8-10 page paper Politics of Violence

FOUNDATION FOR A GENERAL STRAIN THEORY OF CRIME AND DELINQUENCY*

ROBERT AGNEW Emory University

This paper presents a general strain theory of crime and delinquency that is capable of overcoming the criticisms of previous strain theories. In the first section, strain theory is distinguished from social control and dif- ferential association/social learning theory. In the second section, the three major types of strain are described: (1) strain as the actual or antici- pated failure to achieve positively valued goals, (2) strain as the actual or anticipated removal of positively valued stimuli, and (3) strain as the actual or anticipated presentation of negatively valued stimuli. In the third section, guidelines for the measurement of strain are presented. And in the fourth section, the major adaptations to strain are described, and thcwe factors influencing the choice of delinquent versus nondelinquent adaptations are discussed.

After dominating deviance research in the 196Os, strain theory came under heavy attack in the 1970s (Bernard, 1984; Cole, 1975), with several promi- nent researchers suggesting that the theory be abandoned (Hirschi, 1969; Kornhauser, 1978). Strain theory has survived those attacks, but its influence is much diminished (see Agnew, 1985a; Bernard, 1984; Farnworth and Lei- ber, 1989). In particular, variables derived from strain theory now play a very limited role in explanations of crime/delinquency. Several recent causal models of delinquency, in fact, either entirely exchde strain variables or assign them a small role (e.g., Elliott et al., 1985; Johnson, 1979; Massey and Krohn, 1986; Thornberry, 1987; Tonry et al., 1991). Causal models of crime/delinquency are dominated, instead, by variables derived from differ- ential association/social learning theory and social control theory.

This paper argues that strain theory has a central role to play in explana- tions of crime/delinquency, but that the theory has to be substantially revised to play this role. Most empirical studies of strain theory continue to rely on the strain models developed by Merton (1938), A. Cohen (1955), and Cloward and Ohlin (1960). In recent years, however, a wealth of research in several fields has questioned certain of the assumptions underlying those the- ories and pointed to new directions for the development of strain theory. Most notable in this area is the research on stress in medical sociology and psychology, on equity/justice in social psychology, and on aggression in psy- chology-particularly recent versions of frustration-aggression and social

* I would like to thank Helene Raskin I’’hite and Karen Hegtvedt for their comments.

CRIMINOLOGY VOLUME 30 NUMBER 1 1992 47

48 AGNEW

learning theory. Also important is recent research in such areas as the legiti- mation of stratification, the sociology of emotions, and the urban underclass. Certain researchers have drawn on segments of the above research to suggest new directions for strain theory (Agnew, 1985a; Bernard, 1987; Elliott et al., 1979; Greenberg, 1977), but the revisions suggested have not taken full advantage of this research and, at best, provide only incomplete models of strain and delinquency. (Note that most of the theoretical and empirical work on strain theory has focused on delinquency.) This paper draws on the above literatures, as well as the recent revisions in strain theory, to present the outlines of a general strain theory of crime/delinquency.

The theory is written at the social-psychological level: It focuses on the individual and his or her immediate social environment-although the macroimplications of the theory are explored at various points. The theory is also written with the empirical researcher in mind, and guidelines for testing the theory in adolescent populations are provided. The focus is on adoles- cents because most currently available data sets capable of testing the theory involve surveys of adolescents. This general theory, it will be argued, is capa- ble of overcoming the theoretical and empirical criticisms of previous strain theories and of complementing the crime/delinquency theories that currently dominate the field.

The paper is in four sections. In the first section, there is a brief discussion of the fundamental traits that distinguish strain theory from the other two dominant theories of delinquency: social control and differential association/ social learning theory (in the interest of brevity, the term delinquency is used rather than crime and delinquency). In the second section, the three major sources of strain are described. In the third section, guidelines for the mea- surement of strain are provided. And in the final section, the major adapta- tions to strain are listed and the factors influencing the choice of delinquent versus nondelinquent adaptations are discussed.

STRAIN THEORY AS DISTINGUISHED FROM CONTROL AND DIFFERENTIAL

ASSOCIATION/SOCIAL LEARNING THEORY

Strain, social control, and differential association theory are all sociological theories: They explain delinquency in terms of the individual’s social rela- tionships. Strain theory is distinguished from social control and social learn- ing theory in its specification of (1) the type of social relationship that leads to delinquency and (2) the motivation for delinquency. First, strain theory focuses explicitly on negative relationships with others: relationships in which the individual is not treated as he or she wants to be treated. Strain theory has typically focused on relationships in which others prevent the individual

A GENERAL STRAIN THEORY 49

from achieving positively valued goals. Agnew (1 985a), however, broadened the focus of strain theory to include relationships in which others present the individual with noxious or negative stimuli. Social control theory, by con- trast, focuses on the absence of significant relationships with conventional others and institutions. In particular, delinquency is most likely when (1) the adolescent is not attached to parents, school, or other institutions; (2) parents and others fail to monitor and effectively sanction deviance; (3) the adoles- cent’s actual or anticipated investment in conventional society is minimal; and (4) the adolescent has not internalized conventional beliefs. Social learn- ing theory is distinguished from strain and control theory by its focus on positive relationships with deviant others. In particular, delinquency results from association with others who (1) differentially reinforce the adolescent’s delinquency, (2) model delinquent behavior, and/or (3) transmit delinquent values.

Second, strain theory argues that adolescents are pressured into delinquency by the negative afective states-most notably anger and related emotions- that often result from negative relationships (see Kemper, 1978, and Morgan and Heise, 1988, for typologies of negative affective states). This negative affect creates pressure for corrective action and may lead adolescents to (1) make use of illegitimate channels of goal achievement, (2) attack or escape from the source of their adversity, and/or (3) manage their negative affect through the use of illicit drugs. Control theory, by contrast, denies that outside forces pressure the adolescent into delinquency. Rather, the absence of significant relationships with other individuals and groups frees the adoles- cent to engage in delinquency. The freed adolescent either drifts into delin- quency or, in some versions of control theory, turns to delinquency in response to inner forces or situational inducements (see Hirschi, 1969:3 1-34). In differential association/social learning theory, the adolescent commits delinquent acts because group forces lead the adolescent to view delinquency as a desirable or at least justifable form of behavior under certain circumstances.

Strain theory, then, is distinguished by its focus on negative relationships with others and its insistence that such relationships lead to delinquency through the negative affect-specially anger-they sometimes engender. Both dimensions are necessary to differentiate strain theory from control and differential association/social learning theory. In particular, social control and social learning theory sometimes examine negative relationships- although such relationships are not an explicit focus of these theories. Con- trol theory, however, would argue that negative relationships lead to delin- quency not because they cause negative affect, but because they lead to a reduction in social control. A control theorist, for example, would argue that physical abuse by parents leads to delinquency because it reduces attachment to parents and the effectiveness of parents as socializing agents. Likewise,

50 AGNEW

differential associatiodsocial learning theorists sometimes examine negative relationships-even though theorists in this tradition emphasize that imita- tion, reinforcement, and the internalization of values are less likely in nega- tive relationships. Social learning theorists, however, would argue that negative relationships-such as those involving physically abusive parents- lead to delinquency by providing models for imitation and implicitly teaching the child that violence and other forms of deviance are acceptable behavior.

Phrased in the above manner, it is easy to see that strain theory comple- ments the other major theories of delinquency in a fundamental way. While these other theories focus on the absence of relationships or on positive rela- tionships, strain theory is the only theory to focus explicitly on negative rela- tionships. And while these other theories view delinquency as the result of drift or of desire, strain theory views it as the result of pressure.

THE MAJOR TYPES OF STRAIN

Negative relationships with others are, quite simply, relationships in which others are not treating the individual as he or she would like to be treated. The classic strain theories of Merton (1938), A. Cohen (1955), and Cloward and Ohlin (1960) focus on only one type of negative relationship: relation- ships in which others prevent the individual from achieving positively valued goals. In particular, they focus on the goal blockage experienced by lower- class individuals trying to achieve monetary success or middle-class status. More recent versions of strain theory have argued that adolescents are not only concerned about the future goals of monetary success/middle-class sta- tus, but are also concerned about the achievement of more immediate goals- such as good grades, popularity with the opposite sex, and doing well in ath- letics (Agnew, 1984; Elliott and Voss, 1974; Elliott et al., 1985; Empey, 1982; Greenberg, 1977; Quicker, 1974). The focus, however, is still on the achieve- ment of positively valued goals. Most recently, Agnew (1985a) has argued that strain may result not only from the failure to achieve positively valued goals, but also from the inability to escape legally from painful situations. If one draws on the above theories-as well as the stress, equity/justice, and aggression literatures-one can begin to develop a more complete classifica- tion of the types of strain.

Three major types of strain are described-ch refemng to a different type of negative relationship with others. Other individuals may (1) prevent one from achieving positively valued goals, (2) remove or threaten to remove pos- itively valued stimuli that one possesses, or (3) present or threaten to present one with noxious or negatively valued stimuli. These categories of strain are presented as ideal types. There is no expectation, for example, that a factor analysis of strainful events will reproduce these categories. These categories,

A GENERAL STRAIN THEORY 51

rather, are presented so as to ensure that the full range of strainful events are considered in empirical research.

STRAIN AS THE FAILURE TO ACHIEVE POSITIVELY VALUED GOALS

At least three types of strain fall under this category. The first type encom- passes most of the major strain theories in criminology, including the classic strain theories of Merton, A. Cohen, and Cloward and Ohlin, as well as those modem strain theories focusing on the achievement of immediate goals. The other two types of strain in this category are derived from the justicdequity literature and have not been examined in criminology.

