CJTS4
JOURNAL OF RESEARCH IN CRIME AND DELINQUENCYAgnew / GENERAL STRAIN THEORY
BUILDING ON THE FOUNDATION OF GENERAL STRAIN THEORY:
SPECIFYING THE TYPES OF STRAIN MOST LIKELY TO LEAD TO CRIME AND DELINQUENCY
ROBERT AGNEW
General strain theory (GST) is usually tested by examining the effect of strain on crime. Researchers, however, have little guidance when it comes to selecting among the many hundreds of types of strain and have trouble explaining why only some of them are related to crime. This article builds on GST by describing the characteristics of strainful events and conditions that influence their relationship to crime. Strains are said to be most likely to result in crime when they (1) are seen as unjust, (2) are seen as high in magnitude, (3) are associated with low social control, and (4) create some pressure or incentive to engage in criminal coping. Drawing on these character- istics, it is predicted that some types of strain will not be related to crime, including types that have dominated the research on strain theory, and that others will be related to crime, including types that have been neglected by empirical researchers.
General strain theory (GST) argues that strains or stressors increase the likelihood of negative emotions like anger and frustration. These emotions create pressure for corrective action, and crime is one possible response (Agnew 1992). Crime may be a method for reducing strain (e.g., stealing the money you desire), seeking revenge, or alleviating negative emotions (e.g., through illicit drug use). GST builds on previous strain theories in several ways: most notably, by pointing to several new categories of strain, including the loss of positive stimuli (e.g., loss of a romantic partner, death of a friend), the presentation of negative stimuli (e.g., physical assaults and verbal in- sults), and new categories of goal blockage (e.g., the failure to achieve justice goals). Recent research demonstrates that many of the specific strains falling under these categories are related to crime and delinquency (see Agnew 2001a for a summary; Aseltine, Gore, and Gordon 2000; Mazerolle et al. 2000; Piquero and Sealock 2000). The specification of these new categories of strain is GST’s greatest strength.
JOURNAL OF RESEARCH IN CRIME AND DELINQUENCY, Vol. 38 No. 4, November 2001 319-361 © 2001 Sage Publications
319
This strength, however, is also GST’s biggest weakness. GST is so broad that researchers have little guidance as to the specific types of strain to exam- ine in their research. Hundreds of types of strain fall under the major catego- ries of strain listed by GST, as reflected in recent inventories of stressful life events, chronic stressors, and daily life events or hassles (see Cohen, Kessler, and Gordon 1995; Herbert and Cohen 1996 for overviews). And even these inventories do not measure many of the strains described by GST. Further- more, the broadness of GST makes it difficult to falsify. As Jensen (1995) stated, “if strain can be defined in so many different ways, then strain theory is virtually unfalsifiable. There is always a new measure that might salvage the theory” (p. 152).
It is therefore crucial that GST more precisely specify the types of strain most likely to lead to crime and delinquency. This article represents an attempt to do that. First, strain is defined. Although Agnew (1992) presented a general definition of strain, the term has nevertheless been used in different ways by researchers and it is important to clarify its meaning. Second, previ- ous tests of GST are reviewed to determine what they say about the types of strain most likely to lead to crime. Third, the characteristics of those types of strain most likely to lead to crime are described. Briefly, such strains (1) are seen as unjust, (2) are seen as high in magnitude, (3) are associated with low social control, and (4) create some pressure or incentive to engage in crime. Fourth, these characteristics are then used to predict the likelihood that sev- eral types of strain will result in crime. Fifth, suggestions for empirical research are provided.
WHAT IS STRAIN?
Before discussing the types of strain most likely to lead to crime, it is first necessary to clarify what is meant by the term strain. Agnew (1992) stated that strain refers to “relationships in which others are not treating the individ- ual as he or she would like to be treated” (p. 48). Even so, researchers use the term in different ways. Some refer to an objective event or condition (e.g., the infliction of physical abuse, the receipt of poor grades at school), some to the individual’s evaluation of an event or condition (e.g., whether juveniles like the way their parents or teachers treat them), and some to the emotional reaction to an event or condition (e.g., whether respondents are angry at how others treat them). To help clarify the meaning of strain, the following defini- tions are proposed.
Objective strains refer to events or conditions that are disliked by most members of a given group. So, if we state that an individual is experiencing objective strain, we mean that he or she is experiencing an event or condition
320 JOURNAL OF RESEARCH IN CRIME AND DELINQUENCY
that is usually disliked by members of his or her group. Many events and con- ditions are disliked by most people, regardless of group membership (e.g., physical assault, lack of adequate food and shelter). The evaluation of other events and conditions varies with group characteristics, such as gender and age (e.g., Broidy and Agnew 1997; Elder, George, and Shanahan 1996). It is, of course, important for researchers to consider the possibility of such group differences when constructing measures of objective strain.
Empirically, it is possible to determine the objective strains for group members in several ways. Observational research is one method. Anderson (1999), for example, described many of the objective strains in a poor, inner-city, African American community. Surveying a representative sample of group members or people familiar with the group is another method, and both have been employed in the stress research (Turner and Wheaton 1995). In particular, respondents can be asked whether they (or group members) would dislike a range of events and conditions. It is important to present respondents with preestablished lists of events/conditions and to ask them to list events/conditions not on the list. This helps to ensure that a complete list of objective strains is developed.1
Subjective strains refer to events or conditions that are disliked by the peo- ple who are experiencing (or have experienced) them. So, if we state that indi- viduals are experiencing subjective strain, we mean that they are experienc- ing an event or condition that they dislike. One of the key findings to emerge from the stress research is that individuals often differ in their subjective eval- uation of the same objective strains. For example, people differ in how they subjectively evaluate such objective strains as divorce and the death of a fam- ily member. The subjective evaluation of an objective strain is a function of a range of factors, including individual traits (e.g., irritability), personal and social resources (e.g., self-esteem, self-efficacy, social support), goals/values/ identities, and a range of life circumstances (for overviews, see Dohrenwend 1998; Kaplan 1996; Lazarus 1999). Wheaton (1990), for example, found that the quality of ones’prior marriage strongly influenced how people evaluated their divorce, with people in bad marriages evaluating their divorce in posi- tive terms. It is also important to note that an individual’s evaluation of an objective strain frequently changes over time as the individual copes with the strain. So, although there is a relationship between objective and subjective strain, it is far from perfect.
Most of the research on strain theory employs measures of objective strain (although see Agnew and White 1992). Researchers ask individuals whether they have experienced a certain event or condition (e.g., Did you fail any classes? Do your parents yell at you?); no effort is made to measure the indi- vidual’s subjective evaluation of this event/condition. This may cause
Agnew / GENERAL STRAIN THEORY 321
researchers to underestimate the support for strain theory because objective strains sometimes create little subjective strain. This does not mean, however, that researchers should simply employ subjective measures of strain. It is important to examine objective strains as well because this allows us to better distinguish external events from the subjective evaluation of such events. We can then examine individual and group differences in both the exposure to external events/conditions likely to cause strain and the subjective evaluation of those events/conditions. Furthermore, we can explore the factors that influence individual and group differences in the subjective evaluation of the same external events and conditions. This is critical if we are to fully explain individual and group differences in crime. As an illustration, Bernard (1990) argued that poor, inner-city residents have higher rates of violence not only because they experience more objective strains but also because they are more sensitive to such strains (also see Thoits 1995 on individual and group differences in the “vulnerability” to stressors).
The emotional response to an event or condition is closely linked to sub- jective strain. Subjective strain deals with the individual’s evaluation of an event or condition. There are many definitions of emotion, but most state that a central component of an emotion is an evaluation of or an affective response to some object or behavior or idea. Most theorists, however, go on to state that emotions involve more than an evaluation or affective response. For example, they also involve changes in physiological or bodily sensations (see Berkowitz 1993; Smith-Lovin 1995; Thoits 1989). Building on this argu- ment, I would contend that subjective strain is distinct from the full emotional reaction to strain.
Two individuals may evaluate an event/condition in the same way; that is, they may both dislike it an equal amount. So, they have the same level of sub- jective strain. One may become angry in response to the strain, however, whereas the other may become depressed. And they may differ in the degree to which they experience certain emotions, so one may become quite angry, whereas the other may experience only mild anger. So the same subjective strain may result in rather different emotional reactions. Again, a range of individual and environmental factors influences the emotional reaction to subjective strain. The potential utility of distinguishing between subjective strain and the emotional reaction to strain is highlighted by Broidy and Agnew (1997). They argued that males and females often differ in their emo- tional reaction to subjective strains. Although both males and females may experience anger, the anger of females is more likely to be accompanied by feelings of guilt, depression, and anxiety. These additional emotions are said to reduce the likelihood of other-directed crime, thereby helping us explain gender differences in such crime.
322 JOURNAL OF RESEARCH IN CRIME AND DELINQUENCY
RESEARCH ON THE TYPES OF STRAIN MOST LIKELY TO LEAD TO CRIME AND DELINQUENCY
Agnew (1992) described those types of events and conditions most likely to be classified as objective strains and to result in subjective strain. Such events/conditions involve goal blockage, the loss of positive stimuli, and/or the presentation of negative stimuli. They are also high in magnitude (degree), recent, and of long duration. But as indicated earlier, hundreds of events/conditions meet these criteria, and so there are potentially hundreds of objective and subjective strains. Agnew did not discuss whether certain of these strains are more likely to result in crime than others. Rather, he treated these strains as more or less equivalent in terms of their impact on crime. He argued that whether they result in crime is largely a function of the character- istics of the individuals experiencing the strain. In particular, strain is most likely to lead to crime when individuals lack the skills and resources to cope with their strain in a legitimate manner, are low in conventional social sup- port, are low in social control, blame their strain on others, and are disposed to crime. This article builds on Agnew by arguing that the effect of strain on crime is not only a function of individual characteristics but also of the type of strain experienced by the individual. Certain types of strain—either objective or subjective strain—are more likely to result in crime than other types.
Previous research on GST provides some information about the types of strain most likely to lead to crime, although much of this research suffers from two problems that severely limit its utility. First, most tests of GST only examine a small portion of the strains described by Agnew (1992). These tests tend to make use of existing data sets, which were not collected for the purpose of testing GST. As a consequence, many key strain measures are missing—particularly measures of the types of goal blockage described by Agnew and measures of certain types of negative treatment, like peer abuse and experiences with racial discrimination and prejudice. So we have little idea whether these types of strain are related to delinquency. Second, most tests of GST examine the effect of a single, cumulative strain measure on delinquency. In some cases, a measure of stressful life events is employed. Hoffmann and associates, for example, tested GST using a 16- to 18-item measure that focuses on events like “death, illness, or accidents among fam- ily or friends; changes in school or residence; parental divorce or separation; and family financial problems” (Hoffmann and Cerbone 1999; Hoffmann and Miller 1998; Hoffmann and Su 1997; also see Aseltine et al. 2000). In other cases, the cumulative strain measure is a composite of several scales and/or items measuring a range of different types of strain, such as neighbor- hood problems, negative relations with adults, the failure to achieve educa- tional and occupational goals, breaking up with a romantic partner or friend,
Agnew / GENERAL STRAIN THEORY 323
and getting lower grades than you deserve (e.g., Mazerolle 1998; Mazerolle et al. 2000; Mazerolle and Piquero 1997). The use of such cumulative mea- sures means that we lack information on the effect of the individual strain measures.
Researchers employ cumulative measures of strain because Agnew (1992) argued that it is not the effect of one specific strain or stressor that is important; rather, it is the cumulative effect of all the strains experienced by the individual. He recommended combining individual strain measures into a single scale so as to better estimate this cumulative effect (pp. 62-63). It is assumed that all or most of the individual strain measures in the cumulative scale make some contribution to crime. As will be argued below, there is good reason to question this assumption. Most cumulative measures encompass a wide range of strains, and it is likely that some contribute to crime and some do not. Given this fact, it is not surprising that most cumulative measures have only a moderate impact on crime. A consideration of different types of strain, however, might reveal that some have a strong impact on crime, whereas oth- ers have little or no impact.
Some tests of GST do examine the impact of different types of strain on crime among adolescents. Agnew and White (1992) examined the effect of eight strain measures on delinquency, including both general and specific measures. They found that negative life events, life hassles, negative relations with adults, and parental fighting are significantly associated with delin- quency. Neighborhood problems, unpopularity with the opposite sex, occu- pational strain, and clothing strain are not associated with delinquency. Pater- noster and Mazerolle (1994) examined the effect of five strain measures on delinquency. They found that neighborhood problems, negative life events, school/peer hassles, and negative relations with adults are significantly asso- ciated with subsequent delinquency, whereas a measure of educational and occupational expectations is not (see Mazerolle 1998 for information on gen- der differences in the effect of these strain measures). Aseltine et al. (2000) found that family and peer conflict (through anger) are related to selected types of delinquency. Agnew and Brezina (1997) found that poor relations with peers is related to delinquency, whereas unpopularity with peers is not. Piquero and Sealock (2000) found that physical and emotional abuse in the household (toward the juvenile and others) is related to delinquency (also see Brezina 1999). Tests of classic strain theory typically find that the failure to achieve educational and occupational goals is not related to delinquency (see Agnew 1995a). The failure to achieve economic goals, however, may be related to delinquency (Burton and Dunaway 1994).
Many other studies have not set out to test GST but have examined types of strain that fall under the theory. Several studies found that adolescent crime is
324 JOURNAL OF RESEARCH IN CRIME AND DELINQUENCY
significantly related to criminal victimization; parental abuse and neglect; parental rejection; disciplinary techniques that are excessive, very strict, erratic, and/or punitive (e.g., nagging, yelling, threats, insults, and/or hit- ting); family conflict; parental divorce/separation; and negative experiences at school (low grades, poor relations with teachers, and the perception that school is boring and a waste of time). Summaries of these studies are pro- vided in Agnew (1992, 1995b, 1997, 2001a, 2001b). Studies of adults sug- gest that crime is related to marital problems, work in the secondary labor market, unemployment in certain cases, and possibly the failure to achieve economic goals (Agnew et al. 1996; Baron and Hartnagel 1997; Cernkovich, Giordano, and Rudolph 2000; Colvin 2000; Crutchfield and Pitchford 1997; Sampson and Laub 1993; Uggen 2000). There has not been enough good research on other types of strain to draw any firm conclusions about their relationship to crime.
The above studies, then, suggested that certain types of strain are related to crime whereas others are not. At this point, it seems safe to conclude that crime is related to verbal and physical assaults, including assaults by parents, spouses/partners, teachers, and probably peers. Crime is also related to parental rejection, poor school performance, and work problems, including work in the secondary labor market. Crime is not related to the expected fail- ure to achieve educational/occupational success or to unpopularity with peers. Beyond that, the relationship between various strains and crime is unclear.
These data pose a major problem for GST: Why is it that only some types of strain are related to crime? At present, GST offers little guidance in this area. GST, for example, does not allow us to explain why verbal and physical assaults are related to crime, but the failure to achieve educational/ occupational goals and unpopularity with peers is not. All of these strains fall under the categories listed by Agnew (1992), and they are frequently high in magnitude (degree), recent, and of long duration.
Recent versions of GST do argue that certain types of strain are especially relevant to crime (Agnew and Brezina 1997; Broidy and Agnew 1997). Agnew (1997, 2001a, 2001b), for example, argued that although many types of goal blockage may lead to delinquency, the failure to achieve monetary, autonomy, and “masculinity” goals are of special importance. And he argued that although a range of negative or noxious stimuli may cause delinquency, physical and verbal assaults are of special importance. These suggestions, however, are not derived from theory. Rather, they represent ad hoc attempts to explain empirical findings or to incorporate other theoretical and empirical work into GST. Much theoretical and empirical work, for example, suggests that threats to one’s status, particularly one’s masculine status, contribute to
Agnew / GENERAL STRAIN THEORY 325
crime in certain groups (Anderson 1999; Messerschmidt 1993). Likewise, some theoretical and empirical work suggests that the blockage of autonomy goals contributes to delinquency (Agnew 1984; Moffitt 1993; Tittle 1995).
And although empirical research is starting to point to those types of strain that are and are not related to delinquency, it is not wise to depend on such research to fully resolve this issue. There are hundreds of specific types of strain; it will take empirical researchers a long while to determine their rela- tive importance (although observational research and open-ended, intensive interviews can be of some help here). Furthermore, we would still lack an explanation of why some types of strain have a greater effect on crime than other types. The lack of such an explanation might cause us to overlook cer- tain important types of strain. It is therefore important for GST to better explain why some types of strain are more likely to lead to crime than other types.
THE CHARACTERISTICS OF THOSE TYPES OF STRAIN MOST LIKELY TO LEAD TO CRIME
Individuals may cope with strain in a number of ways, only some of which involve crime (see Agnew 1992). Individuals may cope using a variety of cognitive strategies, most of which attempt to redefine strainful events and conditions in ways that minimize their adversity. Individuals may employ behavioral coping strategies that are intended to terminate, reduce, or escape from the strainful events and conditions. Certain of these strategies involve conventional behaviors (e.g., negotiating with the people who harass you), whereas others involve crime (e.g., assaulting the people who harass you). And they may employ emotional coping strategies that are intended to allevi- ate the negative emotions that result from strain. Certain of these strategies involve conventional actions (e.g., listening to music), whereas others involve crime (e.g., illicit drug use). It is argued here that some types of strain are more likely to result in crime than other types because they influence the ability to cope in a noncriminal versus criminal manner, the perceived costs of noncriminal versus criminal coping, and the disposition for noncriminal versus criminal coping. (As indicated above, these factors are also affected by a range of individual characteristics.)
The characteristics of those types of strain most likely to result in crime are discussed in this section, with the discussion referring to both objective and subjective strains. In brief, it is argued that strains are most likely to result in crime when they (1) are seen as unjust, (2) are seen as high in magnitude, (3) are associated with low social control, and (4) create some pressure or incentive to engage in criminal coping. These characteristics are derived
326 JOURNAL OF RESEARCH IN CRIME AND DELINQUENCY
primarily from the stress, justice, and emotions literatures (see references below); the social interactionist theory of coercive behavior (Tedeschi and Felson 1994); defiance theory (Sherman 1993); reintegrative-shaming the- ory (Briathwaite 1989); frustration-aggression theory (Berkowitz 1993); techniques of neutralization or moral disengagement theory (Bandura 1990; Sykes and Matza 1957); differential coercion theory (Colvin 2000); social control theory; social-learning theory; and the routine activities perspective (Cullen and Agnew 1999). There is a discussion of why these characteristics are important and how researchers can determine whether specific types of strain possess these characteristics. In the next section, these characteristics are used to predict the likelihood that several specific types of strain will result in crime.
The Strain Is Seen as Unjust
Agnew (1992) presented unjust treatment as a distinct category of strain, classified under “the failure to achieve positively-valued goals.” In particular, Agnew spoke of the disjunction between just/fair outcomes and actual out- comes. It is here argued that unjust treatment is not a special type of strain dis- tinct from the other types. The issue of injustice applies to all types of strain; that is, it is possible to classify any type of strain according to the extent to which it is seen as unjust. Those types of strain seen as unjust should be more likely to lead to crime, primarily because they are more likely to provoke emotions conducive to crime like anger.
Much data from the emotions and justice literatures indicate that there is a strong link between unjust treatment and anger (see Agnew 1992, 68-69; Averill 1982, 1993; Berkowitz 1993; Hegtvedt and Cook forthcoming; Hegtvedt and Markovsky 1995; Mikula 1986; Mikula, Petri, and Tanzer 1990; Tedeschi and Felson 1994; Tedeschi and Nesler 1993; Tyler 1994; Tyler et al. 1997). And limited data suggest that anger increases the likeli- hood of crime, particularly violent crime (Agnew 1985; Aseltine et al. 2000; Berkowitz 1993; Brezina 1998; Mazerolle et al. 2000; Mazerolle and Piquero 1998; Piquero and Sealock 2000; Tedeschi and Felson 1994; Tyler et al. 1997). Anger fosters crime because it disrupts cognitive processes in ways that impede noncriminal coping; for example, it leads individuals to disre- gard information that may help resolve the situation, and it reduces the ability to clearly express grievances. Anger also reduces the actual and perceived costs of crime; for example, angry individuals are less likely to feel guilt for their criminal behavior because they believe that the injustice they suffered justifies crime. Finally, anger energizes the individual for action, creates a sense of power or control, and creates a desire for revenge or retribution—all of which lead individuals to view crime in a more favorable light (see Agnew
Agnew / GENERAL STRAIN THEORY 327
1992; Averill 1982, 1993; Cloward and Ohlin 1960; Gottfredson and Hirschi 1990; Tedeschi and Felson 1994; Tedeschi and Nesler 1993; see Tyler et al. 1997 on retributive justice).
Measuring injustice. There are several ways to measure the perceived injustice of particular strains. The perceived injustice of objective strains can be estimated by (1) researchers, with such researchers drawing on the justice and attributions literature (see below) and their knowledge of the group being examined; (2) a panel of judges familiar with the group being examined, with such judges being asked to estimate the likelihood that various strains will be seen as unjust by group members; and/or (3) a representative sample of group members, with such members being asked to rate the injustice of various strains (see Mikula 1993; Mikula et al. 1990). The ratings of judges and group members can be averaged. It is best to provide judges and group mem- bers with moderately specific descriptions of the strains being rated because the specific features of the strain can have a large impact on ratings of injus- tice (see below). For example, instead of asking individuals to rate the injus- tice of “a close friend dying,” it is better to ask them to rate the injustice of “a close friend being shot to death by a rival gang.” Data suggest that raters tend to underestimate the extent to which victims perceive the strains they experi- ence as unjust (see Mikula 1986), so these measurement strategies will likely provide conservative estimates of perceived injustice.
The perceived injustice of subjective strains can be estimated by asking victims to rate the injustice of the strains they have experienced. Such ratings will reflect both the characteristics of the strains and the characteristics of the victims. Most notably, victims with attributional biases of the type described by Dodge and Schwartz (1997) will be more likely to rate given strains as unjust. Studies focusing on subjective strains should therefore control for rel- evant individual characteristics when examining the effect of the perceived injustice of strain on crime (see Herbert and Cohen 1996; Turner and Wheaton 1995).2
Factors influencing perceptions of injustice. It is important for GST to describe why some strains are more likely to be perceived as unjust than oth- ers. This allows researchers to better explain individual and group differences in perceptions of injustice, better predict whether given strains will be seen as unjust, and better develop policies that address perceptions of injustice. Sev- eral literatures devote much attention to the factors influencing perceptions of injustice, with the justice and attributions literature being most relevant (for overviews, see Crittenden 1983, 1989; Hegtvedt and Cook forthcoming; Hegtvedt and Markovsky 1995; Mikula 1986, 1993; Tedeschi and Felson 1994; Tedeschi and Nesler 1993; Tyler 1990; Tyler et al. 1997).
328 JOURNAL OF RESEARCH IN CRIME AND DELINQUENCY
These literatures suggest that a strainful event or condition is most likely to be seen as unjust when individuals believe that it involves the voluntary and intentional violation of a relevant justice norm. This belief is influenced by a range of individual characteristics, most of which are described in the justice and attributions literature and by the nature of the strainful event or condition. Most strainful events and conditions involve a perpetrator who does something to a victim in a particular setting or collection of settings. The likelihood that a strainful event will be seen as unjust partly depends on the characteristics of the perpetrator and victim, what the perpetrator does to the victim, what the victim does to the perpetrator, the relationship between the perpetrator and victim, and the setting(s) in which the strain occurs. Per- ceptions of injustice are also influenced by the interpretation of the event/ condition provided by others, especially trusted others, and by (sub)cultural beliefs associated with the event/condition. The contribution of these factors is described below, with the central point being that some strainful events and conditions are more likely than others to be perceived as unjust—holding individual characteristics constant.
Voluntary/intentional. Strainful events and conditions are most likely to be attributed to the voluntary, intentional behavior of others when the follow- ing occurs:
1. There is good evidence that the victim’s strain was in large measure caused by the behavior of others (as opposed to being caused by the victim’s own behav- ior, bad luck or chance, natural/impersonal forces, or forces of uncertain ori- gin). Such evidence includes the following: A perpetrator directly inflicts the strain on the victim (e.g., punches or insults the victim), a perpetrator is identi- fied by trusted others, and/or (sub)cultural beliefs attribute the victim’s strain to the behavior of others.
2. There is good evidence that the perpetrator voluntarily intended to inflict the strain (i.e., freely chose to treat the victim in a way that they knew would proba- bly be disliked). Conversely, there is little evidence that the behavior of the per- petrator was the result of constraint, reasonable accident, or reasonable igno- rance. Such evidence includes the following:
Behavior of the perpetrator. The perpetrator states his or her intention to inflict strain, as sometimes happens in cases involving physical and verbal assault. The perpetrator devotes much effort to or incurs high costs in inflicting the strain. The perpetrator violates normative expectations in inflicting strain. The perpetrator does not excuse, apologize for, or express remorse over the harm he or she has caused. Conversely, the perpetrator expresses pleasure or pride over his behavior (see Averill 1993; Tedeschi and Felson 1994; Tedeschi and Nesler 1993).
Agnew / GENERAL STRAIN THEORY 329
Severity of harm. Attributions of intent are more likely the greater the actual or intended harm to the victim (see Tedeschi and Felson 1994; Tedeschi and Nesler 1993).
Characteristics of the perpetrator and the relationship between the perpetrator and victim. The perpetrator has the personal and social resources to volun- tarily and intentionally inflict the strain (e.g., has sufficient power, is aware of the harmful consequences of his or her behavior). The perpetrator has a known history of intentionally harming the victim or others. The perpetra- tor is disliked by the victim or has a negative reputation, making attribu- tions of malicious intent more likely. This dislike/negative reputation may be related to the characteristics of the perpetrator (e.g., race, gang member- ship; see Tedeschi and Felson 1994; Tedeschi and Nesler 1993).
Audience reaction. Others, especially trusted others, tell the victim that the harm inflicted by the perpetrator was intentional.
(Sub)cultural beliefs or causal schema. (Sub)cultural beliefs or causal schema define the strainful event or condition as one that is usually the result of intent.
Criterion 1 is necessary for attributions of intent, and the factors under cri- terion 2—although not necessary—substantially increase the likelihood of attributions of intent.3
The violation of relevant justice rules. Voluntary and intentional efforts to inflict strain are not necessarily seen as unjust. For example, parents, teach- ers, employers, and the police voluntarily and intentionally inflict strain on a routine basis, but the victims of such strain often do not view the actions of these others as unjust. The intentional infliction of strain is most likely to be seen as unjust when it is believed to violate a relevant justice norm. We must consider norms related to distributive, procedural, interactional, and retribu- tive justice (for overviews, see Hegtvedt and Markovsky 1995: Mikula 1993; Mikula et al. 1990; Tedeschi and Felson 1994; Tyler et al. 1997).4 Drawing on the justice literature, as well as the related literature on the techniques of neu- tralization/rules of moral disengagement (Bandura 1990; Sykes and Matza 1957), it can be argued that the voluntary and intentional infliction of strain is likely to be seen as unjust to the extent that
A. Victims believe their strain is undeserved. In the United States, victims are more likely to believe that their strain is deserved if it is the result of negatively valued behavior on their part (e.g., a child is punished for misbehaving) or if it is the result of the possession of certain negatively evaluated characteris- tics—usually achieved characteristics—deemed relevant in the particular situ- ation (e.g., a job applicant gets turned down because he or she does not possess
330 JOURNAL OF RESEARCH IN CRIME AND DELINQUENCY
relevant work experience). Furthermore, the strain must not be excessive given the negatively evaluated behavior or characteristics of the victim. Violations of these conditions foster the impression that strain is undeserved.
B. Victims believe their strain is not in the service of a higher cause or author- ity—such as God, country, or gang. The infliction of strain is often justified by appeals to higher purposes or authorities; for example, nations may ask indi- viduals to serve in combat to protect their country or gangs may ask members to risk injury for the protection of “turf.”
C. Victims believe their strain will result in much net harm to them. The infliction of strain is often justified by claiming that the strain was minor or negligi- ble—in absolute and/or relative terms. Victims, for example, may be told that they suffered little actual harm or that they suffered much less harm than simi- lar others. Perpetrators may also justify the strain they inflict by claiming that victims will achieve a net benefit from it. Parents, for example, may claim that they need to limit the autonomy of their children to protect them from greater harm. Such claims are most likely to be made and accepted in settings in which personal welfare and development are major goals.
D. Victims believe that the process used to decide whether to inflict their strain was unjust. Victims are more likely to make attributions of procedural injustice when (1) they have no voice in the decision to inflict their strain, (2) they do not accord legitimacy to those who inflict their strain, (3) they do not trust those who inflict their strain—believing they are biased or dishonest, (4) they believe that those inflicting their strain do not make use of accurate or complete infor- mation, (5) they believe that different procedures are followed for similar oth- ers, (6) they are not treated in a polite or respectful manner, (7) the deci- sion-making process is incompatible with fundamental moral and ethical values, (8) no rationale is given for the decision that was made, and/or (9) there are no mechanisms available to correct bad decisions.
E. The strain involves treatment perceived as disrespectful, inconsiderate, or aggressive.
F. The strain violates strongly held social norms, especially those embodied in certain criminal laws (see Tyler et al. 1997).
Perceptions of injustice are likely if criteria A, B, and C are satisfied (hav- ing to do with distributive justice); criterion D is satisfied (procedural jus- tice); criterion E is satisfied (interactional justice); and/or criterion F is satis- fied (retributive justice). The characteristics of the strainful event/condition often allow us to roughly judge the likelihood that victims will hold the be- liefs listed in criteria A through E. For example, a criminal victimization is more likely to generate the beliefs outlined above than is the failure of a poorly educated person to obtain a highly paid job. In addition, the justice lit- erature suggests that we pay special attention to the following factors when trying to estimate the likelihood that individuals hold the above beliefs (e.g., believe their strain is undeserved or involves disrespectful treatment):
Agnew / GENERAL STRAIN THEORY 331
1. Do (sub)cultural beliefs define the strain as just or as unjust for one or more of the reasons listed in A through F? Laws defining the strainful treatment as ille- gal are especially relevant, particularly when the treatment involves the viola- tion of criminal laws with severe penalties.
2. Do others, especially trusted others, support or hinder the adoption of the above beliefs (e.g., tell the victim that their negative treatment is undeserved or disre- spectful)? The actions of family and friends, audience members who witness the negative treatment, and the perpetrator of the negative treatment are espe- cially important. For example, victims are less likely to adopt the above beliefs if the perpetrator is a trusted other who offers a convincing justification for their behavior (see Crittenden 1989).
3. Is the victims’s negative treatment very different from their past treatment in similar circumstances and/or from the treatment of similar others? Compari- sons to past treatment and to the treatment of similar others are especially im- portant in situations in which there are no strong standards defining what is just or fair. Comparison others may include specific others, groups, or more gener- alized others or “referential structures” (see Hegtvedt and Markovsky 1995). Unfortunately, it is often difficult to predict the comparison others that are se- lected and the nature of the comparison process—although some progress is being made in this area (see Hegtvedt and Cook forthcoming).
The Strain Is Seen as High in Magnitude
A second factor influencing the likelihood that strainful events and condi- tions will lead to crime is the perceived magnitude of the strain. Strain that is high in magnitude influences the ability to cope in a noncriminal manner, the perceived costs of noncriminal versus criminal coping, and the disposition to engage in criminal coping. It is more difficult to cognitively minimize the impact of severe strain, emotional coping techniques of a noncriminal nature may be less effective, and behavioral coping of a noncriminal nature may be more difficult (e.g., it is more difficult to legally cope with a large rather than small financial problem). Furthermore, not only is it more difficult to legally cope with severe strain, but such strain often reduces the ability to cope. For example, the victims of severe strain are more likely to suffer from depres- sion, which impedes their ability to cope. Finally, severe strain generates more anger and so also influences the perceived costs of crime and the dispo- sition to engage in crime.
Measuring magnitude. The magnitude of objective strains can be esti- mated by (1) researchers, with such researchers taking account of the factors listed below; (2) a panel of judges familiar with the group, with these judges being asked to estimate the extent to which various strains are likely to be dis- liked (or seen as undesirable, harmful/threatening, etc.); and/or (3) a
332 JOURNAL OF RESEARCH IN CRIME AND DELINQUENCY
representative sample of group members, with such members being asked to rate the extent to which they dislike various strains. Again, it is best to provide individuals with specific information about the strains being rated.
The magnitude of subjective strains can be estimated by asking victims to rate the extent to which they dislike the strains they have experienced. These ratings will reflect both the characteristics of the strains and the characteris- tics of the victims. In particular, the same strainful event/condition might be seen as high in magnitude by one victim but low by another—depending on such things as the victim’s goals/activities/identities, coping ability and resources, and level of social support (see Cohen et al. 1995; Kessler et al. 1995; Lazarus 1999; Taylor and Aspinwall 1996; Thoits 1995; Wheaton 1996). Studies focusing on subjective strains should therefore control for rel- evant individual characteristics when examining the effect of the perceived magnitude of strain on crime (see Herbert and Cohen 1996 and Turner and Wheaton 1995 for a fuller discussion).
Factors influencing perceptions of magnitude. Drawing on Agnew (1992) and the stress literature, there is reason to believe that several features of the strainful event/condition influence perceptions of magnitude. These include the degree or amount of strain inflicted; the duration and frequency of the strain, including the expected duration into the future; the recency of the strain; and the centrality of the strain, which refers to the extent to which the strain threatens the core goals, needs, values, activities, and/or identities of the vic- tim. At present, it is unclear how these factors combine to influence overall judgments of magnitude.
Degree of strain. The degree or amount of strain inflicted influences judg- ments of magnitude. As Agnew (1992) pointed out, it is sometimes possible to measure the degree of strain inflicted in terms of a standard metric, like the severity of the physical injuries inflicted or the amount of money lost. This is not possible for many types of strain, however. Furthermore, the metrics used to measure the degree of strain vary from one type of strain to another, mak- ing it difficult to make comparisons across types of strain. These problems can be dealt with using techniques from the stress research. Individuals can rate the degree or amount of strain inflicted for different types of strain using a common scale. Such ratings likely reflect the objective characteristics of the strain (e.g., amount of money lost, injury inflicted), (sub)cultural beliefs regarding the degree of strain (e.g., beliefs regarding what is a small versus large financial loss, a minor versus serious insult), audience reactions to the strain (where applicable), and individual characteristics (especially when dealing with subjective strains).
Agnew / GENERAL STRAIN THEORY 333
Duration/frequency of strain. The duration and frequency of strain are also likely to influence the perceived magnitude of strain. As Agnew (1992) pointed out, data suggest that strains of long duration (chronic stressors) and/or high frequency have a greater negative impact on the individual (also see Lepore 1995; Turner and Wheaton 1995). Furthermore, data from the stress literature suggest that unresolved strains have a much greater impact on the individual than resolved strains (Herbert and Cohen 1996; Turner and Avison 1992). It is therefore important to determine whether the strain has been resolved. If the strain has not been resolved, it is also important to esti- mate its expected duration. That is, will the strain be resolved shortly or con- tinue for some time, perhaps increasing in frequency and/or degree? The importance of estimating the expected duration of strain is illustrated in the work of Anderson (1999). As Anderson emphasized, seemingly trivial strains like a negative remark or a stare often generate much distress among inner-city residents, partly because they signal future conflicts of a more seri- ous nature.
Recency. As Agnew (1992) noted, the impact of strains or stressors dissi- pates over time. Therefore, recent strains should have a larger impact on judg- ments of magnitude than older strains. At the same time, it is important to note that severe childhood strains may sometimes contribute to later criminal behavior (Elder et al. 1996; Kessler et al. 1995, 1997; Widom 1998).
Centrality of strain. Two individuals might be similar to one another in the degree, duration/frequency, and recency of their strain; yet they may differ dramatically in the perceived magnitude of their strain. One reason for this has to do with the centrality of the strain: Does the strain threaten the core goals, needs, values, activities, and/or identities of the individual? For exam- ple, two individuals may perceive the same monetary loss differently because they differ in the value they place on money.
Centrality is conceived of in different ways depending on the researcher and/or research tradition. Classic strain theorists, frustration-aggression the- orists, and certain stress researchers focus on the importance of the goals, needs, or terminal values that are blocked or threatened (Berkowitz 1993; Cloward and Ohlin 1960; Cohen 1955; Dohrenwend 2000; Kaplan 1996; Lazarus 1999; Merton 1938; Wethington, Brown, and Kessler 1995). Certain stress researchers focus on the extent to which the strain leads to change (or negative change) in the usual or core activities of daily life (e.g., Dohrenwend 1998; Wheaton 1996). Still others—including strain, stress, social interactionist, and identity theorists—focus on the extent to which strains threaten the core identities of individuals or threaten efforts to establish posi- tive identities (e.g., Berkowitz 1993; Burke 1996; Cohen 1997; Kaplan 1996;
334 JOURNAL OF RESEARCH IN CRIME AND DELINQUENCY
Tedeschi and Felson 1994; Thoits 1991; Tyler 1994; Tyler et al. 1997; Wheaton 1996). These perspectives overlap to a large degree (see Burke 1996; Dohrenwend 1998; Kaplan 1996). For example, one’s core identities are in large measure defined in terms of one’s goals, values, and activities. In any event, GST can accommodate all these perspectives: Strain is central to the extent that it threatens core goals, needs, values, activities, and/or identities.
Judgments regarding the centrality of strain are partly influenced by the characteristics of the strain. For example, certain strainful events/conditions are such that they threaten a broad range of goals, values, needs, identities, and activities, so they are likely to be high in centrality for the overwhelming majority of people who experience them. Examples include “extreme stress- ors” (Dohrenwend 1998, 2000) and “traumatic events” (Wheaton, Roszell, and Hall 1997). As Dohrenwend (2000) stated, extreme stressors are such that “all usual activities are disrupted and all of the individual’s goals are in jeopardy” (p. 8). Judgments regarding centrality are also influenced by the (sub)cultural beliefs associated with the strainful event/condition and how the event/condition is interpreted by others (e.g., whether audience members define an insult as trivial or a serious challenge to one’s manhood).
The Strain Is Caused by or Associated with Low Social Control
A third factor affecting the likelihood that strain will lead to crime is the level of social control associated with the strain. Certain strains are caused by or associated with low social control, such as the strain caused by erratic parental discipline (low direct control), parental rejection (low attachment), work in the secondary labor market (low commitment), or homelessness (low direct control, attachment, and commitment). Such strains are more likely to result in crime because the low social control associated with them reduces the costs of crime. Also, low social control may reduce the ability to cope in a noncriminal manner. Individuals low in direct control, conventional attach- ments, and conventional commitments generally lack the social supports and resources that facilitate noncriminal coping.
Conversely, certain strains stem from or are associated with high social control. For example, much adolescent strain stems from the efforts of par- ents to supervise their children (direct control), much parental strain stems from the demands associated with childcare (attachment), and much occupa- tional strain stems from the long working hours and difficult tasks associated with many professional/business jobs (commitment). Such strains are less likely to result in crime because the high social control associated with them increases the costs of crime. High social control may also increase the ability to cope in a noncriminal manner. High control is frequently associated with
Agnew / GENERAL STRAIN THEORY 335
the provision of social support and the possession of personal and financial resources that facilitate noncriminal coping.
An excellent illustration of the association between strain and social con- trol is provided in Hirschi’s (1969) and Kornhauser’s (1978) discussion of classic strain theory. Classic strain theorists focus on one type of strain: the inability to achieve conventional success goals—like educational and occu- pational success—through legitimate channels. Hirschi and Kornhauser argued that the pursuit of such goals implies some level of social control. As Kornhauser stated,
if the child is sufficiently socialized to have a strong desire for conventional goals, he should be well enough socialized also to have the internalized values governing the conventional means of achieving them. . . . He should also be strongly enough attached to conventional persons and institutions to resist the temptation to use nonnormative means. (P. 47)
The pursuit of conventional success goals therefore implies at least moder- ately high levels of attachment, commitment, and belief (in conventional norms). And this may explain why the inability to achieve educational and occupational goals is unrelated to crime in most studies (Agnew 1995a).
Measuring social control. Researchers should estimate the extent to which the type of strain being examined is associated with (1) supervision or direct control by conventional others, (2) attachment to conventional others, (3) commitment to conventional institutions, and (4) the acceptance of con- ventional beliefs, especially beliefs condemning crime. This is easily done in certain cases; for example, the strain being examined stems from or is associ- ated with employment in prestigious, well-paid jobs that indicate a strong commitment to conventional society. In other cases, researchers can employ observational or survey research to determine the association between strain and social control. For example, survey data can be used to determine whether individuals who desire educational and occupational success are high in such types of social control as attachment to conventional others and beliefs condemning crime.
The Strain Creates Some Pressure or Incentive to Engage in Criminal Coping
A final factor affecting the likelihood that strain will lead to crime is the extent to which the strain creates some pressure or incentive to engage in criminal coping. Drawing on social-learning and routine activities theories, it can be argued that the type of strain experienced influences the availability
336 JOURNAL OF RESEARCH IN CRIME AND DELINQUENCY
and appeal of noncriminal and criminal coping options—thereby affecting the pressure/incentive to engage in crime. Certain types of strain are associ- ated with exposure to others who model crime, reinforce crime, and/or pre- sent beliefs favorable to crime (e.g., child abuse, being bullied by peers). More directly, certain types of strain are associated with exposure to others who model criminal coping to that type of strain and present beliefs favorable to criminal coping to that type of strain. Furthermore, criminal coping may be the only or the most effective way to address the perceived injustice and reduce the perceived magnitude of that type of strain (see Brezina 2000). Anderson’s (1999) discussion of life in a poor, inner-city community pro- vides an example.
Anderson (1999) argued that young males in this community are under much pressure to respond to one type of strain—disrespectful treatment— with violence. The perpetrators of disrespectful treatment and others in the community frequently model and present beliefs favorable to criminal cop- ing. And violence is often the only effective way to respond to disrespectful treatment. Efforts to ignore disrespectful treatment or reason with the perpe- trators of such treatment often result in further abuse—by both the perpetra- tor and others in the community. Victims cannot rely on the police or others to intervene on their behalf (also see Black 1983). And the efforts of victims to cognitively reinterpret their strain or engage in emotional coping are also ineffective. The perpetrators of the strain typically escalate their level of abuse, others regularly remind the victim of the disrespectful treatment they have experienced, and subcultural beliefs define such treatment as unjust and high in magnitude. Cognitive reinterpretation is therefore difficult. Violent coping, however, reduces feelings of injustice, reduces the likelihood of fur- ther disrespectful treatment, and allows the victim to protect or enhance their identity/status.
Measuring the pressure or incentive for criminal coping. Researchers should consider the following factors when determining whether a particular instance of strain creates some pressure or incentive for criminal coping.
1. Does the strain stem from or is it associated with exposure to others who model, reinforce, and/or present beliefs favorable to crime?
2. What behavioral options of a noncriminal and criminal nature are available to members of the group experiencing the strain in question? Are these options frequently modeled by others? Do they have (sub)cultural support? How effec- tive will these options be in reducing the perceived injustice and magnitude of the strain?
3. What cognitive options of a noncriminal criminal nature are available to mem- bers of the group experiencing the strain? Efforts to cognitively cope with
Agnew / GENERAL STRAIN THEORY 337
strain usually involve attempts to minimize the injustice, degree, duration/fre- quency, recency, and/or centrality of the strain. It is more difficult to cognitively minimize the injustice and/or magnitude of some types of strain than others. In particular, minimization is more difficult when (1) the victim re- ceives clear and frequent information on the injustice and magnitude of their strain, with this information coming from such sources as trusted others, wit- nesses to the strainful event or condition, and members of the community; (2) (sub)cultural beliefs define the strain as unjust and high in magnitude; (3) there is strong (sub)cultural and structural support for the goals, needs, val- ues, activities, and/or identities being challenged; and (4) the strain is unre- solved, perhaps increasing in frequency and/or degree.
Information in the above areas can be obtained from observational studies, intensive interviews, and surveys.
CLASSIFYING TYPES OF STRAIN ACCORDING TO THEIR LIKELIHOOD OF LEADING TO CRIME
In sum, strainful events and conditions are most likely to lead to crime when they (1) are seen as unjust, (2) are seen as high in magnitude, (3) are associated with low social control, and (4) create some pressure or incentive for criminal coping. At present, I would argue that all four of these character- istics are roughly equal in importance and that the absence of any one charac- teristic substantially reduces the likelihood that strain will result in crime—unless the strain is seen as extraordinarily unjust and high in magni- tude (see below). These characteristics are next used to predict the relative likelihood that different types of strain will result in crime. Drawing on the existing research where possible, I roughly estimate the likelihood that these strains are seen as unjust, are seen as high in magnitude, are associated with low social control, and create some pressure or incentive for criminal coping. It would of course be desirable to verify my judgments using the research strategies described above.
It is not possible in this short article to make predictions for all types of strain. Instead, I consider several broad types of strain. These types of strain were selected for several reasons: They encompass many of the major types of strain that people face—including family, peer, school, and work-related strains; they include most of the strains examined in tests of classic and GST, as well as certain strains neglected by empirical researchers; and most of these strains can be examined with currently available data sets. The focus on broad types of strain, however, does reduce the accuracy of the predictions. As indicated, it is more difficult to classify broadly defined types of strain on
338 JOURNAL OF RESEARCH IN CRIME AND DELINQUENCY
the above characteristics. For example, it is difficult to predict whether unem- ployment will be related to crime. As many researchers argue, the relation- ship between unemployment and crime depends on the circumstances associ- ated with the unemployment. Limited evidence suggests that unemployment is most likely to lead to crime when it is persistent (i.e., high in magnitude) and blamed on others (i.e., seen as unjust) (see Baron and Hartnagel 1997; Box 1987; Colvin 2000; Hagan and McCarthy 1997; Uggen 2000). As a result, the strains below are simply sorted into two groups: those predicted to be unrelated or weakly related to crime and those predicted to be more strongly related to crime.
Types of Strain Unrelated or Weakly Related to Crime
The first condition (strain seen as unjust) allows us to predict that a wide range of strains will be unrelated to crime. At the most general level, these include those types of strain that are clearly the result of reasonable accident or chance, reasonable ignorance, reasonable constraint, the victim’s own behavior, or natural causes like extreme weather and disease (as opposed to those types of strain resulting from the voluntary and intentional violation of justice norms). Many of the strains commonly included in the stressful life events scales used to test GST likely fall into this category, like accident, seri- ous illness or injury, serious illness or injury of brother or sister, brother or sister leaving home for college or a job, and family member dying.
The second condition (strain seen as high in magnitude) allows us to pre- dict that strains that are low in magnitude will be unrelated to crime. Certain types of strain are more likely to be seen as low in magnitude than other types. For example, those strains that threaten peripheral goals are more likely to be seen as low in magnitude than those that threaten core goals. There are data ranking the importance of various goals in the United States as a whole and among certain groups (e.g., Rokeach 1973). Such data can be used as a guide in predicting the likelihood that specific strains will be seen as low or high in magnitude (more below). At the same time, it is important to note that many seemingly serious strains—like the death of a family member—may be per- ceived as low or high in magnitude depending on the circumstances (see Wethington et al. 1995; Wheaton 1990). So, it is important for researchers to estimate the magnitude of the strains they are examining, something that is rarely done in the criminology research.
Considerations of injustice and magnitude—as well the third and fourth conditions (the strain is associated with low social control and creates some pressure or incentive for criminal coping)—allow us to predict that several
Agnew / GENERAL STRAIN THEORY 339
other types of strain will not be related to crime. These include types of strain that have dominated the research on strain theory.
The failure to achieve those goals that result from conventional socializa- tion and that are difficult to achieve through illegitimate channels. These goals include educational success, occupational success, and middle-class status. Although the inability to achieve these goals may result in strain of high magnitude, such strain is unlikely to be seen as unjust. Among other things, the failure to achieve such goals is typically blamed on the victim. As Merton (1968) stated, the cultural system in the United States conveys the message that “success or failure are wholly results of personal qualities; that he who fails has only himself to blame” (p. 222; also see Merton 1968:191, 201-03). And much research on the legitimation of stratification suggests that people tend to accept responsibility for their place in the stratification system (see Agnew 1992; Hegtvedt and Markovsky 1995; see below for the argu- ment that minority-group members in the United States may sometimes blame others for their failure to achieve conventional success goals). Further- more, as argued above, the pursuit of conventional success goals implies some level of social control. Finally, the inability to achieve these goals is not likely to create strong pressure for criminal coping. In particular, these goals are not easily achieved through criminal means, like theft and violence. In fact, criminal behavior may undermine the achievement of these goals. Therefore, criminal coping is not likely to be reinforced. These arguments may help explain why empirical research typically finds that crime is unre- lated to the disjunction between educational and/or occupational aspirations and expectations (see Agnew 1995a for an overview; Jensen 1995).5
Supervision/discipline by parents, teachers, criminal justice officials, and other conventional authority figures that is (1) not overly strict, (2) consis- tent, (3) contingent on negative behavior, (4) not excessive given the infrac- tion, and (5) not verbally or physically abusive. Such supervision/discipline may generate much strain (e.g., juveniles being grounded, offenders being arrested and sent to prison). But this strain is not likely to be seen as unjust because it is deserved, is administered in a fair way by legitimate authority figures, and is not aggressive or disrespectful. Furthermore, such supervi- sion/discipline creates a high level of direct control and reduces the likeli- hood of association with delinquent others. Much data demonstrate that parental and school supervision/discipline of the above type is associated with lower levels of delinquency (Agnew 2001b; Sampson and Laub 1993). And some data suggest that this may be true for supervision/discipline by criminal justice officials as well (Lanza-Kaduce and Radosevich 1987; Sherman 1993, 2000; Tyler 1990).
340 JOURNAL OF RESEARCH IN CRIME AND DELINQUENCY
The burdens associated with the care of conventional others to whom one likely has a strong attachment, like children and sick/disabled spouses. Although such care may create great strain, it is not likely to be viewed as unjust. There is a strong cultural expectation that one is supposed to care for children and sick/disabled spouses, an expectation likely to be supported by others in the person’s network. In fact, one is usually labeled a bad parent or spouse if such care is not provided. This type of strain implies at least a mod- erate level of social control: the “victim” may be closely supervised by others inside and outside the family, the victim likely has a strong emotional bond to conventional others, and cultural beliefs strongly support the provision of adequate care. This type of strain also does not create much pressure or incen- tive for most forms of criminal coping. Caregivers have little opportunity to engage in crime, except for family violence, neglect, certain types of illicit drug use, and possibly shoplifting. Crime is not an effective solution to this type of strain. And the burdens associated with care giving limit association with criminal others.
The impact of this type of strain on crime has not been well examined. Data from the stress literature, however, indicate that females are more likely than males to experience this type of strain (see Broidy and Agnew 1997). This may partly explain gender differences in crime. It may also help explain why such differences are smallest for the crimes of family violence, larceny, and certain types of illicit drug use, such as the misuse of prescription drugs.
The excessive demands associated with conventional pursuits that provide rewards like high pay, prestige, and/or intrinsic satisfaction (or that have a strong likelihood of providing access to such rewards in the future). The prime examples of such pursuits are work in prestigious and/or well-paid jobs (or work in the primary labor market) and attending college. Excessive demands include long working (or studying) hours and work on difficult tasks. Such strain may be seen as high in magnitude, but it is not likely to be seen as unjust. The voluntary or quasi-voluntary nature of these conventional pursuits contributes to self-blame, and the victims of such strain may feel that the excessive demands made on them are justified or offset by the rewards they receive. Such strain is frequently caused by or associated with high social control, including commitment to conventional activities (e.g., one’s job or educational pursuits) and supervision (i.e., much time is spent on struc- tured tasks that are closely monitored). And such strain does not create pres- sure or incentives for criminal coping. The excessive demands limit the opportunity for association with criminal others. Furthermore, crime is typi- cally not an effective solution to such demands (with the exception of cheat- ing and certain types of white-collar crime). This type of strain has not been
Agnew / GENERAL STRAIN THEORY 341
well examined, although we do know that time spent studying is negatively related to crime (Agnew 2001b; Hirschi 1969).
Unpopularity with or isolation from peers, especially criminal peers. Such strain may be high in magnitude and may also be seen as unjust. In par- ticular, individuals may blame their unpopularity/isolation on peers who unfairly reject them or on parents who unfairly limit their social life. Such strain, however, may contribute to an increase in social control by increasing time spent with parents or other conventional figures. Also, such strain does not create much pressure or incentive for crime. Little time is spent with peers who may reinforce crime, model crime, and foster beliefs conducive to crime. And related to this, there are fewer opportunities for crime. Data sup- port this prediction: Crime is less common among juveniles who report that they are unpopular with peers, have few close friends, have few delinquent friends, never or seldom date, or seldom engage in unsupervised social activi- ties with peers (Agnew 2001b; Agnew and Brezina 1997; Agnew and Petersen 1989; Osgood et al. 1996).
Isolation from those situations or environments conducive to crime. Such strain is closely related to strain from unpopularity or isolation from peers because these situations/environments are typically settings where unsuper- vised peers gather. Likewise, this type of strain often stems from peer rejec- tion and the efforts of parents to supervise their children (e.g., setting cur- fews, prohibiting attendance at parties). This type of strain is unlikely to lead to crime for the reasons indicated in the previous paragraph. Data support this prediction (Agnew 2001b; Agnew and Petersen 1989; Osgood et al. 1996). And the fact that females are more likely to experience this type of strain than males may help explain gender differences in crime (Broidy and Agnew 1997; Jensen and Brownfield 1986; Osgood et al. 1996).
Extreme instances of the above types of strain. Certain of the above types of strain may lead to crime in extreme cases. Extraordinary demands at work or school and extraordinary demands for the care of conventional others may be viewed as unjust because they are far outside the range of past experience or the experience of similar others, they may severely tax efforts at conven- tional coping, and they may eventually undermine conventional attachments and commitments. As such, these extraordinary demands may lead to crime. This argument finds indirect support in the work of Wells and Rankin (1988), who found a curvilinear relationship between parental supervision and delin- quency. Increases in parental supervision up to a point reduce delinquency, but very strict supervision increases delinquency.
342 JOURNAL OF RESEARCH IN CRIME AND DELINQUENCY
Types of Strain More Strongly Related to Crime
As indicated, strainful events and conditions are unlikely to lead to crime unless they are (1) seen as unjust, (2) seen as high in magnitude, (3) associ- ated with low social control, and (4) create some pressure or incentive for criminal coping. Such strains are likely to include (but are not limited to) the following.
The failure to achieve core goals that are not the result of conventional socialization and that are easily achieved through crime. Such goals include money—particularly the desire for much money in a short period of time (as opposed to the gradual accumulation of savings), thrills/excitement, high lev- els of autonomy, and masculine status (see Agnew 1997, 2001a, 2001b; Agnew et al. 1996; Anderson 1999; Cernkovich et al. 2000; Colvin 2000; Katz 1988; Matza and Sykes 1961; Messerschmidt 1993; Moffitt 1993; Tittle 1995). These are core goals for at least certain segments of the population. It is difficult to predict whether the failure to achieve these goals will be seen as unjust, although it has been suggested that this is the case where barriers to success are visible and such barriers involve discrimination based on acquired characteristics—like “the mere fact of birth into a particular race, religion, social class, or family” (Cloward and Ohlin 1960:119; also see Anderson 1999; Blau and Blau 1982; Messerschmidt 1993). The pursuit of these goals does not imply conventional socialization or high social control. Rather, the pursuit of these goals frequently stems from the possession of cer- tain individual traits, like sensation seeking (White, Labouvie, and Bates 1985), exposure to “subterranean’ traditions” or subcultural groups (see Matza and Sykes 1961), and structural conditions—like poverty in the midst of plenty (see Kornhauser 1978). In this area, Cernkovich et al. (2000) dem- onstrated that the desire for material success does not function as a form of social control. Furthermore, these goals—unlike educational and occupa- tional success—are easily achieved through crime. Crime is frequently used to get money (Agnew et al. 1996; Cernkovich et al. 2000; Colvin 2000), obtain thrills/excitement (Katz 1988), demonstrate or obtain autonomy (Agnew 1984; Moffitt 1993; Tittle 1995), and “accomplish” masculinity (Anderson 1999; Messerschmidt 1993).
Parental rejection. Parents who reject their children do not express love or affection for them, show little interest in them, provide little support to them, and often display hostility toward them. Parental rejection is likely to create much strain because it may seriously threaten many of the child’s goals, val- ues, needs, activities, and/or identities. Parental rejection is likely to be seen
Agnew / GENERAL STRAIN THEORY 343
as unjust given cultural expectations and the experiences of other children. Parental rejection is associated with very low rather than high social control. And rejection creates some pressure or incentive to engage in crime, largely because rejected children are more likely to be exposed to deviant/aggressive behaviors by their parents and associate with delinquent peers. Data indicate that parental rejection is strongly related to delinquency (Agnew 2001b; Sampson and Laub 1993).
Supervision/discipline that is very strict, erratic, excessive given the infraction, and/or harsh (use of humiliation/insults, threats, screaming, and/or physical punishments). Such supervision/discipline is likely to be seen as high in magnitude, particularly if the individual is exposed to it on a regular basis by parents, school officials, criminal justice officials, or others. It is likely to be seen as unjust because it violates one or more justice norms. It is associated with low social control; sanctions administered in the above manner do not function as effective direct controls, and they frequently undermine attachments and commitments to conventional others and institu- tions. Such supervision/discipline also creates some pressure or incentive for crime because the sanctioning agents frequently model aggressive behavior, implicitly or explicitly foster beliefs conducive to aggression, and sometimes reinforce aggression (see Patterson, Reid, and Dishion 1992). This type of discipline is also likely to promote association with delinquent peers. Data indicate that parents, school officials, and possibly criminal justice officials who employ this type of discipline/supervision increase the likelihood of crime (Agnew 2001b; Colvin 2000; Lanza-Kaduce and Radosevich 1987; Patterson et al. 1992; Sampson and Laub 1993; Sherman 1993, 2000; Tyler 1990).
Child neglect and abuse. Child neglect and abuse represent extreme forms of parental rejection and harsh parental discipline, and abuse/neglect should be related to crime for all the reasons listed above for these forms of strain. Data support this prediction (Smith and Thornberry 1995; Widom 1998).
Negative secondary school experiences. Negative school experiences include low grades, negative relations with teachers (e.g., teachers treat unfairly, belittle/humiliate), and the experience of school as boring and a waste of time. These experiences are likely to be seen as high in magnitude given the central role that school plays in the lives of juveniles. They may be seen as unjust. The compulsory nature of school and the dependent status of juveniles contribute to external blame. Also, juveniles may feel that school personnel ask much of them (several hours of their time and attention each
344 JOURNAL OF RESEARCH IN CRIME AND DELINQUENCY
day) but give little in return—which contributes to feelings of distributive injustice. Feelings of injustice are especially likely when students believe they are discriminated against because of ascribed characteristics. Negative school experiences are associated with low rather than high social control. And negative experiences may foster association with delinquent peers. Data indicate that negative school experiences are related to delinquency (Agnew 2001b; Sampson and Laub 1993).
Work in the secondary labor market. Such work commonly involves unpleasant tasks (e.g., simple, repetitive work; physically demanding work; work that requires a subservient stance), little autonomy, coercive control (e.g., threats of being fired), low pay, few benefits, little prestige, and very limited opportunities for advancement. Furthermore, such work is often intermittent in nature. Such work is likely to create much strain, especially given the central role of work for adults. Such work may be seen as unjust. Although individuals often accept responsibility for their position in the stratification system, the high demands and meager benefits of such work are likely to be seen as unjust by many. Such work is associated with low rather than high social control (Crutchfield and Pitchford 1997). And such work may create some pressure or incentive for criminal coping. Crime is often an effective remedy to the problems associated with work in the secondary labor market. And such work often increases the likelihood of exposure to others who are disposed to crime (Colvin 2000; Crutchfield and Pitchford 1997). Data suggest that work in the secondary labor market is associated with crime (Colvin 2000; Crutchfield and Pitchford 1997).
Homelessness, especially youth homelessness. This type of strain is likely to be seen as very high in magnitude because it represents a major challenge to a broad range of goals, needs, values, activities, and identities. Further- more, homelessness dramatically increases the likelihood that many other types of strain will be experienced, particularly conflicts with and victimiza- tion by others (Baron and Hartnagel 1997; Davis 1999; Hagan and McCarthy 1997). Homelessness may be seen as unfair, particularly among youth—whose homelessness is often the result of parental abuse and neglect (Davis 1999; Hagan and McCarthy 1997). And homelessness is strongly associated with low social control and the social learning of crime, as demon- strated in several recent studies (Baron and Hartnagel 1997; Davis 1999; Hagan and McCarthy 1997). Data indicate that homelessness and its atten- dant problems are associated with crime (Baron and Hartnagel 1997; Hagan and McCarthy 1997).
Agnew / GENERAL STRAIN THEORY 345
Abusive peer relations, especially among youth. Peer abuse has been neglected as a type of strain, although data suggest that it is widespread and that it often has a devastating effect on victims (e.g., Ambert 1994: Lock- wood 1997). Such abuse may involve insults/ridicule, gossip, threats, attempts to coerce, and physical assaults. Peer abuse is likely to be seen as high in magnitude, especially among youth, where peers are of central impor- tance. Peer abuse is likely to be seen as unjust because it frequently violates one or more justice norms (e.g., is excessive given the infraction, is disre- spectful or aggressive, is not administered by legitimate sanctioning agents). Peer abuse is not associated with high social control. Peer abuse among juve- niles, in particular, often occurs away from sanctioning agents like parents and teachers. And such abuse is often associated with some pressure or incen- tive to engage in crime. Peer abuse is especially common in delinquent peer groups and gangs, where the victim is regularly exposed to others who model crime, present beliefs favorable to crime, and reinforce crime (see Agnew 2001b; Colvin 2000). Furthermore, peers often model criminal coping in response to abuse, present beliefs that encourage criminal coping in response to abuse, and differentially reinforce criminal coping in response to abuse.
Criminal victimization. Victimization is typically seen as unjust and high in magnitude. Victimization is not associated with high social control; in fact, victimization is most likely to occur in settings in which social control is low—such as settings where young, unsupervised males gather (Jensen and Brownfield 1986; Lauritsen, Sampson, and Laub 1991; Meier and Miethe 1993). Furthermore, victimization may reduce concern with internal and external sanctions because criminal victimization often provides a justifica- tion for crime in the eyes of the victim and others. Finally, criminal victimiza- tion is often associated with the social learning of crime. Victimization is more common in delinquent peer groups and gangs, and victimization by def- inition involves exposure to a criminal model (Lauritsen et al. 1991). Limited data suggest that criminal victimization is strongly related to criminal offend- ing (see Dawkins 1997; Esbensen and Huizinga 1991; Jensen and Brownfield 1986; Lauritsen et al. 1991; Lauritsen, Laub, and Sampson, 1992; Sampson and Lauritsen 1993).
Experiences with prejudice and discrimination based on ascribed charac- teristics, like race/ethnicity. Data indicate that racial prejudice and discrimi- nation are quite common in the United States (Ambert 1994; Forman, Wil- liams, and Jackson 1997). This type of strain is likely to be seen as unjust and high in magnitude, particularly given the strong cultural emphasis in the United States on egalitarianism. Prejudice/discrimination may reduce social control, particularly attachment and commitment to those individuals and
346 JOURNAL OF RESEARCH IN CRIME AND DELINQUENCY
institutions associated with the prejudice and discrimination. Prejudice/ discrimination may also create some pressure or incentive to engage in crime because the victim is exposed to others who violate strongly held social norms. Data indicate that experiences with prejudice and discrimination con- tribute to psychological distress (Finch, Kolody, and Vega 2000; Schulz et al. 2000), and certain qualitative studies have linked prejudice and discrimina- tion to crime (e.g., Anderson 1999). Quantitative studies, however, have not devoted much attention to experiences with prejudice and discrimination.
Summary
The above list represents the most comprehensive attempt to identify those types of strain that are and are not related to crime. It incorporates and extends the work of classic and contemporary strain theorists. Building on the classic strain theorists, it argues that the inability to achieve certain suc- cess goals—particularly educational and occupational goals—is not related to crime, whereas the inability to achieve other success goals—like the rapid acquisition of much money—is related to crime. The list also includes many of the strains that contemporary researchers have identified—like the denial of autonomy needs (Moffitt 1993; Tittle 1995); threats to masculine status (Anderson 1999; Messerschmidt 1993); disrespectful, unfair, or abusive police practices (Sherman 1993, 2000); and the types of coercion discussed in Colvin’s (2000) theory of differential coercion.6 The list also contains types of strain that have not been extensively discussed in the strain literature— noting which are related to crime and which are not. The general principles listed in the previous section allow us to group all of these strains under one theoretical umbrella.
HOW DO WE TEST THE ABOVE ARGUMENTS?
Strain is most likely to lead to crime when it is seen as unjust, is seen as high in magnitude, is associated with low social control, and creates some pressure or incentive to engage in criminal coping. If these arguments are cor- rect, types of strain that meet these conditions should be more strongly related to crime than types that do not (although the precise relationship between strain and crime is a function of the characteristics of both the strain and the people experiencing the strain). So at the most basic level, researchers should test the above arguments by classifying strains on the above character- istics and then examining the relative impact of these strains on crime. The classification of strains just presented can be used as a staring point for such research. Ideally, researchers should compare the criminal behavior of
Agnew / GENERAL STRAIN THEORY 347
people who have experienced the above strains. As an alternative, researchers can present people with vignettes describing these types of strain and then ask how likely they or others would be to respond to them with crime (see Mazerolle and Piquero 1997 for a model).
The Cumulative Effect of Strain
This strategy for testing strain theory differs from the approach now taken by most researchers, who examine the impact of cumulative measures of strain on crime—with these cumulative measures often containing types of strain that differ widely on the above characteristics. Although researchers should not ignore the argument that strains may have a cumulative effect on crime, it is most important at this point to determine which types of strain are most strongly related to crime. Once this is determined, researchers can then explore the cumulative impact of strain on crime. Cumulative scales can be created by combining those types of strain that have a significant impact on crime—perhaps weighting them by their regression coefficients. A similar strategy has been successfully employed in the stress literature (Herbert and Cohen 1996; Turner and Wheaton 1995; Wheaton et al. 1997; also see Agnew and White 1992). Or researchers can determine whether strains inter- act with one another in their impact on crime through the creation of interac- tion terms (see Wheaton et al. 1997; note the argument that moderate levels of prior stress sometimes reduce the negative effects of current stressors).
Distinguishing Strain from Social Control and Social Learning
Researchers testing GST all confront a major problem: Many of the “strain” measures they use—like low grades or harsh parental discipline— can also be taken as social control or social-learning measures. Researchers usually deal with this problem by assigning some measures to the strain camp, some to the social control camp, and some to the social-learning camp. They then try to justify these assignments, although their arguments are often less than convincing. Agnew (1995c) explained why this is so, noting that most variables have implications for strain, social control, and social- learning theories. Harsh discipline, for example, is often classified as a type of strain, but some claim that it leads to crime by reducing attachment to par- ents or implicitly teaching the child that violence is acceptable under certain conditions (see Brezina 1998). It is therefore difficult to classify an independ- ent variable as a purely strain, social control, or social-learning variable. This article makes the same argument: Most types of strain have implications for social control and the social learning of crime. Furthermore, it is argued that
348 JOURNAL OF RESEARCH IN CRIME AND DELINQUENCY
those types of strain most likely to lead to crime are those that are associated with low social control and the social learning of crime.
This argument raises a major problem: If those types of strain most strongly related to crime are associated with low control and the social learn- ing of crime, how do we know whether these strains affect crime for reasons related to strain, social control, or social-learning theories? Agnew’s (1995c) solution to this problem was to examine the intervening processes described by these theories. Although these theories have many of the same independ- ent variables in common, they differ in terms of their specification of inter- vening processes. Strain theory argues that these variables increase crime through their effect on negative emotions, control theory argues that they lower the perceived costs of crime, and social-learning theory argues that they influence the perceived desirability of crime. A few studies have attempted to examine such intervening processes, and they typically find that the processes associated with all three theories are operative (see Agnew 1985; Brezina 1998).7 Unfortunately, most existing data sets do not allow for the proper examination of these intervening processes (see Schieman 2000 and Stone 1995 for discussions of certain of the problems involved in measur- ing the key negative emotion of anger).
There is a second strategy that may be employed to determine if a strain measure affects crime for reasons related to strain, social control, or social-learning theory. Certain strain measures may affect crime because they reduce social control and/or foster the social learning of crime. As indi- cated, harsh discipline is said to reduce attachment to parents and foster beliefs conducive to violence. In such cases, we can examine the effect of the strain measure on crime while controlling for the relevant social control and social-learning variables. For example, we can examine the effect of harsh discipline on crime while controlling for parental attachment and beliefs con- ducive to violence. Or we can examine the effect of teacher conflicts while controlling for attachment to teachers, attachment to school, and grades. If the strain measure still affects crime after such controls, support for strain theory is increased. This strategy cannot be followed in all cases, however. Certain strain measures—like low grades—directly index the respondent’s level of social control or social learning. Therefore, it is not possible to con- trol for the relevant control or social-learning variables. Also, there is some risk in arguing that the direct effect of the strain measure on crime is best explained by strain theory. Researchers may have failed to control for or properly measure all relevant social control and social-learning variables. And it is possible that the strain measure affects crime for reasons other than those offered by strain, social control, and social-learning theories (e.g., genetic factors may influence both exposure to strain and levels of crime).
Agnew / GENERAL STRAIN THEORY 349
Finally, a third strategy sometimes allows us to determine whether strain variables affect crime for reasons distinct from those offered by social control theory. According to the logic of control theory, neutral relationships with other individuals and groups should have the same effect on crime as negative relationships. For example, a juvenile who does not care about her parents should be just as delinquent as a juvenile who dislikes or hates her parents. Both juveniles are equally free to engage in delinquency; that is, both have nothing to lose through delinquency. According to the logic of strain theory, however, the juvenile who hates her parents should be higher in delinquency than the juvenile who does not care about her parents. This is because the juvenile who hates her parents is under more strain. Her hatred likely stems from unpleasant relations with her parents, and it is stressful to live with peo- ple you hate. This prediction is easily tested with certain data sets, but researchers rarely compare juveniles who dislike/hate their parents with juveniles who neither like nor dislike their parents (see Nye 1958 for an exception). Similar analyses can be conducted in other areas. For example, researchers can compare the criminal behavior of individuals who hate their grades or jobs with those who do not care about their grades or jobs. If strain theory is correct, individuals who hate their grades or jobs should be higher in crime.
None of these strategies allows us to perfectly determine whether strain variables affect crime for reasons related to strain, social control, or social-learning theories, but taken together they can shed much light on this problem.
Measuring Strain
Many current measures of strain are quite simplistic; single-item mea- sures of specific strains are often employed, with these measures providing little information about the magnitude, injustice, or other dimensions of the strain. A similar situation characterizes the stress literature, although stress researchers are starting to collect more detailed information on stressors to better estimate things like their magnitude. For example, some stress researchers have abandoned simple checklist measures and are employing intensive interviews with semistructured probes (see Herbert and Cohen 1996; Wethington et al. 1995; Wheaton 1996). Such techniques were devel- oped because respondents often report trivial stressors when checklist mea- sures are used—even when such checklists attempt to focus on serious stress- ors (Dohrenwend 2000; Herbert and Cohen 1996; Wethington et al. 1995). Also, many stress researchers now recognize that the circumstances associ- ated with the stressor have an important effect on its impact. It is difficult to employ intensive interviews in the large-scale surveys often conducted by
350 JOURNAL OF RESEARCH IN CRIME AND DELINQUENCY
criminologists, but criminologists can do a much better job of measuring strain in such surveys. As an illustration, one need only compare the mea- sures of economic strain typically employed by criminologists with those commonly used in the family research. Economic strain is not simply mea- sured in terms of low income or a two- or three-item index of socioeconomic status. Rather, family researchers examine such things as (1) family per capita income; (2) unstable work history, which includes changing to a worse job, demotions, and being fired or laid off; (3) debt-to-asset ratio; and (4) increases or decreases in family income in the past year. Furthermore, researchers recognize that these types of economic strain do not affect all families in the same way. So, more direct measures of economic strain are sometimes employed as well. For example, parents are asked about the extent to which the family has enough money for clothing, food, medical care, and bills. They are also asked about the changes they have had to make to cope with economic hardship, like moving, taking an additional job, canceling medical insurance, and obtaining government assistance (e.g., Conger et al. 1992; Fox and Chancey 1998; Voydanoff 1990; also see Agnew et al. 1996; Cernkovich et al. 2000).
CONCLUSION
GST is usually tested by examining the effect of selected types of strain on crime. Researchers, however, have little guidance when it comes to selecting among the many hundreds of types of strain that might be examined. And they have trouble explaining why only some of the strains they do examine are related to crime. This article builds on GST by describing the characteris- tics of strainful events and conditions that influence their relationship to crime. As indicated, strains are most likely to lead to crime when they (1) are seen as unjust, (2) are seen as high in magnitude, (3) are associated with low social control, and (4) create some pressure or incentive to engage in criminal coping. Based on these characteristics, it is argued that certain types of strain will be unrelated or only weakly related to crime. Such strains include the failure to achieve educational and occupational success, the types of strain that have dominated the research on strain theory. Such strains also include many of the types of strain found in stressful life events scales, which are commonly used to test GST. And it is argued that other types of strain will be more strongly related to crime, including types that have received much attention in the criminology literature (e.g., parental rejection; erratic, harsh parental discipline; child abuse and neglect; negative school experiences) and types that have received little attention (e.g., the inability to achieve selected goals, peer abuse, experiences with prejudice and discrimination).
Agnew / GENERAL STRAIN THEORY 351
The arguments presented in this article should have a fundamental impact on future efforts to test GST because they identify those types of strain that should and should not be related to crime. And in doing so, these arguments make it easier to falsify GST. Furthermore, these arguments help explain the contradictory results of past research on strain theory; for example, they help explain why the failure to achieve educational and occupational success is usually not related to crime, whereas verbal and physical assaults usually have a relatively strong relationship to crime.
These arguments also have important policy implications. Agnew (1992) argued that two major policy recommendations flow from GST: reduce the exposure of individuals to strain and reduce the likelihood that individuals will cope with strain through crime (by targeting those individual character- istics conducive to criminal coping). This article suggests a third recommen- dation: alter the characteristics of strains in ways that reduce the likelihood they will result in crime. Despite our best efforts, many individuals will be exposed to strain. For example, parents, teachers, and criminal justice offi- cials will continue to sanction individuals in ways that are disliked. We can, however, alter the ways in which these sanctions are administered so as to reduce the likelihood that they will (1) be seen as unjust, (2) be seen as high in magnitude, (3) reduce social control, and (4) create some pressure or incen- tive to engage in crime. In fact, this is one of the central thrusts behind the restorative justice and related movements (see Bazemore 1998; Briathwaite 1989; Sherman 1993, 2000; Tyler 1990). These movements point to ways in which criminal justice officials can increase the perceived justice of sanc- tions, reduce the perceived magnitude of sanctions, sanction in ways that increase rather than reduce social control, and sanction in ways that create lit- tle pressure or incentive for crime. Recommendations in these areas include treating offenders with respect; making them aware of the harm they have caused; giving them some voice in determining sanctions; tempering the use of severe, punitive sanctions; and reintegrating offenders with conventional society through a variety of strategies—like reintegration ceremonies and the creation of positive roles for offenders. Certain parent-training and school-based programs are also structured in ways that reduce the likelihood that strains like disciplinary efforts will be administered in ways that increase the likelihood of criminal coping (see Agnew 1995d, 2001b).
This article, then, extends Agnew’s (1992) GST in a way that substantially improves its ability to explain and control crime. Although Agnew (1992) argues that the reaction to strain is largely a function of individual character- istics, this article argues that the reaction to strain is a function of both indi- vidual characteristics and the characteristics of the strain that is being experi- enced. Strain is most likely to lead to crime when individuals possess characteristics conducive to criminal coping (as described in Agnew 1992)
352 JOURNAL OF RESEARCH IN CRIME AND DELINQUENCY
and they experience types of strain conducive to criminal coping (as described above). This extension of strain theory parallels recent develop- ments in the stress literature. Like Agnew (1992), stress researchers argued that the impact of stressors on outcome variables was largely a function of individual characteristics like coping skills and social support. Stress researchers, however, have increasingly come to realize that stressors do not have comparable impacts on outcome variables. Certain stressors are signifi- cantly related to outcome variables—most often measures of mental and physical health—whereas others are not (e.g., Aseltine et al. 2000; Aseltine, Gore, and Colten 1998; Brown 1998; Wethington et al. 1995; Wheaton et al. 1997; Dohrenwend 1998). So we must consider both the nature of the stressor and the characteristics of the individual experiencing the stressor.
Like Agnew’s (1992) original statement of GST, however, the arguments in this article are in need of further research and elaboration. The predictions regarding the impact of specific types of strain on crime are tentative. Researchers should use the methods described in this article to better deter- mine the extent to which these and other types of strain are seen as unjust, are seen as high in magnitude, are associated with low social control, and create some pressure or incentive for crime. Such research should improve the accu- racy of the predictions that are made. Furthermore, researchers should pay attention to the impact of group membership in such research. For example, it is likely that there are group differences in the extent to which certain strains are seen as unjust or high in magnitude.8 In addition, researchers should examine whether particular strains have a greater impact on some types of crime than other types. For example, some research suggests that certain strains are more strongly related to aggression/violence than to other types of crime (e.g., Agnew 1990; Aseltine et al. 1998, 2000; Mazerolle et al. 2000; Mazerolle and Piquero 1997). (Likewise, the stress research reveals that some stressors are more strongly related to some types of negative outcomes than to others.) The arguments presented in this article, then, are still in need of much development, but that does not diminish their central thrust—some strains are more likely than others to result in crime.
NOTES
1. Most of the research in criminology simply assumes that certain events or conditions are disliked by most of the people being studied. This is probably a reasonable assumption in most cases (e.g., criminal victimization), although it is a more questionable assumption in other cases (e.g., changing schools). A potentially more serious problem with the criminology research is that researchers rarely employ a complete or comprehensive list of objective strains. Researchers usually only examine a few types of objective strain—often overlooking many of the most important types. For example, interviews with adolescents suggest that peer conflict and abuse
Agnew / GENERAL STRAIN THEORY 353
are among the most important types of objective strain in this group, but such conflict/abuse is rarely considered by researchers (although see Agnew 1997; Agnew and Brezina 1997; Ambert 1994; Aseltine, Gore, and Gordon 2000; Seiffge-Krenke 1995). Likewise, experiences with racial prejudice and discrimination are seldom considered by researchers, despite evidence that such experiences are a major type of objective strain among African Americans and others (Ambert 1994; Anderson 1999). Recent research suggests that the failure to examine the full range of stressors can lead researchers to substantially underestimate the effect of stress or strain (Turner, Wheaton, and Lloyd 1995).
2. This is much less of a problem when judges or group members are rating the injustice of objective strains because these ratings are averaged across judges or group members.
3. Attributions of recklessness and negligence may also lead to perceptions of unjust treat- ment, although they result in less blame than attributions of intent. See Tedeschi and Felson (1994) and Tyler et al. (1997) for discussions in this area.
4. The distributive justice literature focuses on norms governing the distribution of out- comes, with outcomes broadly defined. Such outcomes include the types of strain considered in general strain theory (GST): the blockage of goal-seeking behavior, the removal of positively valued stimuli, and the presentation of negatively valued stimuli. Several rules govern the distri- bution of outcomes (e.g., equity, need, equality). And a range of factors influences the choice of the most relevant rule(s) and the determination of whether the rule(s) has been violated—with self-interest being a major factor (Hegtvedt and Cook forthcoming; Hegtvedt and Markovsky 1995; Tyler et al. 1997). The procedural justice literature focuses on the process by which people decide how to distribute outcomes. Several factors have been found to influence judgments about the fairness or justice of this process, although the relative importance of these factors var- ies by type of situation and other variables (Hegtvedt and Markovsky 1995; Lind and Tyler 1988; Sherman 2000; Tyler 1994; Tyler et al. 1997). The interactional justice literature focuses on the norms governing interaction between people, with data indicating that people have a strong desire to be treated in a polite, respectful, considerate, nonaggressive manner (Mikula 1986, 1993; Mikula, Petri, and Tanzer 1990; Tedeschi and Felson 1994). The retributive justice litera- ture focuses on the factors that influence the reaction to people who break social rules, with research indicating that people feel a need to sanction those who intentionally violate rules and with the sanction being proportional to the harm intended or inflicted (Tedeschi and Felson 1994; Tyler et al. 1997). Violations of distributive, procedural, interactional, and retributive jus- tice norms may each influence overall evaluations of justice, although the relative importance of each type of justice varies according to several factors (Tyler et al. 1997).
5. Agnew (1992) argued that the inability to achieve educational and occupational goals would be related to crime if researchers focused on the disjunction between expectations or expected goals and actual achievements. He claimed that expectations are taken more seriously than aspirations. An empirical study by Jensen (1995), however, failed to find support for this argument—although further tests would be useful.
6. Colvin’s (2000) theory of differential coercion essentially described a general type of strain—coercion—said to be especially conducive to crime (the theory also presented excellent discussions of the many ways that coercion may contribute to crime and the cultural and struc- tural sources of coercion). Coercion involves “the use or threat of force and intimidation aimed at creating compliance through fear,” including the “actual or threatened removal of social sup- ports,” and “pressure arising from structural arrangements and circumstances that seem beyond individual control,” creating “a sense of desperation that seems to compel an individual toward immediate action.” This broad definition includes most or all of the types of strain said to lead to crime but may also include many of the strains not predicted to affect crime—such as the inabil- ity to achieve conventional success goals, demands for the care of conventional others, and isola- tion from peer groups and situations conducive to crime.
354 JOURNAL OF RESEARCH IN CRIME AND DELINQUENCY
7. One should also take account of the possibility that anger may indirectly affect crime by reducing the perceived costs of crime and increasing the perceived desirability of crime, as indi- cated earlier in this article.
8. Explaining the origins of such differences is, of course, central to any effort to develop the macro-side of GST (see the excellent discussions in Anderson 1999; Bernard 1990; Colvin 2000; and Messerschmidt 1993).
REFERENCES
Agnew, Robert. 1984. “Autonomy and Delinquency.” Sociological Perspectives 27:219-36. . 1985. “A Revised Strain Theory of Delinquency.” Social Forces 64:151-67. . 1990. “The Origins of Delinquent Events: An Examination of Offender Accounts.”
Journal of Research in Crime and Delinquency 27:267-94. . 1992. “Foundation for a General Strain Theory of Crime and Delinquency.” Criminol-
ogy 30:47-87. . 1995a. “Strain and Subcultural Theories of Criminality.” In Criminology: A Contempo-
rary Handbook, edited by Joseph F. Sheley. Belmont, CA: Wadsworth. . 1995b. “The Contribution of Social-Psychological Strain Theory to the Explanation of
Crime and Delinquency.” In The Legacy of Anomie Theory, Advances in Criminological The- ory, Vol. 6, edited by Freda Adler and William S. Laufer. New Brunswick, NJ: Transaction.
. 1995c. “Testing the Leading Crime Theories: An Alternative Strategy Focusing on Motivational Processes.” Journal of Research in Crime and Delinquency 32:363-98.
. 1995d. “Controlling Delinquency: Recommendations from General Strain Theory.” In Crime and Public Policy, edited by Hugh D. Barlow. Boulder, CO: Westview.
. 1997. “Stability and Change in Crime over the Life Course: A Strain Theory Explana- tion.” In Developmental Theories of Crime and Delinquency, Advances in Criminological Theory, Vol. 7, edited by Terence P. Thornberry. New Brunswick, NJ: Transaction.
. 2001a. “An Overview of General Strain Theory.” In Explaining Criminals and Crime, edited by Raymond Paternoster and Ronet Bachman. Los Angeles: Roxbury.
. 2001b. Juvenile Delinquency: Causes and Control. Los Angeles: Roxbury. Agnew, Robert and Timothy Brezina. 1997. “Relational Problems with Peers, Gender and Delin-
quency.” Youth and Society 29:84-111. Agnew, Robert, Francis T. Cullen, Velmer S. Burton, Jr., T. David Evans, and R. Gregory
Dunaway. 1996. “A New Test of Classic Strain Theory.” Justice Quarterly 13:681-704. Agnew, Robert and David M. Petersen. 1989. “Leisure and Delinquency.” Social Problems
36:332-50. Agnew, Robert and Helene Raskin White. 1992. “An Empirical Test of General Strain Theory.”
Criminology 30:475-99. Ambert, Ann-Marie. 1994. “A Qualitative Study of Peer Abuse and Its Effects: Theoretical and
Empirical Implications.” Journal of Marriage and the Family 56:119-30. Anderson, Elijah. 1999. Code of the Street. New York: Norton. Aseltine, Robert H., Jr., Susan Gore, and Mary Ellen Colten. 1998. “The Co-occurrence of
Depression and Substance Abuse in Late Adolescence.” Development and Psychopathology 10:549-70.
Aseltine, Robert H., Jr., Susan Gore, and Jennifer Gordon. 2000. “Life Stress, Anger and Anxi- ety, and Delinquency: An Empirical Test of General Strain Theory.” Journal of Health and Social Behavior 41:256-75.
Agnew / GENERAL STRAIN THEORY 355
Averill, James R. 1982. Anger and Aggression: An Essay on Emotion. New York: Springer-Verlag.
. 1993. “Illusions of Anger.” In Aggression and Violence, edited by Richard B. Felson and James T. Tedeschi. New York: Springer-Verlag.
Bandura, Albert. 1990. “Selective Activation and Disengagement of Moral Control.” Journal of Social Issues 46:27-46.
Baron, Stephen W. and Timothy F. Hartnagel. 1997. “Attributions, Affect, and Crime: Street Youths’ Reactions to Unemployment.” Criminology 35:409-34.
Bazemore, Gordon. 1998. “Restorative Justice and Earned Redemption.” American Behavioral Scientist 41:768-813.
Berkowitz, Leonard. 1993. Aggression: Its Causes, Consequences, and Control. New York: McGraw-Hill.
Bernard, Thomas J. 1990. “Angry Aggression among the ‘Truly Disadvantaged.’ ” Criminology 28:73-96.
Black, Donald. 1983. “Crime as Social Control.” American Sociological Review 48:34-45. Blau, Judith R. and Peter M. Blau. 1982. “The Cost of Inequality: Metropolitan Structure and
Violent Crime.” American Sociological Review 47:114-29. Box, Steven. 1987. Recession, Crime and Punishment. Basingstoke, UK: Macmillan. Briathwaite, John. 1989. Crime, Shame, and Reintegration. Cambridge, UK: Cambridge Uni-
versity Press. Brezina, Timothy. 1998. “Adolescent Maltreatment and Delinquency: The Question of Inter-
vening Processes.” Journal of Research in Crime and Delinquency 35:71-99. . 1999. “Teenage Violence toward Parents as an Adaptation to Family Strain.” Youth and
Society 30:416-44. . 2000. “Delinquent Problem-Solving: An Interpretative Framework for Criminological
Theory and Research.” Journal of Research in Crime and Delinquency 37:3-30. Broidy, Lisa and Robert Agnew. 1997. “Gender and Crime: A General Strain Theory Perspec-
tive.” Journal of Research in Crime and Delinquency 34:275-306. Brown, George W. 1998. “Loss and Depressive Disorders.” In Adversity, Stress, and
Psychopathology, edited by Bruce P. Dohrenwend. New York: Oxford University Press. Burke, Peter J. 1996. “Social Identities and Psychosocial Stress.” In Psychosocial Stress, edited
by Howard B. Kaplan. San Diego, CA: Academic Press. Burton, Velmer S., Jr. and R. Gregory Dunaway. 1994. “Strain, Relative Deprivation, and Mid-
dle-Class Delinquency.” In Varieties of Criminology, edited by Greg Barak. Westport, CT: Praeger.
Cernkovich, Stephen A., Peggy C. Giordano, and Jennifer L. Rudolph. 2000. “Race, Crime, and the American Dream.” Journal of Research in Crime and Delinquency 37:131-70.
Cloward, Richard A. and Lloyd E. Ohlin. 1960. Delinquency and Opportunity. New York: Free Press.
Cohen, Albert K. 1955. Delinquent Boys. New York: Free Press. . 1997. “An Elaboration of Anomie Theory.” In The Future of Anomie Theory, edited by
Nikos Passas and Robert Agnew. Boston: Northeastern University Press. Cohen, Sheldon, Ronald C. Kessler, and Lynn Underwood Gordon. 1995. Measuring Stress.
New York: Oxford University Press. Colvin, Mark. 2000. Crime and Coercion: An Integrated Theory of Chronic Criminality. New
York: St. Martin’s. Conger, Rand D., Katherine J. Conger, Glen H. Elder, Jr., Frederick O. Lorenz, Ronald L. Simons,
and Lee B. Whitbeck. 1992. “A Family Process Model of Economic Hardship and Adjust- ment of Early Adolescent Boys.” Child Development 63:526-41.
356 JOURNAL OF RESEARCH IN CRIME AND DELINQUENCY
Crittenden, Kathleen S. 1983. “Sociological Aspects of Attribution.” Annual Review of Sociol- ogy 9:425-46.
. 1989. “Causal Attribution in Sociocultural Context.” Sociological Quarterly 30:1-14. Crutchfield, Robert D. and Susan R. Pitchford. 1997. “Work and Crime: The Effects of Labor
Stratification.” Social Forces 76:93-118. Cullen, Francis T. and Robert Agnew. 1999. Criminological Theory: Past to Present. Los
Angeles: Roxbury. Davis, Nanette J. 1999. Youth Crisis. Westport, CT: Praeger. Dawkins, Nicola. 1997. “Striking Back: An Empirical Test of the Impact of Victimization on
Violent Crime.” Presented at the annual meeting of the American Society of Criminology, San Diego, CA.
Dodge, Kenneth A. and D. Schwartz. 1997. “Social Information Processing Mechanisms in Aggressive Behavior.” In Handbook of Antisocial Behavior, edited by David M. Stoff, James Breiling, and Jack D. Maser. New York: John Wiley.
Dohrenwend, Bruce P. 1998. Adversity, Stress, and Psychopathology. New York: Oxford Uni- versity Press.
. 2000. “The Role of Adversity and Stress in Psychopathology: Some Evidence and Its Implications for Theory and Research.” Journal of Health and Social Behavior 41:1-19.
Elder, Glen H., Jr., Linda K. George, and Michael J. Shanahan. 1996. “Psychosocial Stress over the Life Course.” In Psychosocial Stress, edited by Howard B. Kaplan. San Diego, CA: Aca- demic Press.
Esbensen, Finn-Aage and David Huizinga. 1991. “Juvenile Victimization and Delinquency.” Youth and Society 23:202-28.
Finch, Brian Karl, Bohdan Kolody, and William A. Vega. 2000. “Perceived Discrimination and Depression among Mexican-Origin Adults in California.” Journal of Health and Social Behavior 41:295-313.
Forman, Tyrone A., David R. Williams, and James S. Jackson. 1997. “Race, Place, and Discrimi- nation.” Perspectives on Social Problems 9:231-61.
Fox, Greer Linton and Dudley Chancey. 1998. “Sources of Economic Distress.” Journal of Fam- ily Issues 19:725-49.
Gottfredson, Michael R. and Travis Hirschi. 1990. A General Theory of Crime. Stanford, CA: Stanford University Press.
Hagan, John and Bill McCarthy. 1997. Mean Streets. Cambridge, UK: Cambridge University Press.
Hegtvedt, Karen A. and Karen S. Cook. Forthcoming. “Distributive Justice: Recent Theoretical Developments and Applications.” In The Justice Reader, edited by Joseph Sanders and V. Lee Hamilton. New York: Plenum..
Hegtvedt, Karen A. and Barry Markovsky. 1995. “Justice and Injustice.” In Sociological Per- spectives on Social Psychology, edited by Karen S. Cook, Gary Alan Fine, and James S. House. Needham Heights, NY: Allyn and Bacon.
Herbert, Tracy B. and Sheldon Cohen. 1996. “Measurement Issues in Research on Psychosocial Stress.” In Psychosocial Stress, edited by Howard B. Kaplan. San Diego, CA: Academic Press.
Hirschi, Travis. 1969. Causes of Delinquency. Berkeley: University of California Press. Hoffmann, John P. and Felice Gray Cerbone. 1999. “Stressful Life Events and Delinquency
Escalation in Early Adolescence.” Criminology 37:343-74. Hoffmann, John P. and Alan S. Miller. 1998. “A Latent Variable Analysis of Strain Theory.”
Journal of Quantitative Criminology 14:83-110.
Agnew / GENERAL STRAIN THEORY 357
Hoffmann, John P. and S. Susan Su. 1997. “The Conditional Effects of Stress on Delinquency and Drug Use: A Strain Theory Assessment of Sex Differences.” Journal of Research in Crime and Delinquency 34:46-78.
Jensen, Gary F. 1995. “Salvaging Structure through Strain: A Theoretical and Empirical Cri- tique.” In The Legacy of Anomie Theory, Advances in Criminological Theory, Vol. 6, edited by Freda Adler and William S. Laufer. New Brunswick, NJ: Transaction.
Jensen, Gary F. and David Brownfield. 1986. “Gender, Lifestyles, and Victimization: Beyond Routine Activity.” Violence and Victims 1:85-99.
Kaplan, Howard B. 1996. “Psychosocial Stress from the Perspective of Self Theory.” In Psychosocial Stress, edited by Howard B. Kaplan. San Diego, CA: Academic Press.
Katz, Jack. 1988. Seductions of Crime. New York: Basic Books. Kessler, Ronald C., Jacquelyn Gillis-Light, William J. Magee, Kenneth S. Kendler, and
Lindon J. Eaves. 1997. “Childhood Adversity and Adult Psychopathology.” In Stress and Adversity over the Life Course, edited by Ian H. Gottlib and Blair Wheaton. Cambridge, UK: Cambridge University Press.
Kessler, Ronald C., James S. House, Renee R. Anspach, and David R. Williams. 1995. “Social Psychology and Health.” In Sociological Perspectives on Social Psychology, edited by Karen S. Cook, Gary Alan Fine, and James S. House. Needham Heights, NY: Allyn and Bacon.
Kornhauser, Ruth Rosner. 1978. Social Sources of Delinquency. Chicago: University of Chicago Press.
Lanza-Kaduce, Lonn and Marcia J. Radosevich. 1987. “Negative Reactions to Processing and Substance Use among Young Incarcerated Males.” Deviant Behavior 8:137-48.
Lauritsen, Janet L., John Laub, and Robert J. Sampson. 1992. “Conventional and Delinquent Activities: Implications for the Prevention of Violent Victimization among Adolescents.” Violence and Victims 7:91-108.
Lauritsen, Janet L., Robert J. Sampson, and John H. Laub. 1991. “The Link between Offending and Victimization among Adolescents.” Criminology 29:265-92.
Lazarus, Richard S. 1999. Stress and Emotion: A New Synthesis. New York: Springer. Lepore, Stephen J. 1995. “Measurement of Chronic Stressors.” In Measuring Stress, edited by
Sheldon Cohen, Ronald C. Kessler, and Lynn Underwood Gordon. New York: Oxford Uni- versity Press.
Lind, E. Allan and Tom R. Tyler. 1988. The Social Psychology of Procedural Justice. New York: Plenum.
Lockwood, Daniel. 1997. Violence among Middle School and High School Students: Analysis and Implications for Prevention. Washington, DC: National Institute of Justice.
Matza, David and Gresham Sykes. 1961. “Juvenile Delinquency and Subterranean Beliefs.” American Sociological Review 26:713-19.
Mazerolle, Paul. 1998. “Gender, General Strain, and Delinquency: An Empirical Examination.” Justice Quarterly 15:65-91.
Mazerolle, Paul, Velmer S. Burton, Jr., Francis T. Cullen, T. David Evans, and Gary L. Payne. 2000. “Strain, Anger, and Delinquency Adaptations: Specifying General Strain Theory.” Journal of Criminal Justice 28:89-101.
Mazerolle, Paul and Alex Piquero. 1997. “Violent Responses to Strain: An Examination of Con- ditioning Influences.” Violence and Victims 12:323-43.
. 1998. “Linking Exposure to Strain with Anger: An Investigation of Deviant Adapta- tions.” Journal of Criminal Justice 26:195-211.
Meier, Robert F. and Terence D. Miethe. 1993. “Understanding Theories of Criminal Victimiza- tion.” Crime and Justice: A Review of Research 17:465-95.
358 JOURNAL OF RESEARCH IN CRIME AND DELINQUENCY
Merton, Robert K. 1938. “Social Structure and Anomie.” American Sociological Review 3:672-82.
. 1968. Social Theory and Social Structure. New York: Free Press. Messerschmidt, James W. 1993. Masculinities and Crime. Lanham, MD: Rowman and
Littlefield. Mikula, Gerold. 1986. “The Experience of Injustice: Toward a Better Understanding of Its Phe-
nomenology.” In Justice in Social Relations, edited by Hans Werner Bierhoff, Ronald L. Cohen, and Jerald Greenberg. New York: Plenum.
. 1993. “On the Experience of Injustice.” European Journal of Social Psychology 4:223-44.
Mikula, Gerold, Birgit Petri, and Norbert Tanzer. 1990. “What People Regard as Unjust: Types and Structures of Everyday Experiences of Injustice.” European Journal of Social Psychol- ogy 2:133-49.
Moffitt, Terrie E. 1993. “ ‘Life-Course Persistent’ and ‘Adolescent-Limited’ Antisocial Behav- ior: A Developmental Taxonomy.” Psychological Review 100:674-701.
Nye, Ivan. 1958. Family Relationships and Delinquent Behavior. New York: John Wiley. Osgood, D. Wayne, Janet K. Wilson, Patrick M. O’Malley, Jerald G. Bachman, and Lloyd D.
Johnston. 1996. “Routine Activities and Individual Deviant Behavior.” American Sociologi- cal Review 61:635-55.
Paternoster, Raymond and Paul Mazerolle. 1994. “General Strain Theory and Delinquency: A Replication and Extension.” Journal of Research in Crime and Delinquency 31:235-63.
Patterson, Gerald R., John B. Reid, and Thomas J. Dishion. 1992. Antisocial Boys. Eugene, OR: Castalia.
Piquero, Nicole Leeper and Miriam D. Sealock. 2000. “Generalizing General Strain Theory: An Examination of an Offending Population.” Justice Quarterly 17:449-84.
Rokeach, Milton. 1973. The Nature of Human Values. New York: Free Press. Sampson, Robert J. and John H. Laub. 1993. Crime in the Making. Cambridge, MA: Harvard
University Press. Sampson, Robert J. and Janet L. Lauritsen. 1993. “Violent Victimization and Offending: Indi-
vidual-, Situational-, and Community-Level Risk Factors.” In Understanding and Pre- venting Violence, Vol. 3, Social Influences, edited by the National Research Council. Washington, DC: National Research Council.
Schieman, Scott. 2000. “Education and the Activation, Course, and Management of Anger.” Journal of Health and Social Behavior 41:20-39.
Schulz, Amy, David Williams, Barbara Israel, Adam Becker, Edith Parker, Sherman A. James, and James Jackson. 2000. “Unfair Treatment, Neighborhood Effects, and Mental Health in the Detroit Metropolitan Area.” Journal of Health and Social Behavior 41:314-32.
Seiffge-Krenke, Inge. 1995. Stress, Coping, and Relationships in Adolescence. Mahwah, NJ: Lawrence Erlbaum.
Sherman, Lawrence W. 1993. “Defiance, Deterrence, and Irrelevance: A Theory of Criminal Sanctions.” Journal of Research in Crime and Delinquency 30:445-73.
. 2000. “The Defiant Imagination: Consilience and the Science of Sanctions.” Presented at the University of Pennsylvania, Philadelphia.
Smith, Carolyn and Terence P. Thornberry. 1995. “The Relationship between Childhood Mal- treatment and Adolescent Involvement in Delinquency.” Criminology 33:451-81.
Smith-Lovin, Lynn. 1995. “The Sociology of Affect and Emotion.” In “Sociological Perspec- tives on Social Psychology, edited by Karen S. Cook, Gary Alan Fine, and James S. House. Needham Heights, NY: Allyn and Bacon.
Agnew / GENERAL STRAIN THEORY 359
Stone, Arthur A. 1995. “Measurement of Affective Response.” In Measuring Stress, edited by Sheldon Cohen, Ronald C. Kessler, and Lynn Underwood Gordon. New York: Oxford Uni- versity Press.
Sykes, Gresham M. and David Matza. 1957. “Techniques of Neutralization: A Theory of Delin- quency.” American Sociological Review 22:664-70.
Taylor, Shelley E. and Lisa G. Aspinwall. 1996. “Mediating and Moderating Processes in Psychosocial Stress: Appraisal, Coping, Resistance, and Vulnerability.” In Psychosocial Stress, edited by Howard B. Kaplan. San Diego, CA: Academic Press.
Tedeschi, James T. and Richard B. Felson. 1994. Violence, Aggression, and Coercive Actions. Washington, DC: American Psychological Association.
Tedeschi, James T. and Mitchell S. Nesler. 1993. “Grievances: Development and Reaction.” In Aggression and Violence, edited by Richard B. Felson and James T. Tedeschi. Washington, DC: American Psychological Association.
Thoits, Peggy A. 1989. “The Sociology of Emotions.” Annual Review of Sociology 15:317-42. . 1991. “On Merging Identity Theory and Stress Research.” Social Psychology Quarterly
54:101-12. . 1995. “Stress, Coping, and Social Support Processes: Where Are We? What Next?”
Journal of Health and Social Behavior (Extra Issue):53-79. Tittle, Charles R. 1995. Control Balance: Toward a General Theory of Deviance. Boulder, CO:
Westview. Turner, Jay R. and William R. Avison. 1992. “Innovations in the Measurement of Life Stress:
Crisis Theory and the Significance of Event Resolution.” Journal of Health and Social Behavior 33:36-50.
Turner, R. Jay and Blair Wheaton. 1995. “Checklist Measurement of Stressful Life Events.” In Measuring Stress, edited by Sheldon Cohen, Ronald C. Kessler, and Lynn Underwood Gordon. New York: Oxford University Press.
Turner, R. Jay, Blair Wheaton, and Donald A. Lloyd. 1995. “The Epidemiology of Social Stress.” American Sociological Review 60:104-25.
Tyler, Tom. 1990. Why People Obey the Law. New Haven, CT: Yale University Press. . 1994. “Psychological Models of the Justice Process: Antecedents of Distributive and
Procedural Justice.” Journal of Personality and Social Psychology 67:850-63. Tyler, Tom R., Robert J. Boeckmann, Heather J. Smith, and Yuen J. Huo. 1997. Social Justice in
a Diverse Society. Boulder, CO: Westview. Uggen, Christopher. 2000. “Work as a Turning Point in the Life Course of Criminals: A Duration
Model of Age, Employment, and Recidivism.” American Sociological Review 67:529-46. Voydanoff, Patricia. 1990. “Economic Distress and Family Relations: A Review of the Eighties.”
Journal of Marriage and the Family 52:1099-115. Wells, Edward L. and Joseph H. Rankin. 1988. “Direct Parental Controls and Delinquency.”
Criminology 26:263-84. Wethington, Elaine, George W. Brown, and Ronald C. Kessler. 1995. “Interview Measurement
of Stressful Life Events.” In Measuring Stress, edited by Sheldon Cohen, Ronald C. Kessler, and Lynn Underwood Gordon. New York: Oxford University Press.
Wheaton, Blair. 1990. “Life Transitions, Role Histories, and Mental Health.” American Socio- logical Review 55:209-24.
. 1996. “The Domains and Boundaries of Stress Concepts.” In Psychosocial Stress, edited by Howard B. Kaplan. San Diego, CA: Academic Press.
Wheaton, Blair, Patricia Roszell, and Kimberlee Hall. 1997. “The Impact of Twenty Childhood and Adult Traumatic Stressors on the Risk of Psychiatric Disorder.” In Stress and Adversity over the Life Course: Trajectories and Turning Points, edited by Ian H. Gottlib. New York: Cambridge University Press.
360 JOURNAL OF RESEARCH IN CRIME AND DELINQUENCY
White, Helene Raskin, Erich W. Labouvie, and Marsha E. Bates. 1985. “The Relationship between Sensation Seeking and Delinquency: A Longitudinal Analysis.” Journal of Research in Crime and Delinquency 22:197-211.
Widom, Cathy Spatz. 1998. “Childhood Victimization: Early Adversity and Subsequent Psychopathology.” In Adversity, Stress, and Psychopathology, edited by Bruce P. Dohrenwend. New York: Oxford University Press.
Agnew / GENERAL STRAIN THEORY 361
- Editorial Board
- Contents
- Building on the Foundation of General Strain Theory: Specifying the Types of Strain Most likely to Lead to Crime and Delinque
- Student Anger and Aggressive Behavior in School: An Initial Test of Agnew's Macro-Level Strain Theory
- Policing Is Not a Treatment: Alternatives to the Medical Model of Police Research
- The Effects of Substance Use on Specific Types of Criminal Offending in Young Men
- Crime and Risky Behavior in Traffic: An Example of Cross-Situational Consistency
- About the Authors
- Index to Volume 38