1999, Western Criminology Review. All Rights Reserved.

Building the Foundation for a Side-by-Side Explanatory Model:
A General Theory of Crime, the Age-Graded Life Course Theory, and Attachment Theory

Rebecca S. Katz

Citation: Katz, Rebecca S. 1999. "Building the Foundation for a Side-by-Side Explanatory Model: A General Theory of Crime, the Age-Graded Life-Course Theory, and Attachment Theory." Western Criminology Review 1(2). [Online]. Available: http://wcr.sonoma.edu/v1n2/katz.html.



A general theory of crime and the age-graded life-course theory are considered disparate explanatory models of the development of criminal behavior. However, both can be linked in a side-by-side fashion utilizing John Bowlby's theory of attachment. Early theoretical work by Travis Hirschi and Delbert Elliott discussed the necessity of theoretical integration to more thoroughly explain multiple pathways leading to deviance without reconciling divergent theoretical assumptions. Using a longitudinal sample of four hundred and eleven men, this research finds that self-control is not a time stable trait and that attachments to coworkers in early adulthood act as constraints on low self-control, leading to desistance.

Keywords: crime, criminal behavior, self-control, state dependence theory, attachment theory, life-course theory, violence, parental behavior, social capital, empathy

Building the Foundation for a Side-By-Side Explanatory Model: A General Theory
of Crime, the Age-Graded Life Course Theory, and Attachment Theory

The persistent heterogeneity and state dependence paradigms dominate criminological explanations of delinquency and crime (Nagin and Farrington 1992; Nagin and Paternoster 1991). Although both assume a causal relationship between early and later deviant behavior, the primary tenet of the persistent heterogeneity approach is that the relationship reflects an enduring personality characteristic that remains stable over time. In contrast, the state dependence paradigm views it as a product of enduring social circumstances that can be mediated by other factors.

Gottfredson and Hirschi's (1990) general theory of crime and Sampson and Laub's (1993) age-graded life-course perspective are representative of these two respective themes. Gottfredson and Hirschi posit that both imprudent and criminal behaviors can be predicted by a common characteristic: lack of self-control. They explain that lack of self-control does not require crime to exist and that self-control can be modified by opportunities and other constraints (Gottfredson and Hirschi 1990: 89). Numerous tests of Gottfredson and Hirschi's theory support their prediction concerning self-control and crime although some believe the theory does not explain enough variance in criminal behavior (Longshore, Turner, and Stein 1998). Finally, the general theory of crime assumes that although the individual's personality (i.e., the characteristic of self-control) remains stable through time, the relationship between self-control and crime is amenable to change. Conversely, Sampson and Laub's age-graded life-course perspective posits that both continuity and change exist throughout the life-course and that modifications in individual behavior may occur through new experiences or social circumstances.

Thus, these perspectives hold different assumptions about likely offenders. Nonetheless, they share a focus on a pivotal causal process: for Gottfredson and Hirschi, it is the emotional investment of the parents in the child's development of self-control; for Sampson and Laub, it is the emotional attachment of a previous offender to a job or a marriage, which leads to desistance. This similarity allows them to be linked using a psychological theory of deviance called attachment theory (e.g. Bowlby 1944; Horner 1991). This paper combines these approaches using a side-by-side theoretical linking and tests it using data from the Cambridge Youth Study (West 1973).


Side-by-side theoretical integration allows partial theories to be placed side-by-side and segregates the cases to which they are considered applicable (Elliot, Ageton, and Canter 1979; Hirschi 1979). Such an approach addresses the question of the relative importance of each perspective in explaining criminal behavior and allows each theory to maintain its own definition of delinquency. It is not considered comprehensive integration since the compatibility of theoretical assumptions is not addressed (Liska, Kroh, and Messner 1989). This side-by-side model links theories through a common independent variable, in this case infant-parent attachment (Bowlby 1969), emotional investment in the child (Gottfredson and Hirschi 1990), and attachment to a spouse or the workplace (Sampson and Laub 1993).

Attachment Theory

Attachment theory assumes that the development of an attachment between the primary caretaker and the child is the basic foundation of all future development (Bowlby 1969, 1988; Horner 1991), that there is a structure of personality present within each individual that is modified based upon on-going socialization experiences, and that the feeding behavior and the dependency needs of the child are motivational interpersonal focuses (Bowlby 1988: 24-26). Bowlby argues that attachment occurs in early childhood as the child perceives that some person in his or her environment behaves in a protective and nurturing manner. This is someone who is conceived of as being able to "better cope with the world" and, through the provision of safety and security to the child, begins the process of bonding (Bowlby 1988: 27). Thus, a child or infant develops a secure base when "he is nourished physically and emotionally, comforted if distressed, reassured if frightened...." (Bowlby 1988: 11).

Bowlby explains that delinquents are really "affectionless," that is, they have been unable to intimately connect with others. His original work (Bowlby 1947) describes a number of events occurring in the lives of children that appear to interfere with the development of a secure primary attachment to a caregiver, such as multiple placements, traumatic or arduous conditions in early childhood, and the early absence of a parent. The most delinquent boys and girls in Bowlby's original study were unable to intimately connect with others and were insecurely attached to their primary caretakers in early childhood. Since then, other researchers have demonstrated that loving and responsive parents or significant others facilitate a secure attachment style, while parents or significant others who do not demonstrate these qualities will be more likely to have children who develop an insecure attachment style (Howing et al. 1993; Isabella, Belsky, and von Eye, 1989).

Unfortunately, most research has not taken mother-infant attachment, early infant, or childhood attachment with caretakers into account to explain conformity or deviance. Farrington and Hawkins (1991) emphasize that childhood events prior to the age of eight are significant predictors of later adult criminality. Recent evidence (e.g., Raine, Brennan, and Mednick 1997) also indicates that insecurely or poorly attached children are more likely to engage in later violent behavior. An insecure attachment produces low levels of empathic understanding (e.g., Chaffin, 1992; Richardson et al. 1994). Horner (1991) suggests that the central underling factor involved in a secure attachment is the experience of empathy. A child develops self-control and empathy as the result of receiving empathic understanding from a parent or guardian. When potential offenders can perceive others as humans rather than as objects, they are less likely to inflict injury upon them.

Bowlby views the development of a secure attachment as a process by which the primary caretaker emotionally invests in the child. Thus, Gottfredson and Hirschi's emotional investment process and Bowlby's attachment process appear quite similar and allow both theories to be linked together in a side-by-side fashion. This emotional investment or attachment process facilitates the development of self-control by fostering empathic understanding and the development of trust, leading to non-deviant behavior. Such attachments--to a spouse, a workplace, or to coworkers, may occur later in life and repair the original attachment relationship. Only a limited number of studies take empathy into account in explaining criminality and most focus exclusively on sex offenders (Pithers 1994; Chaffin 1992). However, the absence of empathy is not unique to sex offenders (Richardson et al. 1994).

The General Theory of Crime

Although Gottfredson and Hirschi propose that person-level explanations of criminal behavior are all that are required, they also indicate that low self-control is not the "motivating force" leading to criminal behavior and that "..... the link between self-control and crime is not deterministic, but probabilistic, affected by opportunities and other constraints" (Hirschi and Gottfredson 1993:53). The existence of these "other constraints" may include later attachments to significant others or to work that prevent those with low self-control from offending. As mentioned above, Gottfredson and Hirschi also posit that lack of self-control predicts a wide variety of behaviors, including crime and the inability to commit to long-term relationships. While this position seems inconsistent with Sampson and Laub's assumption that attachment to others can lead to desistance, the present study assumes that low self-control will not predict marital or job attachments. If this is substantiated, then linking their construct of constraints to Sampson and Laub's construct of later attachments can be easily accomplished. Thus, by redefining Gottfredson and Hirschi's construct of self-control as a continuous rather than a dichotomous trait, I hope to develop a clear explanation of both change and continuity throughout the life course.

According to Gottfredson and Hirschi, low self-control is evident in early childhood through specific personality characteristics, such as an inability to postpone gratification, a low tolerance for frustration, and a tendency to engage in high levels of risk-taking behavior. Self-control develops through parental emotional investment in the child, monitoring the child's behavior, recognizing deviance when it occurs, and punishing the child. Parental emotional investment is necessary in order to activate the three forms of parental management (Gottfredson and Hirschi 1990: 95-97).

Although Gottfredson and Hirschi propose that children either develop self-control by the end of early childhood or fail to develop it at all, I suggest that self-control is not an "all or none" personality trait but rather that it exists on a continuum. Moreover, while Gottfredson and Hirschi assume that person-level traits are stable predictors of imprudent and criminal behavior, I argue that self-control does not uniformly predict behaviors. Some individuals with higher rather than lower levels of self-control may be more susceptible to the effects of constraints.

A variety of studies support the assertion of a causal relationship between self-control and criminality (McGee and Newcomb 1992; Pulkkinen and Pitkanen 1993). Further, other cross-sectional research indicates that low self-control predicts involvement in some crimes of force and fraud, especially among men (e.g., Grasmick et al. 1993; Nagin and Paternoster 1993). Longshore et al. (1998) nonetheless maintain that self-control is a weak predictor of crime. They note that "self-control seems not to improve predictive power or conceptual clarity regarding the etiology of crime" (Longshore et al. 1998: 179). Others hold that Moffitt's (1997) developmental model has more support than Gottfredson and Hirschi's general theory of crime (Bartusch et al. 1997).

Age-Graded Life-Course Theory

The state dependence theory of Sampson and Laub assumes that the causal relationship between early delinquent offending and later adult deviant behavior is not solely a product of individual characteristics; social events may change some individuals while others continue to offend. Their theory contains three main components. First, the micro-level structural context is mediated by informal family and school social controls, which can explain delinquency in childhood and adolescence. Next, there is continuity in antisocial behavior from childhood through adulthood in a variety of life domains. Finally, informal social bonds to family and employment during adulthood explain changes in criminality over the life span despite early childhood propensities (Laub and Sampson 1993: 7). Sampson and Laub's research and subsequent replications of their work substantiate their hypotheses (Sampson and Laub 1993; Paternoster and Brame 1997; Laub et al. 1998; Sommers et al. 1994; Horney et al. 1995).

Most important for the current study, Sampson and Laub find that attachments or social bonds in adulthood increase some individuals' social capital, leading to desistance from most types of deviant behavior, with the exception of men involved in drunkenness and violence. This paper links job and marital attachments in a side-by-side fashion to Bowlby's construct of infant-parent attachment. Men who become attached to coworkers or a spouse will increase their self-control; alternatively, as Gottfredson and Hirschi hypothesize, constraints in the form of job or marital attachments may prevent those with low self-control from offending.

In summary, combining attachment theory with the general theory of crime and life-course theory -- by linking the constructs of parental emotional investment to attachments and social bonds -- may strengthen the predictive power of each perspective. The development of self-control may be explained without attempting to reconcile the competing assumptions of these distinctive theories. Further, Bowlby's prediction that early secure attachment precludes deviant behavior resonates with Sampson and Laub's findings that later attachments to work or to a partner explain desistance. Linking these perspectives through Bowlby's attachment theory may better explain crime and desistance over the life course.


Prior research shows that family structure and processes influence the likelihood of later delinquent behavior (Sampson and Laub 1993; Rosen 1985). Sampson and Laub (1993) delineate nine structural and five process variables and find that most structural variables have indirect effects on delinquency through family process variables.

Moreover, McCord (1991) finds that the behavior of fathers and mothers are distinct and multiplicative in their impact on the probability of later delinquency. The findings of these studies, and a review of the literature (Wright and Wright 1994), lead to two important conclusions. First, family structure has both direct and indirect effects on delinquency while family function or process mediate some of these structural effects. Second, although the behavior of the mother influences adolescent deviance, some early behaviors of fathers also affect adult deviance.

Figure 1 illustrates the proposed order of early family process and structural variables that are expected to influence the development of self-control. The first four hypotheses reflect the proposed order as illustrated, from the beginning of the child's life (early attachment and absent father) through the development of (low) self-control.

Hypothesis 1: Early absent father is positively related to insecure attachment of the infant.

Substantial research reveals that the absence of a father figure early in male lives increases later delinquency (e.g. Bowlby 1947; Gabel 1992). Moreover, research finds that such an absence will directly affect a boy's ability to develop self-control (Johnson 1986; King 1993; Wenk et al. 1994) over the life-course. The secure attachment or emotional investment process facilitates the child's ability to develop and demonstrate both empathy and self-control. By extension, an insecure attachment will lead to lower levels of empathy and self-control, and to an increase in violent behavior.

Hypothesis 2: Insecure parent-infant attachment is inversely related to self control and empathy and positively related to violent behavior across the life-course.

Hypothesis 3: Insecure parent-infant attachment is positively related to inadequate attention to the child.

A child must receive parental attention in order to be monitored by the parent. I anticipate that the lack of parental attention will increase the rigid administration of parental rules and the over-vigilance of the parents. I also expect that attachment, or the emotional investment of the parent in the child, affects parental attention, which is a precursor to communicate parental rules to the child. Clearly, parental attention and the existence of specific parental rules are necessary in order to observe or monitor the child. Monitoring is measured here by a construct called vigilance. Vigilance will affect parental ability to discipline the child, which is then expected to directly affect self-control. Lack of parental attention is utilized as an indicator which precedes the other parental management behaviors. This is measured by the presence of physical neglect and the absence of parental praise. Physically neglectful parents, and parents who do not reward their children for good behavior, are quick to provide unreasonable and harsh punishments or to over-control their child (Bronstein et al. 1996; Cusinato 1994; Mosley and Thompson 1995). This measure of the lack of parental attention resembles indicators used by Sampson and Laub, such as parental rejection of the boy. Rigid rules and over-vigilance are similar to the indicators used by Laub and Sampson (1988:365) called Ferractic and Merratic. Laub and Sampson utilized these variables to separately measure mother's and father's behaviors that elicited fear and resentment in the child. Thus, this sequential ordering of parental behaviors is an elaboration of previous research on parental behaviors as well as an attempt to replicate and improve upon Sampson and Laub's work.


Hypothesis 4: The rigid and disinterested discipline of mother and father decrease child self-control. Mother's discipline and father's discipline have a reciprocal relationship.

Gottfredson and Hirschi discuss the importance of including what they refer to as the punishment of the child in the development of self control. The parental disciplinary indicators used here are measures of Gottfredson and Hirschi's construct of parental punishment. It is important to include both conjoint and separate indicators of the disciplinary styles of both parents because research finds that mother's and father's harsh, rigid, or lax forms of discipline increase delinquency (e.g., McCord 1979, 1991; Cernkovich and Giordano 1987).

Hypothesis 5: Self-control is inversely related to violent behavior; self-control is unrelated to job or marital attachments.

As noted above, studies find a link between self-control and criminal behavior but not between self-control and long-term commitments (marriage or work).

Hypothesis 6: A work-related attachment is inversely related to violence and positively related to self-control.

Hypothesis 7: Attachment to a woman is unrelated to desistance from violence or self-control.