STRAIN AS THE DISJUNCTION BETWEEN ASPIRATIONS AND EXPECTATIONS/ACTUAL ACHIEVEMENTS

The classic strain theories of Merton, A. Cohen, and Cloward and Ohlin argue that the cultural system encourages everyone to pursue the ideal goals of monetary success and/or middle-class status. Lower-class individuals, however, are often prevented from achieving such goals through legitimate channels. In line with such theories, adolescent strain is typically measured in terms of the disjunction between aspirations (or ideal goals) and expecta- tions (or expected levels of goal achievement). These theories, however, have been criticized for several reasons (see Agnew, 1986, 1991b; Chard, 1964; Hirschi, 1969; Kornhauser, 1978; Liska, 1987; also see Bernard, 1984; Farnworth and Leiber, 1989). Among other things, it has been charged that these theories (1) are unable to explain the extensive nature of middle-class delinquency, (2) neglect goals other than monetary success/middle-class sta- tus, (3) neglect barriers to goal achievement other than social class, and (4) do not fully specify why only some strained individuals turn to delinquency. The most damaging criticism, however, stems from the limited empirical support provided by studies focusing on the disjunction between aspirations and expectations (see Kornhauser, 1978, as well the arguments of Bernard, 1984; Elliott et al., 1985; and Jensen, 1986).

As a consequence of these criticisms, several researchers have revised the above theories. The most popular revision argues that there is a youth sub- culture that emphasizes a variety of immediate goals. The achievement of these goals is further said to depend on a variety of factors besides social class: factors such as intelligence, physical attractiveness, personality, and athletic ability. As a result, many middle-class individuals find that they lack the traits or skills necessary to achieve their goals through legitimate chan- nels. This version of strain theory, however, continues to argue that strain stems from the inability to achieve certain ideal goals emphasized by the (sub)cultural system. As a consequence, strain continues to be measured in

52 AGNEW

terms of the disjunction between aspirations and actual achievements (since we are dealing with immediate rather than future goals, actual achievements rather than expected achievements may be examined).

It should be noted that empirical support for this revised version of strain theory is also weak (see Agnew, 1991b, for a summary). At a later point, several possible reasons for the weak empirical support of strain theories focusing on the disjunction between aspirations and expectations/achieve- ments will be discussed. For now, the focus is on classifying the major types of strain.

STRAIN AS THE DISJUNCTION BETWEEN EXPECTATIONS AND ACTUAL ACHIEVEMENTS

As indicated above, strain theories in criminology focus on the inability to achieve ideal goals derived from the cultural system. This approach stands in contrast to certain of the research on justice in social psychology. Here the focus is on the disjunction between expectations and actual achievements (rewards), and it is commonly argued that such expectations are existentially based. In particular, it has been argued that such expectations derive from the individual’s past experience and/or from comparisons with referential (or generalized) others who are similar to the individual (see Berger et al., 1972, 1983; Blau, 1964; Homans, 1961; Jasso and Rossi, 1977; Mickelson, 1990 Ross et al., 1971; Thibaut and Kelley, 1959). Much of the research in this area has focused on income expectations, although the above theories apply to expectations regarding all manner of positive stimuli. The justice literature argues that the failure to achieve such expectations may lead to such emo- tions as anger, resentment, rage, dissatisfaction, disappointment, and unhap- piness-that is, all the emotions customarily associated with strain in criminology. Further, it is argued that individuals will be strongly motivated to reduce the gap between expectations and achievements-with deviance being commonly mentioned as one possible option. This literature has not devoted much empirical research to deviance, although limited data suggest that the expectations-achievement gap is related to anger/hostility (Ross et al, 1971).

This alternative conception of strain has been largely neglected in criminol- ogy. This is unfortunate because it has the potential to overcome certain of the problems of current strain theories. First, one would expect the disjunc- tion between expectations and actual achievements to be more emotionally distressing than that between aspirations and achievements. Aspirations, by definition, are ideal goals. They have something of the utopian in them, and for that reason, the failure to achieve aspirations may not be taken seriously. The failure to achieve expected goals, however, is likely to be taken seriously since such goals are rooted in reality-the individual has previously exper- ienced such goals or has seen similar others experience such goals. Second,

A GENERAL STRAIN THEORY 53

this alternative conception of strain assigns a central role to the social com- parison process. As A. Cohen (1965) argued in a follow-up to his strain the- ory, the neglect of social comparison is a major shortcoming of strain theory. The above theories describe one way in which social comparison is important: Social comparison plays a central role in the formation of individual goals (expectations in this case; also see Suls, 1977). Third, the assumption that goals are culturally based has sometimes proved problematic for strain theory (see Kornhauser, 1978). Among other things, it makes it difficult to integrate strain theory with social control and cultural deviance theory (see Hirschi, 1979). These latter theories assume that the individual is weakly tied to the cultural system or tied to alternative/oppositional subcultures. The argu- ment that goals are existentially based, however, paves the way for integra- tions involving strain theory. 1

STRAIN AS THE DISJUNCTION BETWEEN JUSTIFAIR OUTCOMFS AND ACTUAL OUTCOMES

The above models of strain assume that individual goals focus on the achievement of specific outcomes. Individual goals, for example, focus on the achievement of a certain amount of money or a certain grade-point average. A third conception of strain, also derived from the justice/equity literature, makes a rather different argument. It claims that individuals do not necessar- ily enter interactions with specific outcomes in mind. Rather, they enter interactions expecting that certain distributive justice rules will be followed, rules specifying how resources should be allocated. The rule that has received the most attention in the literature is that of equity. An equitable relationship is one in which the outcome/input ratios of the actors involved in an exchange/allocation relationship are equivalent (see Adams, 1963, 1965; Cook and Hegtvedt, 1983; Walster et al., 1978). Outcomes encompass a broad range of positive and negative consequences, while inputs encompass the individual's positive and negative contributions to the exchange. Individ- uals in a relationship will compare the ratio of their outcomes and inputs to the ratio@) of specific others in the relationship. If the ratios are equal to one another, they feel that the outcomes are fair or just. This is true, according to equity theorists, even if the outcomes are low. If outcome/input ratios are

1. One need not assume that expectations are existentially based; they may derive from the cultural system as well. Likewise, one need not assume that aspirations derive from the cultural system. The focus in this paper is on types of strain rather than sources of strain, although a consideration of sources is crucial when the macroimplications of the theory are developed. Additional information on the sources of positively valued goals- including aspirations and expectations-can be found in Alves and Rossi, 1978; Cook and Messick, 1983; Hochschild, 1981; Jasso and Rossi, 1977; Martin and Murray, 1983; Mes- sick and Sentis, 1983; Mickelson, 1990; and Shepelak and Alwin, 1986.

54 AGNEW

not equal, actors will feel that the outcomes are unjust and they will experi- ence distress as a result. Such distress is especially likely when individuals feel they have been underrewarded rather than overrewarded (Hegtvedt, 1990).

The equity literature has described the possible reactions to this distress, some of which involve deviance (see Adams, 1963, 1965; Austin, 1977; Wal- ster et al., 1973, 1978; see Stephenson and White, 1968, for an attempt to recast A. Cohen’s strain theory in terms of equity theory). In particular, inequity may lead to delinquency for several reasons-all having to do with the restoration of equity. Individuals in inequitable relationships may engage in delinquency in order to (1) increase their outcomes (e.g., by theft); (2) lower their inputs (e.g., truancy from school); (3) lower the outcomes of others (e.g., vandalism, theft, assault); and/or (4) increase the inputs of others (e.g., by being incorrigible or disorderly). In highly inequitable situations, individuals may leave the field (e.g., run away from home) or force others to leave the field.* There has not been any empirical research on the relation- ship between equity and delinquency, although much data suggest that ineq- uity leads to anger and frustration. A few studies also suggest that insulting and vengeful behaviors may result from inequity (see Cook and Hegtvedt, 1991; Donnerstein and Hatfield, 1982; Hegtvedt, 1990; Mikula, 1986; Sprecher, 1986; Walster et al., 1973, 1978).

It is not difficult to measure equity. Walster et al. (1978:23&242) provide the most complete guide to measurement.3 Sprecher (1986) illustrates how

2. Theorists have recently argued that efforts to restore equity need not involve the specific others in the inequitable relationship. If one cannot restore equity with such spe- cific others, there may be an effort to restore “equity with the world” (Austin, 1977; Ste- phenson and White, 1968; Walster et al., 1978). That is, individuals who feel they have been inequitably treated may try to restore equity in the context of a totally different rela- tionship. The adolescent who is inequitably treated by parents, for example, may respond by inequitably treating peers. The concept of “equity with the world” has not been the subject of much empirical research, but it is intriguing because it provides a novel explana- tion for displayed aggression. It has also been argued that individuals may be distressed not only by their own inequitable treatment, but also by the inequitable treatment of others (see Crosby and Gonzalez-Intal, 1984; Walster et al., 1978.) We may have, then, a sort of vicarious strain, a type little investigated in the literature.

The equity literature has been criticized on a number of points, the most promi- nent being that there are a variety of distribution rules besides equity-such as equality and need (Deutsch, 1975; Folger, 1984; Mikula, 1980; Schwinger, 1980; Utne and Kidd, 1980). Much recent research has focused on the factors that determine the preference for one rule over another (Alves and Rossi, 1978; Cook and Hegtvedt, 1983; Deutsch, 1975; Hegtvedt, 1987, 1991a; Hochschild, 1981; Lerner, 1977; Leventhal, 1976; Leventhal et al., 1980; Schwinger, 1980; Walster et al., 1978). Also, the equity literature argues that individuals compare themselves with similar others with whom they are involved in exchange/alloca- tion relations. However, it has been argued that individuals sometimes compare themselves with dissimilar others, make referential (generalized) rather than local (specific) compari- sons, make internal rather than external comparisons, make group-to-group comparisons,

3.

A GENERAL STRAIN THEORY 55

equity may be measured in social surveys; respondents are asked who contrib- utes more to a particular relationship and/or who “gets the best deal” out of a relationship. A still simpler strategy might be to ask respondents how fair or just their interactions with others, such as parents or teachers, are. One would then predict that those involved in unfair relations will be more likely to engage in current and future delinquency.

The literature on equity builds on the strain theory literature in criminol- ogy in several ways. First, all of the strain literature assumes that individuals are pursuing some specific outcome, such as a certain amount of money or prestige. The equity literature points out that individuals do not necessarily enter into interactions with specific outcomes in mind, but rather with the expectation that a particular distributive justice rule will be followed. Their goal is that the interaction conform to the justice principle. This perspective, then, points to a new source of strain not considered in the criminology litera- ture. Second, the strain literature in criminology focuses largely on the indi- vidual‘s outcomes. Individuals are assumed to be pursuing a specific goal, and strain is judged in terms of the disjuntion between the goal and the actual outcome. The equity literature suggests that this may be an oversimplified conception and that the individual’s inputs may also have to be considered. In particular, an equity theorist would argue that inputs will condition the individual’s evaluation of outcomes. That is, individuals who view their inputs as limited will be more likely to accept limited outcomes as fair. Third, the equity literature also highlights the importance of the social com- parison process. In particular, the equity literature stresses that one’s evalua- tion of outcomes is at least partly a function of the outcomes (and inputs) of

. those with whom one is involved in exchange/allocation relations. A given outcome, then, may be evaluated as fair or unfair depending on the outcomes (and inputs) of others in the exchange/allocation relation.

or avoid social comparison altogether (see Berger et al., 1972; Hegtvedt, 1991b; Martin and Murray, 1983; see Hegtvedt, 1991b, and Suls and Wills, 1991, for a discussion of the factors affecting the choice of comparison objects). Finally, even if one knows what distribution rule individuals prefer and the types of social comparisons they make, it is still difficult to predict whether they will evaluate their interactions as equitable. Except in unambiguous situations of the type created in experiments, it is hard to predict what inputs and outcomes individuals will define as relevant, how they will weight those inputs and outcomes, and how they will evaluate themselves and others on those inputs and outcomes (Austin, 1977; Hegtvedt, 1991a; Messick and Sentis, 1979, 1983; Walster et al., 1973, 1978). Fortunately, however, the above three problems do not prohibit strain theory from taking advantage of certain of the insights from equity theory. While it is difficult to predict whether individu- als will define their relationships as equitable, it is relatively easy to measure equity after the fact.

56 AGNEW

SUMMARY: STRAIN AS THE FAILURE TO ACHIEVE POSITIVELY VALUED GOALS

Three types of strain in this category have been listed: strain as the dis- junction between (1) aspirations and expectations/actual achievements, (2) expectations and actual achievements, and (3) just/fair outcomes and actual outcomes. Strain theory in criminology has focused on the first type of strain, arguing that it is most responsible for the delinquency in our society. Major research traditions in the justice/equity field, however, argue that anger and frustration derive primarily from the second two types of strain. To complicate matters further, one can list still additional types of strain in this category. Certain of the literature, for example, has talked of the disjunc- tion between “satisfying outcomes” and reality, between “deserved” out- comes and reality, and between “tolerance levels” or minimally acceptable outcomes and reality. No study has examined all of these types of goals, but taken as a whole the data do suggest that there are often differences among aspirations (ideal outcomes), expectations (expected outcomes), “satisfying” outcomes, “deserved” outcomes, fair or just outcomes, and tolerance levels (Della Fave, 1974; Della Fave and Klobus, 1976; Martin, 1986; Martin and Murray, 1983; Messick and Sentis, 1983; Shepelak and Alwin, 1986). This paper has focused on the three types of strain listed above largely because they dominate the current literature.4

Given these multiple sources of strain, one might ask which is the most relevant to the explanation of delinquency. This is a difficult question to answer given current research. The most fruitful strategy at the present time may be to assume that all of the above sources are relevant-that there are several sources of frustration. Alwin (1987), Austin (1977), Crosby and Gon- zalez-Intal (1984), Hegtvedt (1991b), Messick and Sentis (1983), and Torn- blum (1977) all argue or imply that people often employ a variety of standards to evaluate their situation. Strain theorists, then, might be best advised to employ measures that tap all of the above types of strain. One might, for example, focus on a broad range of positively valued goals and, for each goal, ask adolescents whether they are achieving their ideal outcomes (aspirations), expected outcomes, and just/fair outcomes. One would expect strain to be greatest when several standards were not being met, with perhaps greatest weight being given to expectations and just/fair outcomes.5

4. To add a still further complication, it has been suggested that anger may result from the violation of procedural as well as distributive justice rules (Folger, 1984, 1986; Lind and Tyler, 1988). Procedural justice does not focus on the fairness of outcomes, but rather on the fairness of the procedures by which individuals decide how to distribute resources. A central issue in procedural justice is whether all individuals have a “voice” in deciding how resources will be distributed. One might, then, ask adolescents about the fairness of the procedures used by parents, teachers, and others to make rules.

This strategy assumes that all standards are relevant in a given situation, which 5 .

A GENERAL STRAIN THEORY 57

STRAIN AS THE REMOVAL OF POSITIVELY VALUED STIMULI FROM THE INDIVIDUAL

The psychological literature on aggression and the stress literature suggest that strain may involve more than the pursuit of positively valued goals. Cer- tain of the aggression literature, in fact, has come to de-emphasize the pursuit of positively valued goals, pointing out that the blockage of goal-seeking behavior is a relatively weak predictor of aggression, particularly when the goal has never been experienced before (Bandura, 1973; Zillman, 1979). The stress literature has largely neglected the pursuit of positively valued goals as a source of stress. Rather, if one looks at the stressful life events examined in this literature, one finds a focus on (1) events involving the loss of positively valued stimuli and (2) events involving the presentation of noxious or nega- tive stimuli (see Pearlin, 1983, for other typologies of stressful life events/ conditions).6 So, for example, one recent study of adolescent stress employs a life-events list that focuses on such items as the loss of a boyfriend/girlfriend, the death or serious illness of a friend, moving to a new school district, the divorce/separation of one’s parents, suspension from school, and the presence of a variety of adverse conditions at work (see Williams and Uchiyama, 1989, for an overview of life-events scales for adolescents; see Compas, 1987, and Compas and Phares, 1991, for overviews of research on adolescent stress).’

Drawing on the stress literature, then, one may state that a second type of strain or negative relationship involves the actual or anticipated removal (loss) of positively valued stimuli from the individual. As indicated above, numerous examples of such loss can be found in the inventories of stressful life events. The actual or anticipated loss of positively valued stimuli may lead to delinquency as the individual tries to prevent the loss of the positive stimuli, retrieve the lost stimuli or obtain substitute stimuli, seek revenge

may not always be the case. In certain situations, for example, one may make local com- parisons but not referential comparisons (see Brickman and Bulman, 1977; Crosby and Gonzales-Intal, 1984). In other situations, social comparison processes may not come into play at all; outcomes may be evaluated in terms of culturally derived standards (see Folger, 1986).

6. The stress literature has also focused on positive events, based on the assumption that such events might lead to stress by overloading the individual. Accumulating evi- dence, however, suggests that it is only undesirable events that lead to negative outcomes such as depression (e.g., Gersten et al., 1974; Kaplan et al., 1983; Pearlin et al., 1981; Thoits, 1983).

Certain individuals have criticized the stress literature for neglecting the failure of individuals to achieve positively valued goals. In particular, it has been charged that the stress literature has neglected “nonevents,” or events that are desired or anticipated but do not occur (Dohrenwend and Dohrenwend, 1974; Thoits, 1983). One major distinction between the strain literature in criminology and the stress literature in medical sociology, in fact, is that the former has focused on “nonevents” while the latter has focused on “events.”

7.

58 AGNEW

against those responsible for the loss, or manage the negative affect caused by the loss by taking illicit drugs. While there are no data bearing directly on this type of strain, experimental data indicate that aggression often occurs when positive reinforcement previously administered to an individual is with- held or reduced (Bandura, 1973; Van Houten, 1983). And as discussed below, inventories of stressful life events, which include the loss of positive stimuli, are related to delinquency.

STRAIN AS THE PRESENTATION OF NEGATIVE STIMULI

The literature on stress and the recent psychological literature on aggres- sion also focus on the actual or anticipated presentation of negative or nox- ious stimuli.* Except for the work of Agnew (1985a), however, this category of strain has been neglected in criminology. And even Agnew does not focus on the presentation of noxious stimuli per se, but on the inability of adoles- cents to escape legally from noxious stimuli. Much data, however, suggest that the presentation of noxious stimuli may lead to aggression and other negative outcomes in certain conditions, even when legal escape from such stimuli is possible (Bandura, 1973; Zillman, 1979). Noxious stimuli may lead to delinquency as the adolescent tries to (1) escape from or avoid the negative stimuli; (2) terminate or alleviate the negative stimuli; (3) seek revenge against the source of the negative stimuli or related targets, although the evi- dence on displaced aggression is somewhat mixed (see Berkowitz, 1982; Ber- nard, 1990; Van Houten, 1983; Zillman, 1979); and/or (4) manage the resultant negative affect by taking illicit drugs.

A wide range of noxious stimuli have been examined in the literature, and experimental, survey, and participant observation studies have linked such stimuli to both general and specific measures of delinquency-with the exper- imental studies focusing on aggression. Delinquency/aggression, in particu- lar, has been linked to such noxious stimuli as child abuse and neglect (Rivera and Widom, 1990), criminal victimization (Lauritsen et al., 1991), physical punishment (Straus, 199 l), negative relations with parents (Healy and Bon- ner, 1969), negative relations with peers (Short and Strodtbeck, 1965), adverse or negative school experiences (Hawkins and Lishner, 1987), a wide range of stressful life events (Gersten et al., 1974; Kaplan et al., 1983; Linsky

8. Some researchers have argued that it is often difficult to distinguish the presenta- tion of negative stimuli from the removal of positive stimuli (Michael, 1973; Van Houten, 1983; Zillman, 1979). Suppose, for example, that an adolescent argues with parents. Does this represent the presentation of negative stimuli, (the arguing) or the removal of positive stimuli (harmonious relations with one’s parents)? The point is a valid one, yet the distinc- tion between the two types of strain still seems useful since it helps ensure that all major types of strain are considered by researchers.

A GENERAL STRAIN THEORY 59

and Straus, 1986; Mawson, 1987; Novy and Donohue, 1985; Vaux and Rug- giero, 1983), verbal threats and insults, physical pain, unpleasant odors, dis- gusting scenes, noise, heat, air pollution, personal space violations, and high density (see Anderson and Anderson, 1984; Bandura, 1973, 1983; Berkowitz, 1982, 1986; Mueller, 1983). In one of the few studies in criminology to focus specifically on the presentation of negative stimuli, Agnew (1985a) found that delinquency was related to three scales measuring negative relations at home and school. The effect of the scales on delinquency was partially mediated through a measure of anger, and the effect held when measures of social con- trol and deviant beliefs were controlled. And in a recent study employing longitudinal data, Agnew (1989) found evidence suggesting that the relation- ship between negative stimuli and delinquency was due to the cuusul effect of the negative stimuli on delinquency (rather than the effect of delinquency on the negative stimuli). Much evidence, then, suggests that the presentation of negative or noxious stimuli constitutes a third major source of strain.

Certain of the negative stimuli listed above, such as physical pain, heat, noise, and pollution, may be experienced as noxious largely for biological rea- sons (i.e., they may be unconditioned negative stimuli). Others may be condi- tioned negative stimuli, experienced as noxious largely because of their association with unconditioned negative stimuli (see Berkowitz, 1982). Whatever the case, it is assumed that such stimuli are experienced as noxious regardless of the goals that the individual is pursuing.

THE LINKS BETWEEN STRAIN AND DELINQUENCY Three sources of strain have been presented: strain as the actual or antici-

pated failure to achieve positively valued goals, strain as the actual or antici- pated removal of positively valued stimuli, and strain as the actual or anticipated presentation of negative stimuli. While these types are theoreti- cally distinct from one another, they may sometimes overlap in practice. So, for example, the insults of a teacher may be experienced as adverse because they (1) interfere with the adolescent’s aspirations for academic success, (2) result in the violation of a distributive justice rule such as equity, and (3) are conditioned negative stimuli and so are experienced as noxious in and of themselves. Other examples of overlap can be given, and it may sometimes be difficult to disentangle the different types of strain in practice. Once again, however, these categories are ideal types and are presented only to ensure that all events with the potential for creating strain are considered in empirical research.

Each type of strain increases the likelihood that individuals will experience one or more of a range of negative emotions. Those emotions include disap- pointment, depression, and fear. Anger, however, is the most critical emo- tional reaction for the purposes of the general strain theory. Anger results when individuals blame their adversity on others, and anger is a key emotion

60 AGNEW

because it increases the individual’s level of felt injury, creates a desire for retaliationhevenge, energizes the individual for action, and lowers inhibi- tions, in part because individuals believe that others will feel their aggression is justified (see Averill, 1982; Berkowitz, 1982; Kemper, 1978; Kluegel and Smith, 1986: Ch. 1 0 Zillman, 1979). Anger, then, affects the individual in several ways that are conducive to delinquency. Anger is distinct from many of the other types of negative affect in this respect, and this is the reason that anger occupies a special place in the general strain theory.9 It is important to note, however, that delinquency may still occur in response to other types of negative affect-such as despair, although delinquency is less likely in such cases. 10 The experience of negative affect, especially anger, typically creates a desire to take corrective steps, with delinquency being one possible response. Delinquency may be a method for alleviating strain, that is, for achieving positively valued goals, for protecting or retrieving positive stimuli, or for terminating or escaping from negative stimuli. Delinquency may be used to seek revenge; data suggest that vengeful behavior often occurs even when there is no possibility of eliminating the adversity that stimulated it (Berkowitz, 1982). And delinquency may occur as adolescents try to manage their negative affect through illicit drug use (see Newcomb and Harlow, 1986). The general strain theory, then, has the potential to explain a broad range of delinquency, including theft, aggression, and drug use.

Each type of strain may create a predisposition for delinquency or function as a situational event that instigates a particular delinquent act. In the words of Hirschi and Gottredson (1986), then, the strain theory presented in this paper is a theory of both “criminality” and “crime” (or to use the words of Clarke and Cornish [1985], it is a theory of both “criminal involvement” and “criminal events”). Strain creates a predisposition for delinquency in those cases in which it is chronic or repetitive. Examples include a continuing gap

9. The focus on blame/anger represents a major distinction between the general strain theory and the stress literature. The stress literature simply focuses on adversity, regardless of whether it is blamed on another. This is perhaps appropriate because the major outcome variables of the stress literature are inner-directed states, like depression and poor health. When the focus shifts to outer-directed behavior, like much delinquency, a concern with blarne/anger becomes important.

Delinquency may still occur in the absence of blame and anger (see Berkowitz, 1986; Zillman, 1979). Individuals who accept responsibility for their adversity are still subject to negative affect, such as depression, despair, and disappointment (see Kemper, 1978; Kluegel and Smith, 1986). As a result, such individuals will still feel pressure to take corrective action, although the absence of anger places them under less pressure and makes vengeful behavior much less likely. Such individuals, however, may engage in inner- directed delinquency, such as drug use, and if suitably disposed, they may turn to other forms of delinquency as well. Since these individuals lack the strong motivation for revenge and the lowered inhibitions that anger provides, it is assumed that they must have some minimal disposition for deviance before they respond .o their adversity with outer- directed delinquency (see the discussion of the disposition to delinquency).

10.

A GENERAL STRAIN THEORY 61

between expectations and achievements and a continuing pattern of ridicule and insults from teachers. Adolescents subject to such strain are predisposed to delinquency because (1) nondelinquent strategies for coping with strain are likely to be taxed; (2) the threshold for adversity may be lowered by chronic strains (see Averill, 1982:289); (3) repeated or chronic strain may lead to a hostile a t t i t udea general dislike and suspicion of others and an associated tendency to respond in an aggressive manner (see Edmunds and Kendrick, 1980:2 1); and (4) chronic strains increase the likelihood that individuals will be high in negative affect/arousal at any given time (see Bandura, 1983; Ber- nard, 1990). A particular instance of strain may also function as the situa- tional event that ignites a delinquent act, especially among adolescents predisposed to delinquency. Qualitative and survey data, in particular, sug- gest that particular instances of delinquency are often instigated by one of the three types of strain listed above (see Agnew, 1990; also see Averill, 1982, for data on the instigations to anger).

MEASURING STRAIN As indicated above, strain theory in criminology is dominated by a focus

on strain as goal blockage. Further, only one type of goal blockage is typi- cally examined in the literature-the failure to achieve aspirations, especially aspirations for monetary success or middle-class status. The general strain theory is much broader than current strain theories, and measuring strain under this theory would require at least three sets of measures: those focus- ing on the failure to achieve positively valued goals, those focusing on the loss of positive stimuli, and those focusing on the presentation of negative stimuli. It is not possible to list the precise measures that should be employed in these areas, although the citations above contain many examples of the types of measures that might be used. Further, certain general guidelines for the mea- surement of strain can be offered. The guidelines below will also highlight the limitations of current strain measures and shed further light on why those measures are typically unrelated to delinquency.

DEVELOPING A COMPREHENSIVE LIST OF NEGATIVE RELATIONS

Strain refers to negative or adverse relations with others. Such relations are ultimately defined from the perspective of the individual. That is, in the final analysis adverse relations are whatever individuals say they are (see Berkowitz, 1982). This does not mean, however, that one must employ an idiosyncratic definition of adverse relations4efining adverse relations anew for each person one examines. Such a strategy would create serious problems for (1) the empirical study of delinquency, (2) the prediction and control of delinquency, and (3) efforts to develop the macroimplications of the general

62 AGNEW

strain theory. Rather, one can employ a strategy similar to that followed by stress researchers.

First, one can draw on theory and research to specify those objective situa- tions that might reasonably be expected to cause adversity among adoles- cents. This parallels stress research, which relies on inventories of stressful life events, and several standard inventories are in wide use. The items in such inventories are based, to varying degrees, on the perceptions and judg- ments of researchers, on previous theory and research, and on reports from samples of respondents (see Dohrenwend, 1974). In developing inventories of strainful events, criminologists must keep in mind the fact that there may be important group differences in the types of strain or negative relations most frequently encountered. A list of negative relations developed for one group, then, may overlook certain negative relations important for another group (see Dohrenwend, 1974). It may eventually be possible, however, to develop a comprehensive list of negative relations applicable to most samples of adolescents.

Second, criminologists must recognize that individuals and groups may experience the strainful events in such inventories differently (see Thoits, 1983). Limited data from the stress literature, for example, suggest that the impact of family stressors is greatest among young adolescents, peer stressors among middle adolescents, and academic stressors among old adolescents (Compas and Phares, 1991). Stress researchers have responded to such find- ings not by abandoning their inventories, but by investigating those factors that determine why one group or individual will experience a given event as stressful and another will not. And researchers have identified several sets of variables that influence the perception and experience of negative events (e.g., Compas and Phares, 1991; Pearlin, 1982; Pearlin and Schooler, 1978). Many of the variables are discussed in the next section, and they represent a major set of conditioning variables that criminologists should consider when exam- ining the impact of strainful events on delinquency.

EXAMINING THE CUMULATIVE IMPACT OF NEGATIVE RELATIONS In most previous strain research in criminology, the impact of one type of

negative relation on delinquency is examined with other negative relations ignored or held constant. So, for example, researchers will examine the impact of one type of goal blockage on delinquency, ignoring other types of goal blockage and other potential types of strain. This stands in sharp con- trast to a central assumption in the stress literature, which is that stressful life events have a cumulative impact on the individual. Linsky and Straus (1986: 17), for example, speak of the “accumulation theory,” which asserts that “it is not so much the unique quality of any single event but the cumulu- tion of several stressful events within a relatively short time span” that is

A GENERAL STRAIN THEORY 63

consequential. As a result, it is standard practice in the stressful life-events literature to measure stress with a composite scale: a scale that typically sums the number of stressful life events experienced by the individual.

The precise nature of the cumulative effect, however, is unclear. As Thoits (1983:69) points out, stressful events may have an additive or interactive effect on outcome variables. The additive model assumes that each stressor has a fixed effect on delinquency, an effect independent of the level of the other stressors. Somewhat more plausible, perhaps, is the interactive model, which assumes that “a person who has experienced one event may react with even more distress to a second that follows soon after the first . . . two or more events . . . results in more distress than would be expected from the simple sum of their singular effects.”

Whether the effect is additive or interactive, there is limited support for the idea that the level of stredstrain must pass a certain threshold before nega- tive outcomes result (Linsky and Straus, 1986; Thoits, 1983). Below that level, stresdstrain is unrelated to negative outcomes. Above that level, stredstrain has a positive effect on negative outcomes, perhaps an additive effect or perhaps an interactive effect.

Given these arguments, one should employ a composite index of strain in all analyses or examine the interactions between strain variables. Examining interactions can become very complex if there are more than a few indicators of strain, although it does allow one to examine the differential importance of various types of strain. If stressors have an interactive effect on delinquency, the interaction terms should be significant or the composite index should have a nonlinear effect on delinquency (see the discussion of interactions and non- linear effects in Aiken and West, 1991). If the effect is additive, the interac- tion terms should be insignificant or the composite index should have a linear effect on delinquency (after the threshold level is reached). These issues have received only limited attention in the stress literature (see the review by Thoits, 1983), and they should certainly be explored when constructing meas- ures of strain for the purpose of explaining delinquency. At a minimum, however, as comprehensive a list of negative events/conditions as possible should be examined.

There is also the issue of whether positive events/experiences should be examined. If prior stressors can aggravate the negative effect of subsequent stressors, perhaps positive events can mitigate the impact of stressors. Lim- ited evidence from the stress literature suggests that lists of negative events predict better than lists examining the balance of negative and positive events (usually negative events minus positive events) (see Thoits, 198358-59; Wil- liams and Uchiyama, 1989:lOl; see Gersten et al., 1974, for a possible excep- tion). This topic, however, is certainly in need of more research. In addition to looking at the diference between desirable and undesirable events,

64 AGNEW

researchers may also want to look at the ratio of undesirable to desirable events.

I t should be noted that tests of strain theory in criminology typically examine the disjunction between aspirations and expectations for one or two goals and ignore all of the many other types of strain. The tests also typically assume that strain has a linear effect on delinquency, and they never examine positive as well as negative events. These facts may constitute additional rea- sons for the weak empirical support given to strain theory in criminology.

EXAMINING THE MAGNITUDE, RECENCY, DURATION, AND CLUSTERING OF ADVERSE EVENTS

Limited research from the stress and equity literatures suggest that adverse events are more influential to the extent that they are (1) greater in magnitude or size, (2) recent, (3) of long duration, and (4) clustered in time.

MAGNITUDE

The magnitude of an event has different meanings depending on the type of strain being examined. With respect to goal blockage, magnitude refers to the size of the gap between one’s goals and reality. With respect to the loss of positive stimuli, magnitude refers to the amount that was lost. And with respect to the presentation of noxious stimuli, magnitude refers to the amount of pain or discomfort inJlicted.11 In certain cases, magnitude may be mea- sured in terms of a standard metric, such as dollars or volts delivered. In most cases, however, there is no standard metric available for measuring magnitude and one must rely on the perceptions of individuals (see Jasso, 1980, on quality versus quantity goods). To illustrate, researchers in the stress literature have asked judges to rate events according to the amount of readjustment they require or the threat they pose to the individual (see Thoits, 1983, for other weighting schemes). Such judgments are then aver- aged to form a magnitude score for each event. There is evidence, however, of subgroup differences in weights assigned (Thoits, 198353 -55).

Magnitude ratings are then sometimes used to weight the events in com- posite scales. A common finding, however, is that lists of life events weighted by magnitude do not predict any better than unweighed lists (e.g., Gersten et al., 1974). This is due to the fact that the correlation between lists of

1 1 . As Empey (1956) and others have pointed out, magnitude may also be measured in relative terms. For example, suppose an individual earning $10,OOO a year and an indi- vidual earning $lOO,OOO both lose $100 in a burglary. In absolute terms, the magnitude of their loss is the same. Relative to their current income, however, the magnitude of their loss is quite different. In most cases, it would be difficult to develop precise measures of relative magnitude. Nevertheless, researchers should at the very least be sensitive to this issue when analyzing and interpreting data.

A GENERAL STRAIN THEORY 65

weighted and unweighted events is typically so high (above .90) that the lists can be considered virtually identical (Thoits, 1983). Williams and Uchiyama (1989:99-1OO) explain this high correlation by arguing that severe life events, which are heavily weighted, have a low base rate in the population and so do not have a significant impact on scale scores. Studies that consider major and minor events separately tend to find that major events are in fact more conse- quential than minor events (Thoits, 1983:66).

It should be noted that the previous research on strain theory has paid only limited attention to the dimension of magnitude, even in those cases in which standard metrics for measuring magnitude were available. Samples, in fact, are often simply divided into strained and nonstrained groups, with little con- sideration of variations in the magnitude of strain.

RECENCY

Certain data suggest that recent events are more consequential than older events and that events older than three months have little effect (Avison and Turner, 1988). Those data focus on the impact of stress on depression, and so are not necessarily generalizable to the strain-delinquency relationship. Nev- ertheless, the data suggest that the recency of strain may be an important dimension to consider, and findings in this area might be of special use in designing longitudinal studies, in which the issue of the appropriate lag between cause and effect is central (although the subject of little research and theory).

DURATION

Much theory and data from the equity and stress literatures suggest that events of long duration (chronic stressors) have a greater impact on a variety of negative psychological outcomes (Folger, 1986; Mark and Folger, 1984; Pearlin, 1982; Pearlin and Lieberman, 1979; Utne and Kidd, 1980). Some evidence, in fact, suggests that discrete events may be unimportant except to the extent that they affect chronic events (Cummings and El-Sheikh, 1991; Gersten et al., 1977; Pearlin, 1983). Certain researchers in the equity/justice literature have suggested that the expected duration of the event into the future should also be considered (Folger, 1986; Mark and Folger, 1984; Utne and Kidd, 1980 see especially the “likelihood of amelioration” concept).

CLUSTERING

Data from the stress literature also suggest that events closely clustered in time have a greater effect on negative outcomes (Thoits, 1983). Such events, according to Thoits (1983), are more likely to overwhelm coping resources than evmts spread more evenly over time. Certain data, in particular, trace

66 AGNEW

negative outcomes such as suicide and depression to a series of negative events clustered in the previous few weeks (Thoits, 1983).

ADAPTATIONS TO (COPING STRATEGIES FOR) STRAIN

The discussion thus far has focused on the types of strain that might pro- mote delinquency. Virtually all strain theories, however, acknowledge that only some strained individuals turn to delinquency. Some effort has been made to identify those factors that determine whether one adapts to strain through delinquency. The most attention has been focused on the adoles- cent’s commitment to legitimate means and association with other strained/ delinquent individuals (see Agnew, 199 lb).

The following discussion builds on this effort and is in two parts. First, the major adaptations to strain are described. This discussion points to a number of cognitive, emotional, and behavioral coping strategies that have not been considered in the criminology literature. Second, those factors that influence whether one adapts to strain using delinquent or nondelinquent means are described. This discussion also expands on the criminology literature to include several additional factors that affect the choice of adaptation.

ADAPTATIONS TO STRAIN

What follows is a typology of the major cognitive, emotional, and behav- ioral adaptations to strain, including delinquency.

COGNITIVE COPING STRATEGIES

Several literatures suggest that individuals sometimes cognitively reinter- pret objective stressors in ways that minimize their subjective adversity. Three general strategies of cognitive coping are described below; each strat- egy has several forms. These strategies for coping with adversity may be summarized in the following phrases: “It’s not important,” “It’s not that bad,” and “I deserve it.” This typology represents a synthesis of the coping strategies described in the stress, equity, stratification, and victimization liter- atures (Adams, 1963, 1965; Agnew, 1985b; Agnew and Jones, 1988; Averill, 1982; Della Fave, 1980; Donnerstein and Hatfield, 1982; Pearlin and Schooler, 1978; Walster et al., 1973, 1978). The stress literature, in particu- lar, was especially useful. Stress has been found to have a consistent, although weak-to-moderate, main effect on outcome variables. Researchers have tried to explain this weak-to-moderate effect by arguing that the impact of stressors is conditioned by a number of variables, and much of the atten- tion has been focused on coping strategies (see Compas and Phares, 1991; Thoits, 1984).

A GENERAL STRAIN THEORY 67

Ignore/Minimize the Importance of Adversity. The subjective impact of objective strain depends on the extent to which the strain is related to the central goals, values, and/or identities of the individual. As Pearlin and Schooler (1978:7) state, individuals may avoid subjective strain “to the extent that they are able to keep the most strainful experiences within the least val- ued areas of their life.” Individuals, therefore, may minimize the strain they experience by reducing the absolute and/or relative importance assigned to goaWvalues and identities (see Agnew, 1983; Thoits, 1991a).

In particular, individuals may claim that a particular goal/value or identity is unimportant in an absolute sense. They may, for example, state that money or work is unimportant to them. This strategy is similar to Merton’s adapta- tions of ritualism and retreatism, and it was emphasized by Hyman (1953). Individuals may also claim that a particular goal/value or identity is unim- portant in a relative sense-relative to other goaldvalues or identities. They may, for example, state that money is less important than status or that work is less important than family and leisure activities.

The strategy of minimizing strain by reducing the absolute and/or relative emphasis placed on goals/values and identities has not been extensively examined in the strain literature. Certain evidence, however, suggests that it is commonly employed and may play a central role in accounting for the limited empirical support for strain theory. In particular, research on goals suggests that people pursue a wide variety of different goals and that they tend to place the greatest absolute and relative emphasis on those goals they are best able to achieve (Agnew, 1983; McClelland, 1990; Rosenberg, 1979:265-269; Wylie, 1979).

Maximize Positive Outcomes/Minimize Negative Outcomes. In the above adaptation, individuals acknowledge the existence of adversity but relegate such adversity to an unimportant area of their life. In a second adaptation, individuals attempt to deny the existence of adversity by maximizing their positive outcomes and/or minimizing their negative outcomes. This may be done in two ways: lowering the standards used to evaluate outcomes or dis- torting one’s estimate of current and/or expected outcomes.

Lowering one’s standards basically involves lowering one’s goals or raising one’s threshold for negative stimuli (see Suls, 1977). Such action, of course, makes one’s current situation seem less adverse than it otherwise would be. Individuals may, for example, lower the amount of money they desire (which is distinct from lowering the importance attached to money). This strategy is also related to Merton’s adaptations of ritualism and retreatism, and many of the critics of strain theory in criminology have focused on it. Hyman (1953) and others have argued that poor individuals in the United States are not strained because they have lowered their success goals-bringing their aspira- tions in line with reality. The data in this area are complex, but they suggest

AGNEW

that this adaptation is employed by s o m e b u t not all-lower-class individu- als (see Agnew, 1983, 1986; Agnew and Jones, 1988; see Cloward and Ohlin, 1960, and Empey, 1956, for data on “relative” aspirations).

In addition to lowering their standards, individuals may also cognitively distort their estimate of outcomes. As Agnew and Jones (1988) demonstrate, many individuals exaggerate their actual and expected levels of goal achieve- ment. Individuals with poor grades, for example, often report that they are doing well in school. And individuals with little objective chance of attending college often report that they expect to attend college. (See Wylie, 1979, for additional data in this area.) In addition to exaggerating positive outcomes, individuals may also minimize negative outcomes+laiming that their losses are small and their noxious experiences are mild.

The self-concept literature discusses the many strategies individuals employ to accomplish such distortions (see Agnew and Jones, 1988; Rosen- berg, 1979). Two common strategies, identified across several literatures, are worth noting. In “downward comparisons,” individuals claim that their situ- ation is less worse or at least no worse than that of similar others (e.g., Brick- man and Bulman, 1977; Gruder, 1977; Pearlin and Schooler, 1978; Suls, 1977). This strategy is compatible with the equity literature, which suggests that one’s evaluation of outcomes is conditioned by the outcomes of compari- son others. Temporal comparisons may also be made, with individuals claim- ing that their situation is an improvement over the past. Recent research on the social comparison process suggests that individuals often deliberately make downward comparisons, especially when self-esteem is threatened (Gruder, 1977; Hegtvedt, 1991b; Suls, 1977). In a second strategy, “compen- satory benefits,” individuals cast “about for some positive attribute or circum- stance within a troublesome situation . . . the person is aided in ignoring that which is noxious by anchoring his attention to what he considers the more worthwhile and rewarding aspects of experience” (Pearlin and Schooler, 1978:&7). Crime victims, for example, often argue that their victimization benefited them in certain ways, such as causing them to grow as a person (Agnew, 1985b).

Accept Responsibility for Adversity. Third, individuals may minimize the subjective adversity of objective strain by convincing themselves that they deserve the adversity they have experienced. There are several possible rea- sons why deserved strain is less adverse than undeserved strain. Undeserved strain may violate the equity principle, challenge one’s “belief in a just world” (see Lerner, 1977), and-if attributed to the malicious behavior of another- lead one to fear that it will be repeated in the future. Such reasons may help explain why individuals who make internal attributions for adversity are less distressed than others (Kluegel and Smith, 1986; Mirowsky and Ross, 1990).

A GENERAL STRAIN THEORY 69

Drawing on equity theory, one may argue that there are two basic strate- gies for convincing oneself that strain is deserved. First, individuals may cog- nitively minimize their positive inputs or maximize their negative inputs to a relationship. Inputs are conceived as contributions to the relationship and/or status characteristics believed to be relevant to the relationship (see Cook and Yamagishi, 1983). Second, individuals may maximize the positive inputs or minimize the negative inputs of others. Della Fave (1980) uses both of these strategies to explain the legitimation of inequality in the United States. Those at the bottom of the stratification system are said to minimize their own traits and exaggerate the positive traits and contributions of those above them. They therefore come to accept their limited outcomes as just (also see Kluegel and Smith, 1986; Shepelak, 1987).

BEHAVIORAL COPING STRATEGIES

There are two major types of behavioral coping: those that seek to mini- mize or eliminate the source of strain and those that seek to satisfy the need for revenge.

Maximizing Positive OutcomedMinimizing Negative Outcomes. Behavioral coping may assume several forms, paralleling each of the major types of strain. Individuals, then, may seek to achieve positively valued goals, protect or retrieve positively valued stimuli, or terminate or escape from negative stimuli. Their actions in these areas may involve conventional or delinquent behavior. Individuals seeking to escape from an adverse school environment, for example, may try to transfer to another school or they may illegally skip school. This rather broad adaptation encompasses Merton’s adaptations of innovation and rebellion, as well as those coping strategies described in the equity literature as “maximizing one’s outcomes,” “minimizing one’s inputs,” and “maximizing the other’s inputs.”

Vengeful Behavior. Data indicate that when adversity is blamed on others it creates a desire for revenge that is distinct from the desire to end the adver- sity. A second method of behavioral coping, then, involves the taking of revenge. Vengeful behavior may also assume conventional or delinquent forms, although the potential for delinquency is obviously high. Such behav- ior may involve efforts to minimize the positive outcomes, increase the nega- tive outcomes, and/or increase the inputs of others (as when adolescents cause teachers and parents to work harder through their incorrigible behavior).

EMOTIONAL COPING STRATEGIES Finally, individuals may cope by acting directly on the negative emotions

that result from adversity. Rosenberg (1990), Thoits (1984, 1989, 1990,

70 AGNEW

1991b), and others list several strategies of emotional coping. They include the use of drugs such as stimulants and depressants, physical exercise and deep-breathing techniques, meditation, biofeedback and progressive relaxa- tion, and the behavioral manipulation of expressive gestures through playact- ing or “expression work.” In all of these examples, the focus is on alleviating negative emotions rather than cognitively reinterpreting or behaviorally alter- ing the situation that produced those emotions. Many of the strategies are beyond the reach of most adolescents (Compas et al., 1988), and data indicate that adolescents often employ illicit drugs to cope with life’s strains (Labouvie, 1986a, 1986b; Newcomb and Harlow, 1986). Emotional coping is especially likely when behavioral and cognitive coping are unavailable or unsuccessful.

It should be noted that individuals may employ more than one of the above coping strategies (see Folkman, 199 1). Also, still other coping strategies, such as distraction, could have been listed. It is assumed, however, that the above strategies constitute the primary responses to strain.

PREDICTING THE USE OF DELINQUENT VERSUS NONDELINQUENT ADAPTATIONS

The above typology suggests that there are many ways to cope with strain-nly some of which involve delinquency. And data from the stress literature suggest that individuals vary in the extent to which they use the different strategies (Compas et al., 1988; Menaghan, 1983; Pearlin and Schooler, 1978). These facts go a long way toward explaining the weak sup- port for strain theory. With certain limited exceptions, the strategies are not taken into account in tests of strain theory.

The existence of the above coping strategies poses a serious problem for strain theory. If strain theory is to have any value, it must be able to explain the selection of delinquent versus nondelinquent adaptations. This issue has, of course, been raised before. Critics contend that Merton and other strain theorists fail to explain adequately why only some strained individuals turn to delinquency. This issue, however, is all the more pressing when one considers the full range of nondelinquent adaptations to strain listed above. It is there- fore important to specify those factors that influence the choice of delinquent versus nondelinquent coping strategies.

The following discussion of influencing factors draws on the aggression, equity, and stress literatures (see especially Adams, 1965; Menaghan, 1982; Pearlin and Shooler, 1978; Walster et al., 1978). The aggression literature in psychology is especially useful. Adversity is said to produce a general state of arousal, which can facilitate a variety of behaviors. Whether this arousal results in aggression is said to be determined by a number of factors, many of which are noted below (see Bandura, 1973, 1983; Berkowitz, 1978, 1982).

A GENERAL STRAIN THEORY 71

Those factors affect the choice of coping strategies by affecting (1) the con- straints to nondelinquent and delinquent coping and (2) the disposition to engage in nondelinquent versus delinquent coping.

CONSTRAINTS TO NONDELINQUENT AND DELINQUENT COPING

While there are many adaptations to objective strain, those adaptations are not equally available to everyone. Individuals are constrained in their choice of adaptation(s) by a variety of internal and external factors. The following is a partial list of such factors.

Initial Goals/Values/Identities of the Individual. If the objective strain affects goals/values/identities that are high in absolute and relative impor- tance, and if the individual has few alternative goals/values/identities in which to seek refuge, it will be more difficult to relegate strain to an unimpor- tant are8 of one's life (see Agnew, 1986; Thoits, 1991a). This is especially the case if the goals/values/identities receive strong social and cultural support (see below). As a result, strain will be more likely to lead to delinquency in such cases.

Individual Coping Resources. A wide range of traits can be listed in this area, including temperament, intelligence, creativity, problem-solving skills, interpersonal skills, self-efficacy, and self-esteem. These traits affect the selec- tion of coping strategies by influencing the individual's sensitivity to objective strains and ability to engage in cognitive, emotional, and behavioral coping (Agnew, 1991a; Averill, 1982; Bernard, 1990; Compas, 1987; Edmunds and Kendrick, 1980; Slaby and Guerra, 1988; Tavris, 1984). Data, for example, suggest that individuals with high self-esteem are more resistant to stress (Averill, 1982; Compas, 1987; Kaplan, 1980; Pearlin and Schooler, 1978; Rosenberg, 1990; Thoits, 1983). Such individuals, therefore, should be less likely to respond to a given objective strain with delinquency. Individuals high in self-efficacy are more likely to feel that their strain can be alleviated by behavioral coping of a nondelinquent nature, and so they too should be less likely to respond to strain with delinquency (see Bandura, 1989, and Wang and Richarde, 1988, on self-efficacy; see Thoits, 1991b, on perceived control).

Conventional Social Support. Vaux (1988) provides an extended discussion of the different types of social support, their measurement, and their effect on outcome variables. Thoits (1984) argues that social support is important because it facilitates the major types of coping. The major types of social support, in fact, correspond to the major types of coping listed above. Thus, there is informational support, instrumental support, and emotional support

72 AGNEW

(House, 198 1). Adolescents with conventional social supports, then, should be better able to respond to objective strains in a nondelinquent manner.

Constraints to Delinquent Coping. The crime/delinquency literature has focused on certain variables that constrain delinquent coping. They include (1) the costs and benefits of engaging in delinquency in a particular situation (Clarke and Cornish, 1985), (2) the individual’s level of social control (see Hirschi, 1969), and (3) the possession of those “illegitimate means” necessary for many delinquent acts (see Agnew, 1991a, for a full discussion).

Macro-Level Variables. The larger social environment may affect the probability of delinquent versus nondelinquent coping by affecting all of the above factors. First, the social environment may affect coping by influencing the importance attached to selected goals/values/identities. For example, certain ethnographic accounts suggest that there is a strong social and cul- tural emphasis on the goals of money/status among certain segments of the urban poor. Many poor individuals, in particular, are in a situation in which (1) they face strong economic/status demands, (2) people around them stress the importance of money/status on a regular basis, and (3) few alternative goals are given cultural support (Anderson, 1978; MacLeod, 1987; Sullivan, 1989). As such, these individuals should face more difficulty in cognitively minimizing the importance of money and status.

Second, the larger social environment may affect the individual’s sensitivity to particular strains by influencing the individual’s beliefs regarding what is and is not adverse. The subculture of violence thesis, for example, is predi- cated on the assumption that young black males in urban slums are taught that a wide range of provocations and insults are highly adverse. Third, the social environment may influence the individual’s ability to minimize cogni- tively the severity of objective strain. Individuals in some environments are regularly provided with external information about their accomplishments and failings (see Faunce, 1989), and their attempts at cognitively distorting such information are quickly challenged. Such a situation may exist among many adolescents and among those who inhabit the “street-comer world” of the urban poor. Adolescents and those on the street comer live in a very “public world”; one’s accomplishments and failings typically occur before a large audience or they quickly become known to such an audience. Further, accounts suggest that this audience regularly reminds individuals of their accomplishments and failings and challenges attempts at cognitive distortion.

Fourth, certain social environments may make it difficult to engage in behavioral coping of a nondelinquent nature. Agnew (1985a) has argued that adolescents often find it difficult to escape legally from negative stimuli, espe- cially negative stimuli encountered in the school, family, and neighborhood. Also, adolescents often lack the resources to negotiate successfully with

A GENERAL STRAIN THEORY 73

adults, such as parents and teachers (although see Agnew, 1991a). Similar arguments might be made for the urban underclass. They often lack the resources to negotiate successfully with many others, and they often find it difficult to escape legally from adverse environments-by, for example, quit- ting their job (if they have a job) or moving to another neighborhood.

The larger social environment, then, may affect individual coping in a vari- ety of ways. And certain groups, such as adolescents and the urban under- class, may face special constraints that make nondelinquent coping more difficult. This may explain the higher rate of deviance among these groups.

FACTORS AFFECTING THE DISPOSITION TO DELINQUENCY

The selection of delinquent versus nondelinquent coping strategies is not only dependent on the constraints to coping, but also on the adolescent’s dis- position to engage in delinquent versus nondelinquent coping. This disposi- tion is a function of (1) certain temperamental variables (see Tonry et al., 1991), (2) the prior learning history of the adolescent, particularly the extent to which delinquency was reinforced in the past (Bandura, 1973; Berkowitz, 1982), (3) the adolescent’s beliefs, particularly the rules defining the appropri- ate response to provocations (Bernard’s, 1990, “regulative rules”), and (4) the adolescent’s attributions regarding the causes of his or her adversity. Adoles- cents who attribute their adversity to others are much more likely to become angry, and as argued earlier, that anger creates a strong predisposition to delinquency. Data and theory from several areas, in fact, suggest that the experience of adversity is most likely to result in deviance when the adversity is blamed on another.12 The attributions one makes are influenced by a vari- ety of factors, as discussed in recent reviews by Averill (1982), Berwin (1988), R. Cohen (1982), Crittenden (1983, 1989), Kluegel and Smith (1986), and Utne and Kidd (1980). The possibility that there may be demographic and subgroup differences in the rules for assigning blame is of special interest (see Bernard, 1990; Crittenden, 1983, 1989).

A key variable affecting several of the above factors is association with delinquent peers. It has been argued that adolescents who associate with delinquent peers are more likely to be exposed to delinquent models and beliefs and to receive reinforcement for delinquency (see especially, Akers, 1985). It may also be the case that delinquent peers increase the likelihood that adolescents will attribute their adversity to others.

12. This is a major theme in the psychological research on aggression, in much of the recent research on equity, and in the emotions literature, and it is a central theme in Cloward and Ohlin’s (1960) strain theory (e.g., Averill, 1982; Berkowitz, 1982; R. Cohen, 1982; Crosby and Gonzales-Intal, 1984; Garrett and Libby, 1973; Kemper, 1978; Leventhal, 1976; Mark and Folger, 1984; Martin and Murray, 1984; Weiner, 1982; Zillman, 1979).

74 AGNEW

The individual’s disposition to delinquency, then, may condition the impact of adversity on delinquency. At the same time, it is important to note that continued experience with adversity may create a disposition for delin- quency. This argument has been made by Bernard (1990), Cloward and Ohlin (1960), A. Cohen (1959, Elliott et al. (1979), and others. In particular, it has been argued that under certain conditions the experience of adversity may lead to beliefs favorable to delinquency, lead adolescents to join or form delinquent peer groups, and lead adolescents to blame others for their misfortune.

Virtually all empirical research on strain theory in criminology has neglected the constraints to coping and the adolescent’s disposition to delin- quency. Researchers, in particular, have failed to examine whether the effect of adversity on delinquency is conditioned by factors such as self-efficacy and association with delinquent peers. This is likely a major reason for the weak empirical support for strain theory.

CONCLUSION Much of the recent theoretical work in criminology has focused on the

integration of different delinquency theories. This paper has taken an alterna- tive track and, following Hirschi’s (1979) advice, has focused on the refine- ment of a single theory. The general strain theory builds upon traditional strain theory in criminology in several ways. First, the general strain theory points to several new sources of strain. In particular, it focuses on three cate- gories of strain or negative relationships with others: (1) the actual or antici- pated failure to achieve positively valued goals, (2) the actual or anticipated removal of positively valued stimuli, and (3) the actual or anticipated presen- tation of negative stimuli. Most current strain theories in criminology only focus on strain as the failure to achieve positively valued goals, and even then the focus is only on the disjunction between aspirations and expectations/ actual achievements. The disjunctions between expectations and achieve- ments and @/fair outcomes and achievements are ignored. The general strain theory, then, significantly expands the focus of strain theory to include all types of negative relations between the individual and others.

Second, the general strain theory more precisely specifies the relationship between strain and delinquency, pointing out that strain is likely to have a cumulative effect on delinquency after a certain threshold level is reached. The theory also points to certain relevant dimensions of strain that should be considered in empirical research, including the magnitude, recency, duration, and clustering of strainful events.

Third, the general strain theory provides a more comprehensive account of the cognitive, behavioral, and emotional adaptations to strain. This account sheds additional light on the reasons why many strained individuals do not

A GENERAL STRAIN THEORY 75

turn to delinquency, and it may prove useful in devising strategies to prevent and control delinquency. Individuals, in particular, may be taught those non- delinquent coping strategies found to be most effective in preventing delinquency.

Fourth, the general strain theory more fully describes those factors affect- ing the choice of delinquent versus nondelinquent adaptations. The failure to consider such factors is a fundamental reason for the weak empirical support for strain theory.

Most of the above modifications in strain theory were suggested by research in several areas outside of traditional criminology, most notably the stress research in medical sociology and psychology, the equity/justice research in social psychology, and the aggression research in psychology. With certain exceptions, researchers in criminology have tended to cling to the early strain models of Merton (1938), A. &hen (1955), and Cloward and Ohlin (1960) and to ignore the developments in related fields. And while these early strain models contdn much of value and have had a major influ- ence on the general strain theory in this paper, they do not fully exploit the potential of strain theory.

At the same time, it is important to note that the general strain theory is not presented here as a fully developed alternative to earlier theories. First, the macroimplications of the theory were only briefly discussed. It would not be difficult to extend the general strain theory to the macro level, however; researchers could focus on (1) the social determinants of adversity (for an example, see Bernard, 1990, on the urban underclass) and (2) the social deter- minants of those factors that condition the effect of adversity on delinquency. Second, the theory did not concern itself with the nonsocial determinants of strain, such as illness. It seems doubtful that adversity caused by nonsocial sources is a major source of delinquency because, among other things, it is unlikely to generate anger (see Averill, 1982). Nevertheless, nonsocial sources of adversity should be investigated. Third, the relationship between the general strain theory and other major theories of delinquency must be more fully explored. As hinted earlier, the relationship is rather complex. While the general strain theory is clearly distinct from control and differential association theory, strain may lead to low social control and association with delinquent others. Further, variables from the three theories may interact with one another in producing delinquency. Individuals with delinquent friends, for example, should be more likely to respond to strain with delin- quency. The general strain theory then, is presented as a foundation on which to build.

It is not possible to test the general strain theory fully with currently avail- able data sets, but it is possible to test core sections of the theory. Most data sets dealing with delinquency contain at least some measures of adversity and at least some measures of those factors said to condition the effect of adversity

76 AGNEW

on delinquency. Given this fact, researchers could focus on the following core hypotheses derived from the theory:

First, adverse relations with others will have a positive effect on both gen- eral and specific measures of delinquency, with measures of social control and differential association held constant. This is especially true of adverse rela- tions that are severe and that provide limited opportunities for nondelinquent coping. Prime examples, as discussed earlier, are adverse relations involving family, school, and neighborhood. It is hoped research will point to several measures of strain that are especially relevant to delinquency. Such measures can then be made a routine part of delinquency research, just as the elements of the social bond and measures of differential association are now routinely included in empirical studies.

Second, adverse relations will have a cumulative impact on delinquency after a certain threshold level is reached. Further, this cumulative impact will likely be interactive in nature; each additional increment in strain will have a greater impact than the one before.

Third, the impact of strain or adverse relations on delinquency will be con- ditioned by several variables, as listed above.

Strain theory is the only major theory to focus explicitly on negative rela- tions with others and to argue that delinquency results from the negative affect caused by such relations. As such, it complements social control and differential association/social learning theory in a fundamental way. It is hoped that the general strain theory will revive interest in negative relations and cause criminologists to “bring the bad back in.”

REFERENCES

Adams, J. Stacy 1963

1965

Toward an understanding of inequity. Journal of Abnormal and Social

Inequity in social exchange. In Leonard Ekrkowitz (ed.), Advances in Experimental Social Psychology. New York: Academic Press.

1983 Social class and success goals: An examination of relative and absolute aspirations. Sociological Quarterly 24:435-452.

1984 Goal achievement and delinquency. Sociology and Social Research

1985a A revised strain theory of delinquency. Social Forces 64:151-167. 1985b Neutralizing the impact of crime. Criminal Justice and Behavior

1986

Psychology 67~422-436.

Agnew, Robert

68:435-45 1.

12~221-239. Challenging strain theory: An examination of goals and goal-blockage. Paper presented at the annual meeting of the American Society of Criminology, Atlanta. A longitudinal test of the revised strain theory. Journal of Quantitative Criminology 5:373 -387.

1989

A GENERAL STRAIN THEORY 77

1990 The origins of delinquent events: An examination of offender accounts. Journal of Research in Crime and Delinquency 27:267-294.

199 la Adolescent resources and delinquency. Criminology 28535-566. 1991b Strain and subcultural crime theory. In Joseph Sheley (ed.), Criminology: A

Contemporary Handbook. Belmont, Calif.: Wadsworth.

Adapting to deprivation: An examination of inflated educational expecta- tions. Sociological Quarterly 29:3 15-337.

Multiple Regression: Testing and Interpreting Interactions. Newbury Park, Calif.: Sage.

Agnew, Robert and Diane Jones 1988

Aiken, Leona S. and Stephen G. West 199 1

Akers, Ronald L. 1985 Deviant Behavior: A Social Learning Approach. Belmont, Calif.:

Wadsworth.

Who should get what? Fairness judgments of the distribution of earnings. American Journal of Sociology 84541-564.

Distributive justice and satisfaction with material well-being. American Sociological Review 52:83-95.

A Place on the Comer. Chicago: University of Chicago Press.

Ambient temperature and violent crime: Tests of the linear and curvilinear hypotheses. Journal of Personality and Social Psychology 46:9 1-97.

Equity theory and social comparison processes. In Jerry M. Suls and Richard L. Miller (eds.), Social Comparison Processes. New York: Hemisphere.

Anger and Aggression. New York: Springer-Verlag.

Stressful life events and depressive symptoms: Disaggregating the effects of acute stressors and chronic strains. Journal of Health and Social Behavior

Alves, Wayne M. and Peter H. Rossi 1978

Alwin, Duane F. 1987

Anderson, Elijah

Anderson, Craig A. and Dona C. Anderson 1978

1984

Austin, William 1977

Averill, James R.

Avison, William R. and R. Jay Turner 1982

1988

29:253-264. Bandura, Albert

Aggression: A Social Learning Analysis. Englewood Cliffs, N.J.: Prentice- Hall. Psychological mechanisms of aggression. In Russell G. Geen and Edward Donnerstein (eds.), Aggression: Theoretical and Empirical Reviews. New York: Academic Press. Human agency and social cognitive theory. American Psychologist

1973

1983

1989 44~1175-1184.

Berger, Joseph, Morris Zelditch, Jr., Bo Anderson, and Bernard Cohen 1972 Structural aspects of distributive justice: A status-value formulation. In

Joseph Berger, Morris Zelditch, Jr., and Bo Anderson (eds.), Sociological Theories in Progress. New York: Houghton Mifflin.

AGNEW

Berger, Joseph, M. Hamit Fisck, Robert Z. Norman, and David G. Wagner 1983 The formation of reward expectations in status situations. In David M.

Messick and Karen S. Cook (eds.), Equity Theory: Psychological and S&iological Perspectives. New York: Praeger.

Whatever happened to the frustration-aggression hypothesis? American Behavioral Scientist 21 :691-708. Aversive conditions as stimuli to aggression. In Leonard Berkowitz (ed.), Advances in Experimental Social Psychology. Vol. 15. New York: Academic Press. A Survey of Social Psychology. New York: Holt, Rinehart & Winston.

Control criticisms of strain theories: An assessment of theoretical and empirical adequacy. Journal of Research in Crime and Delinquency

Testing structural strain theories. Journal of Research in Crime and Delinquency 24:262-280. Angry aggression among the “truly disadvantaged.” Criminology 28:73 -96.

Exchange and Power in Social Life. New York: John Wiley & Sons.

Explanation and adaptation in adversity. In Shirley Fisher and James Reason (eds.), Handbook of Life Stress, Cognition and Health. Chichester, England: John Wiley & Sons.

Pleasure and pain in social comparison. In Jerry M. Suls and Richard L. Miller (eds.), Social Comparison Processes. New York: Hemisphere.

Modeling offenders’ decisions: A framework for research and policy. In Michael Tonry and Norval Morris (eds.), Crime and Justice: An Annual Review of Research. Vol. 6. Chicago: University of Chicago Press.

Berkowitz, Leonard 1978

1982

1986 Bernard, Thomas J.

1984

21 :353-372. 1987

1990 Blau, Peter

1964 Brewin, Chris R.

1988

Brickman, Philip and Ronnie Janoff Bulman 1977

Clarke, Ronald V. and Derek B. Cornish 1985

Chard , Marshall B.

Cloward, Richard A. and Lloyd E. Ohlin

Cohen, Albert K.

1964 Anomie and Deviant Behavior. New York: Free Press.

1960 Delinquency and Opportunity. New York: Free Press.

1955 Delinquent Boys. New York: Free Press. 1965 The sociology of the deviant act: Anomie theory and beyond. American

Sociological Review 3 0 5 1 4 .

Perceiving justice: An attributional perspective. In Jerald Greenberg and Ronald L. Cohen (eds.), Equity and Justice in Social Behavior. New York: Academic Press.

The growth of scientific knowledge: Theories of deviance as a case study. In Lewis A. Coser (ed.), The Idea of Social Structure: Papers in Honor of Robert K. Merton. New York: Harcourt Brace Jovanovich.

Cohen, Ronald L. 1982

Cole, Stephen 1975

A GENERAL STRAIN THEORY 79

Compas, Bruce E. 1987 Coping with stress during childhood and adolescence. Psychological Bulletin

101 ~393-403.

Compas, Bruce E., Vanessa L. Malcarne, and Karen M. Fondacaro 1988 Coping with stressful events in older children and young adolescents.

Journal of Consulting and Clinical Psychology 56405-41 1.

Stress during childhood and adolescence: Sources of risk and vulnerability. In E. Mark Cummings, Anita L. Greene, and Katherine H. Karraker (eds.), Life-Span Developmental Psychology: Perspectives on Stress and Coping. Hillsdale, N.J.: Lawrence Erlbaum.

Distributive justice, equity, and equality. Annual Review of Sociology

Empirical evidence of the sense of justice. In Margaret Gruter, Roger D. Masters, Michael T. McGuire (eds.), The Sense of Justice: An Inquiry into the Biological Foundations of Law. New York: Greenwood Press.

Psychological and sociological perspectives on distributive justice: Conver- gent, divergent, and parallel lines. In David M. Mesick and Karen S. Cook (eds.), Equity Theory: Psychological and Sociological Perspectives. New York: Praeger.

Cook, Karen S. and Toshio Yamagishi Social determinants of equity judgments: The problem of multidimensional input. In David M. Messick and Karen S. Cook (eds.), Equity Theory: Psychological and Sociological Perspectives. New York: Praeger.

Sociological aspects of attribution. Annual Review of Sociology 9:425-446. Causal attribution in sociocultural context: Toward a self-presentational theory of attribution processes. Sociological Quarterly 3 0 1-14.

Relative deprivation and equity theories: Felt injustice and the undeserved benefits of others. In Robert Folger (ed.), The Sense on Injustice: Social Psychological Perspectives. New York: Plenum.

Children’s coping with angry environments: A process-oriented approach. In E. Mark Cummings, Anita L. Greene, and Katherine H. Karraker (eds.), Life-Span Developmental Psychology: Perspectives on Stress and Coping. Hillsdale, N.J.: Lawrence Erlbaum.

Success values: Are they universal or class-differentiated? American Journal of Sociology 80:153-169. The meek shall not inherit the earth: Self-evaluations and the legitimacy of stratification. American Sociological Review 45955-97 1.

Della Fave, L. Richard and Patricia Klobus Success values and the value stretch: A biracial comparison. Sociological Quarterly 17:491-502.

Compas, Bruce E. and Vicky Phares 1991

Cook, Karen S., and Karen A. Hegtvedt 1983

1991 9:2 17-24 1.

Cook, Karen S. and David Messick 1983

1983

Crittenden, Kathleen S. 1983 1989

Crosby, Faye and A. Miren Gonzales-Intal 1984

Cummings, E. Mark and Mona El-Sheikh 199 1

Della Fave, L. Richard 1974

1980

1976

80 AGNEW

Deutsch, Morton 1975 Equity, equality, and need: What determines which value will be used as the

basis of distributive justice. Journal of Social Issues 31:137-149.

Problems in defining and sampling the relevant population of stressful life events. In Barbara Snell Dohrenwend and Bruce P. Dohrenwend (eds.), Stressful Life Events: Their Nature and Effects. New York: John Wiley & Sons.

Overview and prospects for research on stressful life events. In Barbara Snell Dohrenwend and Bruce P. Dohrenwend (eds.), Stressful Life Events: Their Nature and Effects. New York: John Wiley & Sons.

Aggression and equity. In Jerald Greenberg and Ronald L. Cohen (eds.), Equity and Justice in Social Behavior. New York: Academic Press.

The Measurement of Human Aggressiveness. New York: John Wiley & Sons.

Delinquency and Dropout. Lexington, Mass.: Lexington Books.

An integrated theoretical perspective on delinquent behavior. Journal of Research in Crime and Delinquency 16:3-27.

Dohrenwend, Bruce P. 1974

Dohrenwend, Barbara Snell and Bruce P. Dohrenwend 1974

Donnerstein, Edward and Elaine Hatfield 1982

Edmunds, G. and D.C. Kendrick 1980

Elliott, Delbert and Harwin Voss

Elliott, Delbtrt, Suzanne Ageton, and Rachel Canter 1974

1979

Elliott, Delbert, David Huizinga, and Suzanne Ageton

Empey, LaMar 1985 Explaining Delinquency and Drug Use. Beverly Hills, