Concept of Self-Control

As Gottfredson and Hirschi use the term, self-control refers to the ability to forego immediate or near-term pleasures that have some negative consequences and to the ability to act in favor of longer-term interests. It is related to concepts such as self-regulation and impulsivity in psychology (Baumeister & Heatherton, 1996; Moffitt, Poulton, & Caspi, 2013), time discounting and skill formation in economics (Heckman, 2006), and social control in sociology (Hirschi, 1969). For self-control theory, important negative consequences can include physical harm, legal sanctions, removal from educational institutions, or disappointment or disapproval of family, teachers, and friends. Gottfredson and Hirschi argue that most crime and delinquency can be seen as the pursuit of relatively immediate and easy benefits or immediate and momentary pleasures, and therefore acts of delinquency and crime tend to be disproportionately undertaken by individuals with relatively low self-control.

Self-control is not regarded as either a predisposition to crime or a personality trait for crime and delinquency. Rather, self-control is understood as an inclination to focus on the short term rather than the long term, on immediate gratification of needs, and on wants and desires (whatever they may be), and not on the longer-term negative consequences of behavior. Self- control theory is a choice theory rather than a deterministic one. In fact, self-control is not a concept specifically focused on crime—low self-control does not require delinquency and crime, nor does it compel it. High self-control can be described as part of social capital or social advantage, since it helps to create successful outcomes for many life experiences and beneficial results from social institutions, including education, the labor market, and interpersonal relations (such as marriage).

Control Theories

Self-control theory belongs to a general class of crime theories, which include social control theory (Hirschi, 1969) and deterrence theory, each of which builds on the assumptions of the classical school in criminology (Beccaria, 1764; Bentham, 1789). According to these theories, people tend to act in accordance with the principles of rationality and self-interest (Gottfredson, 2011a). These theories do not assume that people are inherently bad or immoral; rather, they assume that all people seek to pursue common motivations in accordance with their own view of self-interest and to maximize pleasure and avoid pain.

Control theories in criminology build on these assumptions by focusing on the constellation of controls (personal, social, legal, and situational) that inhibit the pursuit of self-interest via antisocial and problem behaviors. Control theories try to predict differences among individuals and groups in the constellation of these controls, which in turn will indicate which people, groups, and settings will have the most delinquency and crime (Gottfredson & Hirschi, 1990). Gottfredson and Hirschi argue that personal and social controls (as opposed to legal controls, emphasized by deterrence theories) are the most important factors in causing delinquency and crime. Control theories are sometimes referred to as restraint theories because it is the absence of effective restraints (from self, friends, family, and social institutions) that causes differences among people in crime and delinquency, rather than differences in motivations or incentives for crime.

According to self-control theory, people are not inherently criminal, nor are they socialized into crime; rather, people differ in the extent to which they have developed self-control and attend to the controls in their environment which inhibit crime and delinquency. Self-control and social control theories are appropriately regarded as socialization theories, since they focus on the factors that teach adherence to norms and social rules, assuming that children require training in how to conform to these expectations.

According to self-control theory, there is variability in the extent to which people are effectively taught these principles in childhood (Gottfredson & Hirschi, 1990; Hirschi & Gottfredson, 2003). The theory of self-control assumes that socialization differences in childhood produce a continuum among people (not discrete categories of “high and low” self-control) in their ability to focus on longer-term goals and that those with less self-control are more likely to engage in delinquent and criminal behavior than those with greater self-control. Once developed, individual differences in self-control remain relatively stable throughout life. It is not argued, however, that an individual’s self-control cannot change, particularly before adulthood, or that it is necessarily permanently “fixed” at an early age—stability is first an empirical observation.

Definition of Crime for Self-Control Theory

The theory of crime outlined by Gottfredson and Hirschi has a number of components that are integral to the theory and that distinguish it from other perspectives. One is the definition of the dependent variable for the theory—the definition of crime. For self-control theory, crime is defined as behaviors (events) that provide momentary or immediate satisfactions, but that have subsequent negative consequences. They have argued that crimes are essentially acts of force or fraud undertaken in pursuit of self-interest. Gottfredson and Hirschi thus use a behavioral rather than a legal definition of crime—although most criminal and delinquent acts qualify, not all do. Moreover, many noncriminal acts, such as many kinds of accidents, substance abuse, bullying, and school misconduct, qualify as acts with immediate benefits but subsequent costs (Gottfredson & Hirschi, 1990; Gottfredson, 2011b).

According to their general theory, most delinquent and criminal acts are highly opportunistic, momentary or adventitious, and require little by way of planning. Typically, they are easily dissuaded by obstacles such as locks, lights, or the presence of other people. They often involve momentary advantage in personal relations (many assaults) or assertion of self-interests. They typically promise little gain for the offender (although they often have a high cost to the victim); they require little ingenuity (breaking a window, bullying themselves to the front of the line, hitting with an available instrument); they are not a path to success or status or the satisfaction of some deep-seated psychological issue. Rather, they provide common or normal human satisfactions or wants in what appears to be an easy way, but only by ignoring costs.

Armed with this definition, it may be seen that self-control explains why delinquencies, crimes, and other problem behaviors “go together”: they all disproportionately appeal to individuals with relatively low self-control. It also accounts for the lack of specialization in types of crimes and for the versatility effect: interpersonal violence, stealing, drug use, accidents, and school misbehavior are commonly found to be in association as a result of individual differences in self-control. All the acts associated with these problems provide some immediate benefit for the actor (money, pleasure, the end of a troubling dispute), and carry with them the possibility of harmful consequences to the actor or others.

Origins of Self-Control: Early Childhood and the Family

Self-control theory begins with the assumption that human nature shares the general tendency to pursue satisfaction of individual needs and desires. Left unregulated, the pursuit of these needs and desires causes inevitable conflict with others and, consequently, potentially harmful consequences to the actor. As a result, those who care about the child seek to train the child to restrict the pursuit of acts of self-interest that also causes harm to the self or to others, and to attend to the needs and wants of others. For self-control theory, this process is what socialization entails: “parental concern for the welfare or behavior of the child is a necessary condition for successful child-rearing” (Gottfredson & Hirschi, 1990, p. 98). As the child develops, concerned and affectionate caregivers (parents, other relatives, friends and neighbors, and schools) monitor and sanction behavior harmful to the child and others. As a result, children are taught to pay attention to the longer-term consequences of their actions. When a caring adult is present in the developing child’s environment and takes an active role in socialization, high levels of self-control are established and appear to become a stable characteristic of the individual over the life-course. Sometimes, however, such early caregiving is not present in the child’s environment, because an adult who cares about the long-term interests of the child is not around or because the caregiver who is around lacks the skills or resources necessary to create self-control in the child.

As the child moves into formal school environments, the school can play an important role in developing self-control (Gottfredson & Hirschi, 1990, p. 105). Of course, self-control also greatly enhances prospects for successful school experiences. The theory postulates differences among groups, nations, and over time in the level and success of this socialization process. According to control theory, these differences produce differences in levels of crime, violence, and other problem behaviors among individuals, communities, and cultures, and in different time periods.

Emphasis on the learning of self-control in early childhood and on the important roles of the family and school is consistent both with the results of a large research literature on family effects on delinquency (see, e.g., Hirschi, 1969; Gottfredson & Hirschi, 1990; Hirschi & Gottfredson, 2003; Glueck & Glueck, 1950; West & Farrington, 1973; Loeber & Dishion, 1983; Loeber & Stouthamer-Loeber, 1986; Brannigan et al., 2002; Vazsonyi & Huang, 2010) and with demonstrations from control group studies that differences in family socialization practices affect both self-control and delinquency (Piquero et al., 2010, 2016).

Some researchers question the strength of these environmental causes and claim to have discovered strong biological causes for self-control (e.g., Cullen, Unneaver, Wright, & Beaver, 2008; Wright & Beaver, 2005; see Vazsonyi et al., 2017 for a careful review). Gottfredson and Hirschi argue that the evidence does not support the claim that biology mandates or determines that particular individuals must engage in delinquency or crime or in the specific manifestations of the versatility effect (what they refer to as “biological positivism” (Gottfredson & Hirschi, 1990, Ch. 3). Their self-control theory explicitly rejects the idea of a biological predisposition to crime or to delinquency (Gottfredson & Hirschi, 1990, pp. 61–62), even as it has always allowed for biologically caused individual differences in amenability to environmental causes of self-control. But the strong evidence for family effects and the lack of support for biological compulsion would seem to support the claim of self-control theory that socialization is nearly always possible, given an amenable environmental setting conducive to development of self-control in childhood. (Gottfredson & Hirschi, 1990; Hirschi & Gottfredson, 2003; Gottfredson, 2006).

Increasingly, the literature in education and economics (Cunha & Heckman, 2007) also demonstrates the importance of early family environment for development of individual skills such as self-control. According to Heckman (2007, p. 13250):

[T]hese skills are important determinants of educational attainment, crime, earnings, and participation in risky behavior. . . . Evidence of the importance of early environments on a spectrum of health, labor market, and behavioral outcomes suggests that common developmental processes are at work.

Self-control theory was influenced by the observation that people differ considerably in their tendency to ignore the long-term costs of their actions and that these differences appear before adolescence. When self-control becomes established, concern about parental disappointment, shame from family and friends, loss of affection, respect, and approval of significant others are the sanctions of greatest moment. With time, such concerns become a consistent and forceful part of the self and are carried throughout life. Self-control governs actions both consciously (some of the time) and preconsciously (much of the time), restraining unfettered self-interest, including commission of delinquent and criminal acts. A substantial body of recent scholarship from psychology, sociology, education, and child development is consistent with this model (Gottfredson, 2006; Baumeister & Heatherton, 1996; Moffitt et al., 2013; Heckman, 2006).

Foundational Facts for Self-Control Theory

Self-control theory was initially constructed with an appreciation for decades of research and literature on crime and delinquency. This literature represents an important foundation for theory, and as such the empirical status of self-control theory is tied ineluctably to the continuing validity of these correlates of crime and delinquency. They include the following (see, Gottfredson, 2006):

[T]the age distribution of crime, whatever the form of crime and whatever the characteristics of the person, peaks in late adolescence or early adulthood (around age 17–22) and declines sharply and continuously from this peak.

[T]hat delinquents are less likely than nondelinquents to be closely tied to their parents is one of the best documented findings of delinquency research.

There is enormous variability in the types of crime and delinquency committed by those who engage in crime and delinquency. . . . This versatility extends into analogous behavioral manifestations of low self-control such as truancy, dropping out of school, employment instability, alcohol abuse, drug abuse, child and spouse abuse, motor vehicle accidents, and unwanted pregnancy.

The correlation between the delinquency of the subject and the delinquency of his or her friends is one of the strongest in the field.

According to virtually all reviews of developmental research on antisocial and delinquent behavior . . . differences in antisocial behavior remain reasonably stable from the time they are first identified. Early antisocial behavior predicts antisocial behavior in adulthood.

School performance . . . strongly predicts involvement in delinquent and criminal activities. Those who do well in school are unlikely to get into trouble with the law.

(Gottfredson & Hirschi, 1991, p. 526; see also 1983);(Hirschi, 1969, p. 85);(Gottfredson & Hirschi, 1991, p. 528);(Gottfredson & Hirschi, 1990, p. 154);(Gottfredson & Hirschi, 1991, p. 527); and(Gottfredson & Hirschi, 1990, p. 107).

The fact that “early delinquency predicts later delinquency” provides two critical elements for the general theory that distinguishes it from other perspectives—the period of early childhood is extremely critical in the causation of crime and delinquency and the likelihood that childhood causes for problem behaviors will influence other causes coming later in life. The nature of self-control helps to account for the versatility effect—the fact that many delinquencies, crimes, and other problem behaviors seem to “go together” and that interpersonal violence, stealing, drug use, accidents, and school misbehavior are commonly found in association. The acts associated with these problems all provide some immediate benefit for the actor (money, pleasure, the end of a troubling dispute). But each also carries with it the possibility of harmful consequences. What differentiates people is not that such acts may provide them with benefits, but that some people routinely ignore the potential costs attendant on the acts and perform them anyway.

The strong and persistent correlates between attachment to parents (and from parents to children) and delinquency, and attachment to school and teachers and success in school, all strongly suggest that self-control is fostered by these relationships and by the success (or lack thereof) of parents and schools to effectively teach self-control or to teach children to care about and attend to their longer-term interests.

The empirical status of these foundational facts has not been in serious dispute among empirically oriented criminologists for decades. It seems safe to conclude that recent research continues to validate them (e.g., Vold et al., 2002; Farrington, 2003; Gottfredson, 2006).

Studies of the Direct Relationship between Self-Control and Crime

The general conclusion from contemporary research is that measures of self-control in childhood are regularly related, at a moderately strong level, to problem behaviors using a wide variety of measurement methods and study designs and in several disciplines. As Franken, Moffitt, and colleagues (2016) recently put it, “Impaired childhood self-control is highly important as it is associated with an abundance of negative life experiences, such as substance use, criminal offending, school dropout, or unplanned teenage pregnancies, and with negative long term health and financial outcomes . . . [also] adolescents with lower self-control are more likely to have deviant friends.”

The direct relationship literature has focused on the relationship between self control and characteristics of people, on the strength of the relationship for different places or cultures, for different types of crime or other problem behaviors, and over time. Vazsonyi et al. (2001) show self-control effects for males and females from four different countries and for five different age groups. Vazsonyi and Crosswhite (2003) show similar results for African American and Caucasian adolescents. DeLisi (2001a, 2001b) shows self-control effects among offender samples, and Baron (2003) provides them for property crime, drug use, and violent crime among homeless youths. Tittle et al. (2003) indicate that the self-control relationship holds for nonstudent adults, college students, youths, males and females, those with and without official criminal backgrounds, and among people in various countries and places. Tittle and colleagues also argue that “many types of measures of self-control predict a variety of acts. At least some measures of self-control predict some misbehavior for cross-sectional and longitudinal samples, as well as for experimental subjects” (2003, p. 144).

Vazsonyi and colleagues (2001) show common self-control effects for adolescent samples in the United States, Switzerland, Hungary, and the Netherlands. Moffitt et al. (2013) demonstrate the relevance of self-control in New Zealand. Self-control has been used to explain differences within Japan (Vazsonyi et al., 2004) and in Spain (Romero, Luengo, & Sobral, 2001). Vazsonyi et al., (2015) show self-control effects for 11 different countries.

Perrone et al. (2004) cite 13 studies of different crime types and analogous behaviors supporting the relationship between self-control and crime. Vazsonyi et al. (2001) show effects for theft and assault, alcohol and drugs, vandalism, and general deviance. Misconduct for active offenders (failure to appear, probation and parole arrests) is studied by DeLisi (2001a, 2001b), serious delinquency by Junger and Tremblay (1999), intimate violence by Sellers (1999), crime by Brownfield and Sorenson (1993) and Gibbs et al. (1998), occupational delinquency among juveniles by Wright and Cullen (2000), a wide variety of delinquent acts and drug use in French-speaking Canadian samples by LeBlanc and Girard (1997), general delinquency in a national probability sample of adolescents by DeLisi (2004), white-collar offending by Blickle et al. (2006), and general deviance by Gibbs et al. (1998, 2003).

The literature includes impressive demonstrations of the scope of versatility effects and of the connection between self-control and problem behaviors generally. An excellent example is Junger and Tremblay (1999), who provide evidence of the relation between accidents and delinquency and the relation between self-control and other problem behaviors (see also Junger et al., 2001). In general, the connection between self-control and the wide variety of analogous acts is documented by Perrone et al. (2004) for a list that includes cheating, drugs, accidents, traffic risks, school truancy, misbehavior, and dropout; school problems by Nakhaie, Silverman, and LaGrange (2000); accidents by Keane et al. (1993); cheating by Gibbs and Giever (1995); drinking, drug use and delinquency among adolescents by Zhang et al. (2002); grade point averages, self-esteem, binge eating, alcohol abuse, interpersonal relationships, and optimal emotional responses by Tangney et al.(2004); cigarette use, early unwed parenthood and early marriage by Martino et al. (2004); attention-deficit/hyperactivity disorder and bullying by Unnever and Cornell (2003); risky sex, driving behavior, academic dishonesty and gambling in college samples by Jones and Quisenberry (2004); and unemployment and homelessness by Baron (2003). Pinker (2011) argues that self-control changes substantially explain the general decline in violence across centuries.

Meta-analyses and General Reviews of Self-Control Research

A number of meta-analyses have been undertaken that are specifically focused on the relationship between self-control as described by Gottfredson and Hirschi and a dependent variable of delinquency or crime (Pratt & Cullen, 2000; Engel, 2012; Vazsonyi et al., 2017; see also the general review by Schulz, 2006; a related meta-analysis using a related concept of self-control and a broad range of outcome variables has been provided by de Ridder et al., 2012). Each of these studies finds consistent evidence that self-control is associated with delinquency, crime, and other problem behaviors. Pratt and Cullen’s (2000) widely cited study examined research selected for rigor published between 1993 and 1999. This included 21 studies and 49,727 individual cases. They concluded that their estimated effect size in excess of .20 “would rank self-control as one of the strongest known correlates of crime” and that “future research omitting self-control from its empirical analyses risks being misspecified” (2000, p. 952).

Vazsonyi et al. (2017) substantially updates the Pratt and Cullen study and includes a larger number of qualified studies, published between 2000 and 2010. A total of 99 studies with over 200,000 subjects were included. They found random effects mean correlation between self-control and crime and deviance of .415 for cross-sectional studies and .345 for longitudinal studies (design difference was not significant). The strongest effects of low self-control in their study were discovered for “general deviance” (r = .56) and for physical violence (r = .46). They concluded that their study of some of the best available research

provided strong and convincing evidence, based on about 100 cross-sectional and longitudinal studies, that a strong link between low self-control and deviance or crime exists and that it does not greatly vary across modes of assessment, across study designs (cross-sectional versus longitudinal), across measures of deviance, across different populations within the United States, but also across samples across cultures. In this sense, self-control theory has established itself as one of the most influential pieces of theoretical scholarship during the past century, as it continues to stand up to a plethora of rigorous empirical tests.

(Vazsonyi et al., 2017, p. 30).

Thus, with respect to reviews of some of the most rigorous research, reviewers consistently report strong validity for self-control theory from several disciplines and methodologies. In fact, reviews place the empirical support for the theory as among the strongest known to criminology. According to DeLisi (2005, p. 91):

Empirically, the relationship between low self-control and various antisocial outcomes has been nothing short of spectacular. Dozens of studies have explicitly tested the theory and found that low self-control was predictive of failure in family relationships, dating, attachment to church, educational attainment, and occupational status; risky traffic behavior; work-related deviance; having criminal associates and values; residing in a neighborhood perceived to be disorderly; and noncompliance with criminal justice system statuses. Moreover, persons with low self-control are significantly more likely to engage in drinking alcohol; substance abuse; smoking; gambling; violent, property, white-collar, and nuisance offending, and they are more likely to be victimized.

Contemporary Research on Age, Generality, and Stability Effects

In part owing to the conflicting expectations of general theory with the criminal career perspective (cf. Gottfredson & Hirschi, 1986, 1990, 2016 with Blumstein et al., 1986), a great deal of research interest has centered on depicting age, generality, and stability effects over the last several decades. This has resulted in a large empirical literature, using widely different methods, definitions, and samples. Several recent reviews in each of these areas have sought to summarize the literature on these topics in terms of the theory of self-control.

Results from the decades-long search for meaningful discrepancies from the now standard “age–crime curve,” particularly for “serious offenders” or for individuals whose high level of offending does not decline substantially with age, are compatible with the age-invariance hypothesis and its implications for theory and policy. For example, a recent report from the National Research Council of the National Academies on the causes and consequences of incarceration summarized the evidence:

The criminal career model assumes that the offending rate is constant over the course of the criminal career. However, large percentages of crimes are committed by young people, with rates peaking in the midteenage years for property offences and the late teenage years for violent offenses, followed by rapid declines . . . application of group-based trajectory modeling . . . show(s) that the offending trajectories of all identified groups decline sharply with age

(2014, p. 143).

In arguably the most important study of criminal careers to date, Laub and Sampson (2003, pp. 565–569) offer more specific information on this point: “[T]he classic age pattern (Hirschi & Gottfredson, 1983) is replicated even within a population that was selected for their serious, persistent delinquent activity.” Laub and Sampson describe the pattern of crime with age for serious offenders in their data as “fractal” of the overall distribution. (See also results and discussion of Danish data in Kyvsgaard, 2003, Ch. 17.)

In some samples, however, interpretation of the age data for offenders is controversial, particularly studies employing statistical “trajectory” methods (see, e.g., Macmillan, 2008). Recent, independent reviews of this statistical taxonomy literature (which searches for different “trajectories” of offending as subjects age) do not find consistent support for meaningful differential distributions by age (e.g., Skardhamar, 2009, p. 874). Summarizing two decades of taxonomic research using trajectory methodology in an effort to find significant groups of serious offenders who deviate from the standard age distribution of offending, Erosheva, Matsueda, and Telesca (2014) concluded that

the estimated trajectory groups in criminology exhibit weakly unimodal shapes with some differences in the level and timing of offending. This finding is consistent with the age-invariance thesis proposed by Hirschi and Gottfredson (1983). Empirically identified trajectory groups do not reveal the life-course persistent group with a constant rate of offending that criminal career and dual taxonomy approaches predict.

(2014, p. 316).

The self-control thesis for age in self-control theory is that the effect of age on crime and analogous behaviors is invariant across social and cultural conditions and that it applies to all demographic groups. Stability and versatility effects can be derived from it, and it enables a general theory by showing that diverse acts spread over the life course have common causes. It has implications for social policy and research design. It implies that conceptually similar measures of self-control will have similar effects at different ages and that policies like incapacitation will not be effective ways to lower crime rates. It thus “organizes in a consistent way an enormously diverse body of criminological findings” (Hirschi & Gottfredson, 1994, pp. 12–14). This diverse body of research to which age invariance pertains continues to suggest that it is the most tenable reading of the scientific research.

Generality or versatility continues to be regularly reported in research, even though some modest “specialization” in crime-type can be discovered statistically in large samples of recidivists (DeLisi, 2005). As summarized by Farrington (2003, p. 223), “offending is versatile rather than specialized . . . including heavy drinking, reckless driving, sexual promiscuity, bullying, and truancy.” An impressive body of empirical literature has extended the versatility finding well beyond the traditional definitions of crime and delinquency, to accidents, health and welfare behaviors (Cunha & Heckman, 2007; Heckman, 2006, 2007; Moffitt et al., 2011; Junger et al., 2001; Donovan et al.,1991). DeLisi and Piquero (2011, p. 291) put it this way: “There is a large stock of research on offense specialization and/or versatility (23 citations omitted). A main conclusion is that the preponderance of offenders, and by preponderance we mean virtually all offenders, are generalists” (see also DeLisi, 2005, p. 40).

The stability concept of self-control theory has also received research attention, with some scholars questioning the strength of the finding as a basis for self-control theory (e.g., Pratt, 2016). However, self-control theory is based on the well-substantiated observation of the substantial correlation over time between measures of early delinquency and subsequent offending. The relative stability of individual differences in crime, delinquency, and problem behaviors over the life-course is one of the most consistently reported findings in the field—although it is not by any means a perfect correlation (see the preceding discussion). The empirical observation is clearly substantial enough to indicate that a persistent individual characteristic or skill (self-control) is a major source. It is not, as Gottfredson and Hirschi repeatedly point out, a fact that implies that self-control is not caused, cannot change, can be measured by the same indicators at all points in life, or cannot be purposefully manipulated. The evidence remains substantial and consistent with the theory that self-control differences can be measured early in life and that these differences help predict later offending and movement into and out of many social, educational, and interpersonal roles throughout life.

The stability of the personal characteristic of self-control used by Gottfredson and Hirschi and the “lifelong impact of self-control” argued by Moffitt and colleagues (Moffitt et al., 2013 is inferred from the number of problem behaviors measured at different points in time. Researchers who report short-term instability in self-control typically use self-reported attitudinal or personality instruments that themselves have substantial unreliability. This unreliability is confounded with the “stability” of self-control and results in misleading estimates of stability. In any event, recent research does not alter Farrington’s claim that “there is marked continuity in offending and antisocial behaviors from childhood to teenage years to adulthood [citations omitted]. This means there is relative stability of ordering of people on some measure of antisocial behavior over time, and that people who commit relatively many offenses during one age range have a high probability of also committing relatively many offenses during another age range”(2003, p. 223).

Measurement of self-control is an active and important area of research and commentary about self-control theory. Strong arguments can be made for behavioral measurement of self-control rather than measures based on self-reported characteristics. How self-control should be measured at different points in the life-course (childhood, adolescence, young adulthood, and so on) is a key issue for future research. (For discussions of other measurement issues in self-control theory, see Hirschi & Gottfredson, 1994; Marcus, 2000, 2003, 2004; Schulz, 2006; and Duckworth & Kern, 2011.)

Research on Self-Control Selection Effects versus “Treatment” Effects

Self-control theory assumes that the individual characteristic of self-control affects decisions, associations, and affiliations to some degree throughout life. The selection of friends, partners, and institutional affiliations (e.g., continued schooling and jobs) will be influenced by differences in personal attributes like self-control. As a result, nonexperimental research without random assignment has difficulty sorting out the influence of affiliations (like peers or marriage partners or jobs) on individual crime rates as opposed to the influence of self-control (which may well have substantially caused the selection) or simply the effects of expected declines with age. Absent random allocation studies (difficult for obvious reasons), researchers look for various matching designs or noncontrol group methods to ascertain causal effects for such “treatments.” Presently, independent reviews argue that the results of this research regarding adult transitions (such as marriage effects) can be used to support both interpretations equally well (see, e.g., Skardhamar et al., 2015). The role of “peer groups” is similarly active and controversial in the empirical literature. Gottfredson and Hirschi (1990, Ch. 7) point out several implications of self-control theory for the study of peer effects, ranging from inherent weaknesses with self-reported measures of peer delinquency to selection effects. The strongest contemporary research is consistent with their arguments (Matsueda & Anderson, 1998; Franken et al., 2016), but given the centrality of peer effects to various theories of delinquency, research on this topic will remain active.