Sampson and Laub's research clearly shows that attachments to marriage and work decrease most types of offending. However, attachment to wives has no effect on desistance from drunkenness and violence. They explain this finding through their qualitative analysis, which reveals that some of the substance-abusing men abuse their spouses (Sampson and Laub 1993). Thus, although work attachments have desistance effects upon public violence, marital attachments may provide men with a private arena in which to commit violent acts. Attachments to wives will have no effect on desistance from private violent behavior.

Figure 1 shows how the early family-of-origin indicators are expected to influence self-control. First, the absence of the father will positively influence insecure attachment. The development of an early insecure parent attachment will lead to low levels of both empathy and self-control. Empathy has a positive effect on self-control. The remainder of the model representing the sequential order of parenting attempts to follow Gottfredson and Hirschi's explication of the development of self-control.


The data for this research are taken from the Cambridge Study in Delinquent Development, a twenty-year longitudinal study of males living in a working class neighborhood in London, England (West 1973). Data collection began in 1961 and 1962, and the last wave available for study was collected in 1980. The men remaining in the study at that point were approximately twenty-four years of age. This secondary data set is ideal for this research as early family-of-origin variables are available.

Initial interviews were conducted with 411 boys at ages eight, ten, and fourteen by male and female psychologists. The respondents were also interviewed at ages sixteen, eighteen, twenty-one, and twenty-four by young male social science graduate students. Family members were interviewed throughout the boys' adolescence. This data set is restricted primarily to white males, as only twelve of the original group of 411 males are nonwhite.

Independent variables include early parental management indicators, and early and later attachment variables. Tables 1-A to 1-E (follow link to examine tables--Ed.) show the coding for each variable as well as the means, standard deviations, and factor loadings for each item or scale. An attempt has been made to include some of the same indicators of parenting used in Cernkovich and Giordano's (1987) analysis of delinquent behavior; Gottfredson and Hirschi's constructs of parental monitoring, parental recognition of deviant behavior, and punishment; and Laub and Sampson's constructs of Ferratic and Merratic.

The indicators of early infant-parent attachment include the fretfulness of the baby and feeding difficulties with the infant. Both items were retrospectively reported by the mothers when the boys were approximately age eight. Although these may not be the best indicators of insecure attachment, they are the only available measures of the early nature of the mother's relationship with her infant son.

Two items were combined to measure (low) self-control. The first was a combination of peer judgments about the daring of the boy at age ten to eleven and an earlier measure of the adventuresomeness of the boy as judged by his parents at age eight and nine. The second was the singular measure of the adventuresomeness of the boy. These two variables were not so highly correlated as to pose a problem of mulitcolinearity. Principle components factor analysis was conducted on these two variables and revealed that they loaded together into one dimension (alpha=.877, see Table 1 for details).

This paper assumes that changes in empathy are a result of receiving empathy, leading to increases in self-control. The indicator used to represent empathy is measured at a point in time later than the two measures used as indicators of self-control. The indicator of empathy was a questionnaire item that asked respondents what kept them out of trouble at age sixteen. Three types of responses were recorded, the first included the boy's mention of conscience and fear as most important in preventing his "getting into trouble." The second response reflected that their conscience was more important and the third response indicated that neither was important. Responses were coded such that a zero reflects both fear and conscience as important and singularly only conscience as important. A code of one represents fear of being caught as important or that neither fear nor conscience was important. The implications of the selection of conscience are that the boys experienced guilt or shame when considering deviant behavior. Such a selection is assumed to represent the capacity of the boy to have some sense of empathy. It is assumed here that those without empathic abilities are less likely to choose conscience as a more important consideration.

The parenting variables used are those measured early in the child's history. Although most delinquency research measures parental behaviors during the later years of the child's life, concurrent with indicators of delinquency, here I examine earlier parental behaviors which occur prior to the onset of delinquency in adolescence. As previous research indicates, these also influence adult criminal behavior (McCord 1991). First, parental behaviors of the mother were factor analyzed. This reveals that the harsh discipline of the mother as well as her strict discipline loaded on one dimension (alpha=.235). Another variable measuring the physical neglect of the boy and a second representing poor levels of parental praise loaded onto another dimension labeled as parental neglect (alpha=.273). Rigid rules of the parents and parental over-vigilance loaded on another factor and are described in Table 1-B (alpha=.368).

A scale that represents discipline of the father is also included. The first variable in the scale measures the strictness of the father's discipline; higher scores indicate stricter discipline. The second variable measures the father's overall quality of discipline; high scores represented disinterested discipline or harsh discipline (alpha=.325). The high factor loadings of these two item scales may preclude any serious problems with their use. Early analysis of the data also revealed that a very large proportion of these boys had already experienced the absence of their father for at least one month by the time they reached the age of five. Therefore, the structural variable of absent father is also included in the analysis.

There is strong evidence that adolescent deviant behavior is greatly influenced by the behavior of their peer group even after parental variables are taken into account (for reviews see Warr 1993; Moffitt 1997). Therefore, a measure of group deviance at age fourteen is included in the model that attempts to explain interpersonal violence at age fourteen. Previous research also shows that the best predictor of future deviance is previous deviance. Therefore, a measure of deviance from the preceding age category is included as an explanatory variable in each subsequent time period analyzed.

