder, 1977), desistance had always been defined in a static way as denoting the end of the criminal career (see Cusson & Pinsonneault, 1986; Farrington, 1986; Farrington & Hawkins, 1991; Glueck & Glueck, 1943). This trend has continued in more recent years, too: Shover (1996:121), in a study in which he explored criminal career changes among more than 50 persistent thieves in the United States, defined desistance as âthe voluntary termination of serious criminal participationâ; similarly, Stouthamer-Loeber et al. (2004:897) offered the following definition in their study of the factors triggering and inhibiting desistance: âdesistance is defined here as an individualâs stopping the commission of crimeâ.
The first limitation of studying desistance as having a static end point arises when we attempt to determine precisely how and when an individualâs criminal career has stopped. In strict terms, the study of desistance in this sense would need to follow individuals until their death to establish whether or not they had truly stopped committing crimes. Since this is not feasible, most studies have operationalised desistance as the presence of a certain period of time without any criminal offence. Nevertheless, this operationalisation implies at least three key methodological issues that need to be considered.
First, it is difficult to ascertain a sufficient amount of follow-up time, because some offenders show intermittent crime-free gaps throughout their criminal careers (Kazemian, 2007). This is not a new finding â indeed, several decades ago, Glaser (1969) argued that offenders follow a zigzag path in which they alternate between crime and non-crime along their lives. Even prior to that, West (1963:34) had referred to intermittency as âinterludes of honestyâ, arguing that these periods coincided with circumstances (material or relationship-based) that provided offenders with shelter or support. More recently, several further studies gave support to the fact that individuals can experience periods of temporary desistance (see Barnett et al., 1989; Elliott, Huizinga, & Menard, 1989; Farrington, Barnett, & Blumstein; 2020; Horney et al., 1995), with advanced statistical techniques being developed to assess intermittency in criminal careers (Nagin & Land, 1993).
Second, the length of follow-up is also relevant to determine correctly the age from which desistance takes place. For example, Le Blanc and Fréchete (1989) found, when following individuals up to 25 years of age, that the average age of those who had achieved desistance was 19.9 years. However, when following individuals up to the age of 40, the average age at which the last conviction took place was 31 years. This occurred because individuals who were followed until the age 40 had more exposure time to commit crimes. More recently, Farrington (2019) with data from the Cambridge Study in Delinquent Development observed that the mean age of desistance tended to increase, as the duration of the criminal career increased.
Time of exposure can also be a problem when scholars define the cut-off time in their studies. In simple terms, when operationalising desistance as a period free of crime, scholars normally divide their samples into two time periods: to be defined as desisters, individuals who have offended in period one must not have offended in period two. The difficulty is that the age of individuals in both periods differs considerably between studies, making the timing and prevalence of desistance difficult to define and compare (Bushway, Thornberry, & Krohn, 2003).
The third key issue is the type of data available: desistance rates can vary markedly within a sample when considering official criminal records versus self-reported crime (Brame et al., 2004; Farrington, 1989; Farrington, Auty, Coid, & Turner, 2013; Le Blanc & FrĂ©chete, 1989; Sampson & Laub, 2003; Stouthamer-Loeber et al., 2004). A study in the mid-20th century found that many instances of self-reported crime were not detected by the criminal justice system (see Short & Nye, 1957) â a discrepancy which occurs variously because some crimes are not reported, the perpetrator is not found by the police, and/or the case is not officially processed and is subsequently diverted from the criminal justice system (Loeber & Le Blanc, 1990; Piquero, Hawkins, & Kazemian, 2012). The issue is starkly highlighted by Elliott (1994), who estimates that, per 100 self-reported violent offences, merely two arrests are made. Furthermore, in a study that analysed data from the Seattle Social Development Project,1 Farrington et al. (2003) observed that while 85.9% of the juveniles admitted committing at least one offence, only 34% were referred to court for at least one offence. In a more recent article, Farrington, Auty, Coid, and Turner (2013) using data from the Cambridge Study in Delinquent Behaviour observed that on average there were 38 self-reported offences per conviction among the participants of the study, and that this ratio decreased with age.
One would, therefore, expect to find that desistance would occur later in life when assessed using self-reported rather than official data (Farrington, 2007), and that official records may over-represent desistance rates compared to self-reported crime (Farrington, 1989; LeBlanc & FrĂ©chette, 1989; Stouthamer-Loeber et al., 2004). Indeed, Massoglia and Uggen (2007), in a study drawing on data from the Youth Development Study,2 introduced different conceptualisations of desistance through comparison of official records with what they defined as âbehavioural desistanceâ.3 When using official records, 85% of the sample had desisted at the age of 29â30, but this figure dropped to 65% when measuring behavioural desistance.
A further limitation of defining desistance as having a definite end point arises when we try to determine how that termination occurs. Does it happen when an individual commits their last crime? As mentioned previously, it is difficult (if not impossible) to establish when an individual commits their final offence â even if they are followed until their death.
Some scholars have argued that, instead of focussing on the last offence, the individualâs decision to stop committing crimes should be considered as the moment at which desistance occurs (see Cusson & Pinson...