Several scales were developed to measure later attachment to marriage or work. At age eighteen two variables loaded on one factor, one of which measures job stability and the other positive feelings about the job (alpha=.612). Another scale measures job attachment at age twenty-four; it consists of two variables measuring job stability and job satisfaction. These loaded well on another factor (alpha=.505). Next, two forms of attachment to women are utilized, each represented by separate variables that measure long-term relationships with females at age eighteen and twenty-four (see Tables 1-C and 1-E).

Using ordinary least squares regression, self-control is regressed on the early parenting and early empathy indicators. Mother's discipline is regressed on the early parenting variable of lack of attention, insecure attachment, rigid rules, and over-vigilance of the parents, the scale measuring the father's disciplinary style, and absence of father. Similarly, father's disciplinary style is regressed on the same variables as mother's disciplinary style with mother's discipline now included as an independent variable. Lack of parental attention is regressed on early attachment and absent father. Rigid rules and over-vigilance of the parents are regressed on early insecure attachment, absent father, and the absence of parental attention.

In addition, regression analysis is used to examine predictors of self-reported involvement in fighting at different ages. Initial equations include early family-of-origin predictors and low self-control while the final equations include the later attachment variables (both marriage and job). These analyses examine the assumption of Gottfredson and Hirschi regarding the stability of self-control, as well as the assumption of the age-graded life-course theory that both change and stability are likely. Concurrent attachment relationships are utilized in regressions on violence. Current marital relationship is used to predict current violence and desistance. Listwise deletion of data is used to deal with loss of cases in later waves of the study.


The bivariate results indicate that there is no problem with multicolinearity among the independent variables (table not shown but available upon request). There is a positive relationship between absent father and early insecure attachment but it is not statistically significant. Thus, Hypothesis 1 is not supported.

Early empathy is positively but insignificantly related to attachment. It was expected that early insecure attachment and an absent father would be inversely related to empathy. However, when empathy is regressed, using logistic regression, on early insecure attachment and absent father, the model is a poor fit. Thus, Hypothesis 2 is unsupported here.

To test Hypothesis 2, the absence of parental attention is regressed on insecure attachment and absent father. The effects are in the expected direction but only absent father increases the likelihood of poor parental attention. Next, the parental rules scale is regressed on absent father, insecure attachment, and the absence of parental attention. Only insecure attachment increases the likelihood of the parent's use of rigid rules. Then, father's rigid discipline was regressed on the same independent variables listed previously along with mother's strict and disinterested disciplinary style and rigid parental rules and vigilance. Both the lack of parental attention and the harsh disciplinary style of the mother significantly increase the strict discipline of the father. When the parent's rigid rules and over-vigilance are included, eighteen percent of the adjusted variance is explained, whereas the other models only explain about one to five percent of the variance (results not shown but available upon request).

Next, the predictors of self-control are examined in the order specified in Figure 1. The results show that absent father and the discipline of the father decrease self-control while the rigid and over-vigilant rules of the parents significantly increase self-control. The effect of insecure attachment is marginal. Thus, Hypotheses 2 through 4 enjoy some support with the exception of the finding that rigid rules and parental over-vigilance act to increase the child's self-control. It may be that rigid parental rules and vigilance provide better monitoring of youth with insecure attachments while it is the father's rigid discipline that is most detrimental to the youth's self-control.

Explaining Self-Reported Violent Behavior

Fighting at age fourteen is regressed on all the early family-of-origin measures and peer involvement in delinquent activity (see Table 2). Twenty-two percent of the variance in aggressive behavior is explained; however, only two effects are significant. Early insecure attachment and peer delinquency increase involvement in fighting at fourteen, although the effect of the former is marginal. Thus, Hypothesis 2 gains some support.

Table 2

Regression of Self-Control and Violent Behavior
on Independent Variables

Independent Variables


Fighting at 14

Fighting at 18

Fighting at 18

Fighting at 21

Fighting at 21

Fighting at 24

























Absent dad








Lack of parents' attention








Rigid rules and over-vigilance








Dad's rigid discipline








Mom's rigid discipline








Peer deviance at age 14








Previous violence






Job attachment




Female attachment




Female attachment



















*p <.05

.10 < p >.10



As expected, fighting at eighteen is inversely related to self-control and empathy, respectively (see Table 2). Those young men with lower levels of self-control and empathy are more likely to be involved in aggressive behavior at eighteen. Before including the later attachment indicators in the model, it was necessary to determine whether self-control is significantly related to job or marital attachments. As predicted, the bivariate relationship between self-control and these attachment indicators is insignificant, thus partially substantiating Hypothesis 5 (correlations not shown but available upon request). The next model adds job and female attachments. As expected, job attachment significantly decreases fighting while attachment to a female increases fighting. It appears that marital attachments do not lead violent men to desistance, partially substantiating Hypothesis 7 and the previous work of Sampson and Laub. A recent review of the literature on violence against women also supports this expectation (Crowell and Burgess, 1996). Empathy and self-control continue to have a positive effect on violent behavior, although the latter becomes marginally significant. The decrease in the effect of self-control is contrary to expectations, it may be that Gottfredson and Hirschi are mistaken when they insist that self-control can not be changed. The adjusted explained variance increases to eight percent with both types of attachment included in the model.

In the next model, fighting at twenty-one is regressed on the early family-of-origin variables and fighting at eighteen. The model is not significant, and only previous involvement in fighting ("previous violence" in Table 2) has a positive effect on fighting at twenty-one. Similarly, when attachment to a female is included in the model, only previous involvement in fighting increases fighting at twenty-one, and the model is a poor fit (model not shown). However, job attachment significantly decreases fighting and decreases the strength of the effect of previous involvement in violent behavior. Only six percent of the variance in violent behavior is explained by this model.

Finally, using logistic regression, the dichotomous variable of involvement in fighting at twenty-four is regressed on all the early family-of-origin variables and involvement in fighting at twenty-one. The initial model was a poor fit; only self-control predicts involvement in violent behavior at twenty-four (data not shown). Next, marital attachment at age twenty-four is included in the model. Although the effect of self-control is suppressed, marital attachment at age twenty-four increases male involvement in violence, but the effect is not significant (p=.11). However, the effect of marital attachment is greater than the other items in the model (data not shown).

Job attachment at age twenty-four is not included in this model because of the small sample size. The equation is estimated using job attachment at age eighteen rather than at age twenty-four. However, even after this adjustment the sample size remains quite small. Nonetheless, the model is significan: only marital attachment has statistically significant effect on violence (see Table 2).

Overall, the evidence from each of the models reveals substantial evidence to refute Gottfredson and Hirschi's assumptions that low self control is a stable trait and that it consistently predicts deviant behavior. There is more support for Sampson and Laub's age-graded life-course theory that change through time is possible. Furthermore, Sampson and Laub's previous findings are supported in regard to the lack of desistant effects of marriage on violent offending and the desistant effects of job attachment among violent offenders.


Sampson and Laub's age-graded life-course theory appears to be a viable explanation of how change occurs in the lives of some men, while Gottfredson and Hirschi's perspective also has some support. However, contrary to Gottfredson and Hirschi's predictions, self-control appears to be a trait that exists on a continuum; it can be directly modified and does not consistently predict crime and imprudent behavior. Additionally, low self-control does not preclude the later likelihood of job or relational attachments among formerly persistent offenders. Clearly, these perspectives are capable of being linked in a side-by-side fashion through the use of the common construct of attachment. What this research also suggests is that empathy may be critical to the development of an attachment relationship.

Future research must continue to investigate the role empathy plays in the process of attachment and desistance. To accomplish this, criminologists must utilize findings from sociological and psychological research on marital quality, martial satisfaction, and work satisfaction in order to better understand how these relationships can supplant a damaged attachment template through the provision of empathic understanding. This may be the key to more clearly explaining how the early experience of empathic understanding leads to secure attachment, the development of empathy, and desistance. Clouding this process are the findings noted both in this research and in Sampson and Laub's work, which reveal that marital attachments do not lead to desistance among substance abusing men or violent men. Although not investigated here, Sampson and Laub discovered that these men did not desist from violence partly as the result of their history of substance abuse. Also, Sampson and Laub found, through their qualitative analysis, that violent men who were substance abusers physically assaulted their wives. Thus, future work must include indicators of substance abuse in investigating violent behavior, empathic understanding, and attachments to partners. Such work may find that there is something about substance abusers which precludes their accurately receiving empathic understanding.

James Messerschmidt's (1993) structured action theory may be useful in explaining the relationship between violence against women, substance abuse among men, and these men's inability to receive empathy. Messerschmidt proposes that adherence to an ideology of hegemonic masculinity interferes with men's relationships with women and men's capacity to adequately parent their own children. This hegemonic ideology may interfere with the receipt of empathic understanding, thus suppressing the effects of otherwise desistance-producing relationships. Therefore, future work should include hegemonic masculinity and empathic understanding, as well as both early family-of-origin and family of procreation indicators, into explanations of adult deviance and change through the life-course.

The inclusion of both early family-of-origin and later family-of-procreation indicators with hegemonic masculine ideology, substance abuse, and empathic understanding should increase the explained variance in this side-by-side model. Although the inclusion of early family-of-origin behaviors did not substantially increase the explained variance in this test, such early parental behaviors appear to affect later adult behavior through their effects on self-control. Additionally, a distinctive sequential order of parenting was uncovered which should be further explored through the use of path analysis techniques. Although the order of these parenting behaviors is somewhat different than expected, a sequential order clearly exists. The sequential order of parenting behaviors is best illustrated as shown in Figure 2: absent father affects poor parental attention, which then affects father's and mother's disciplinary style. Moreover, all the parenting variables examined, with the exception of the mother's harsh discipline, appear to have direct effects on self-control. Also as expected, most of the early family variables affect later violent behavior through their effect on self-control. Obviously, if later marital attachments have effects on violence and self-control, later family-of-procreation indicators may also have similar effects. For example, recent work by Farrington and West (1995) examines the desistance effects of fatherhood within the context of marital relationships and finds that fatherhood decreases the likelihood of subsequent conviction. Thus, families of procreation may play as large as a role in the development of desistance as early family-of-origin processes play in the development of self-control and delinquency.

Although Gottfredson and Hirschi are accurate in their outline of how parental management behaviors affect self-control, linking their work in a side-by-side fashion with Bowlby's attachment theory and Sampson and Laub's age-graded-life-course theory clearly improves their model. Finally, although this research leads the way for a further elaboration of all three perspectives, some problems remain which must be addressed before future tests are conducted. These relate to both measurement and generalizability issues and require the accomplishment of three objectives in order to clarify the complex effects of attachment and empathy on deviance. First, early infant-parent attachment, and later attachment relationships, must be more carefully operationalized through the ethnographic measurement of parental behaviors, including parental nurturing, empathic responsiveness of the parent to the infant, and the timeliness of these responses in accordance with the child's needs. Second, path analysis will help untangle the nature, direction, and strength of these effects. Third, although this data set contained only white males, the approach taken is developed to provide a good fit for females as well as males and for racial/ethnic minorities. Deviance among women and minority groups also need to be thoroughly examined using this model.


1. I would would like to thank Constance Hardesty and Shelley White for their editorial assistance on earlier drafts of this paper.


Bartusch, Jeglum, J. Dawn, Donald R. Lynam, Terrie E. Moffitt, and Phil A. Silva. 1997. "Is Age Important? Testing a General Versus a Developmental Theory of Antisocial Behavior." Criminology 35:13-48.

Bowlby, John. 1947. Forty-four Juvenile Thieves: Their Characters and Home Life. London: Bailliere, Tindall, and Cox.

________. 1969. Attachment and Loss, Volume I. Basic Books.

________ . 1988. A Secure Base: Parent-Child Attachment and Healthy Human Development. Basic Books Inc.

Bronstein, Phyllis, Paula Duncan, Adele D'Ari, Jean Pieniadz, Martha Fitzgerald, Craig L. Abrams, Barbara Franskowski, Oscare Franco, Connie Hunt, and Susan Y. Oh Cha. 1996. "Family and Parenting Behaviors Predicting Middle School Adjustment: A Longitudinal Study." Family Relations 45:415-426.

Cernkovich, Stephen A., and Peggy C. Giordano. 1987. "Family Relationships and Delinquency." Criminology 25:295-313.

Chaffin, Mark. 1992. "Factors Associated with Treatment Completion and Progress Intrafamilial Sexual Abusers." Child Abuse and Neglect 16:251-265.

Cusinato, Mario. 1994. "Parenting Over the Family LIfe Cycle." Chapter 5 in Handbook of Developmental Family Psychology and Psychopathology, edited by Lucinao L'Abate. New York: John Wiley and Sons, Inc.

Elliot, Delbert, Suzanne S. Ageton, and Rachelle Canter. 1979. "An Integrated Theoretical Perspective on Delinquent Behavior." Journal of Research in Crime and Delinquency 16:3-27.

Farrington, David and J. David, Hawkins. 1991. "Predicting Participation, Early Onset and Later Persistence in Officially Recording Offending." Criminal Behavior and Mental Health 1:1-33.

Farrington, David P. and Donald J. West. 1995. "Effects of Marriage, Separation, and Children on Offending by Adult Males." Current Perspectives on Aging and the Life Cycle 4:249-281.

Gabel, Stewart, M.D. 1992. "Behavioral Problems in Sons of Incarcerated or Otherwise Absent Fathers: The Issue of Separation." Family Process 31:303.

Gottfredson, Michael and Travis Hirschi. 1990. A General Theory of Crime. Stanford University Press.

Grasmick, Harold, G., Charles R. Tittle, Robert J. Bursik, and Bruce J. Arneklev. 1993. "Testing the Core Empirical Implications of Gottfredson and Hirschi's General Theory of Crime." Journal of Research in Crime and Delinquency 30:5-29.

Hirschi, T. 1979. " Separate and Unequal is Better." Journal of Research in Crime and Delinquency 16:35-38.

________ and Michael Gottfredson. 1993. "Commentary: Testing The General Theory of Crime." Journal of Research in Crime and Delinquency 30:47-54.

Horner, Althea J. 1991. Psychoanalytic Object Relations Therapy. Janson Aronson Inc..

Horney, Julie, Wayne D. Osgood, and Haen Ineke Marshall. 1995. "Criminal Careers in the Short-Term: Intra-Individual Variability in Crime and Its Relation to Local Life Circumstances." American Sociological Review 60:655-673.

Howing, Phyllis, T., John S. Wodarski, David P. Kurtz, and Martin Gaudin, Jr. 1993. Maltreatment and the School-Age Child. The Haworth Press Inc.

Isabella, Russell A., Jay Belsky and Alexander von Eye. 1989. "Origins

of Infant-Mother Attachment: An Examination of Interactional Synchrony

During the Infant's First Year." Developmental Psychology 25:12-21.

Johnson, Richard E. 1986. "Family Structure and Delinquency: General Patterns and Gender Differences." Criminology 24:65-84.

King, Valarie. 1993. "Nonresident Father Involvement and Child Well-Being. Can Dads Make a Difference?" Journal of Family Issues 18:78-96.

Laub, John H., Daniel S. Nagin, and Robert J. Sampson. 1998. "Trajectories of Change in Criminal Offending: Good Marriages and the Desistance Process." American Sociological Review 68:225-238.

Laub, John H., and Robert J. Sampson. 1993. "Turning Points in the Life Course: Why Change Matters to the Study of Crime." Criminology 31:301-325.

Longshore, Douglas, Susan Turner, and Judith A. Stein. 1998. "Reliability and Validity of Self-Control Measure: Rejoinder." Criminology 34:175-182.

McCord, Joan. 1991. "Family Relationships, Juvenile Delinquency, and Adult Criminality." Criminology 29:399-417.

McCord, Joan. 1979. "Some Child Rearing Antecedents of Criminal Behavior in Adult Men." Journal of Personality and Social Psychology 37:1471-1486.

McGee, Lina, and Michael D. Newcomb. 1992. "General Deviance Syndrome: Expanded Hierarchical Evaluations at Four Ages from Early Adolescence to Adulthood." Journal of Consulting and Clinical Psychology 60:766-776.

Messerschmidt, James W. 1993. Masculinities and Crime: Critique and Reconceptualization of Theory. Rowman and Littlefield Publishers, Inc.

Minty, Brian. 1988. "Public Care or Distorted Family Relationships: The Antecedents of Violent Crime." The Harvard Journal 27(3):172-187.

Moffitt, Terrie E. 1997. "Adolescent-Limited and Life-Course Persistent Offending: A Complementary Pair of Developmental Theories," Pp. 11-54 in Developmental Theories of Crime and Delinquency, Advances in Criminological Theory, edited by Terence P. Thornberry. Transaction Publishers.

Mosley, Jane and Elizabeth Thompson. 1995. "Fathering Behavior and Child Outcomes: The Role of Race and Poverty." Chapter 8 in Fatherhood Contemporary Theory, Research, and Policy: Research on Men and Masculinities, edited by William Marsiglio. Sage Publications.

Nagin, Daniel, S. and David P. Farrington. 1992. "The Onset and Persistence of Offending." Criminology 30:501-528.

Nagin, Daniel S. and Raymond Paternoster. 1991. "On the Relationship of Past to Future Participation in Delinquency." Criminology 29:163-189.

Nagin, Daniel S. and Raymond Paternoster. 1993. "Enduring Individual Differences and Rational Choice Theories of Crime." Law and Society Review 27:467-496.

Paternoster, Raymond, and Robert Brame. 1997. "Multiple Routes to Delinquency? A Test of Developmental and General Theories of Crime." Criminology 35:49-84.

Pithers, William. 1994. "Process Evaluation of a Group Therapy Component Designed to Enhance Sex Offenders' Empathy for Sexual Abuse Survivors." Behavior Research and Therapy 32:656-571.

Pulkkinen, Lea and Tuuli Pitkanen. 1993. "Continuities in Aggressive Behavior From Childhood to Adulthood." Aggressive Behavior 19:249-263.

Raine, Adrian, Patricia Brennan, and A. Sarnoff Mednick. 1997. "Interaction Between Birth Complications and Early Maternal Rejection in Predisposing Individuals to Adult Violence: Specificity to Serious, Early-Onset Violence." American Journal of Psychiatry 154:1265-1271.

Richardson, Deborah R., Georgina S. Hammock, Stephen M. Smith, and Manuel Gardner. 1994. "Empathy as a Cognitive Inhibitor of Interpersonal Aggression." Aggressive Behavior 20: 275-290.

Sampson Robert, J. and John H. Laub. 1993. Crime in the Making Pathways and Turning Points Through Life. Harvard University Press.

Sampson Robert, J. and John H. Laub. 1996. "Socioeconomic Achievement in the Life Course of Disadvantage Men: Military Service as a Turning Point, Circa 1940-1965." American Sociological Review 61:347-367.

Sommers, Ira, Deborah R. Baskin, and Jeffrey Gagan. 1994. "Getting Out of the Life: Crime Desistance by Female Street Offenders." Deviant Behavior 15:125-149.

Warr, Mark. 1993. "Parents, Peers and Delinquency." Social Forces 72: 247-264.

West, D.J. 1973. Who Becomes Delinquent? Second Report of the Cambridge Study in Delinquent Development. London, Heinemann Educational Center: Publishers for the Cambridge Institute of Criminology.

Wenk, DeeAnn, Constance L. Hardesty, Carolyn S. Morgan, and L. Blair Sampson. 1994. "The Influence of Parental Involvement on the Well-Being of Sons and Daughters." Journal of Marriage and the Family 56:229-234.

Wright, Kevin, N. and Karen E. Wright. 1994. Family Life, Delinquency, and Crime: A Policymaker's Guide, Research Summary. Washington, D.C.: Office of Juvenile Justice.



Return to Current Issue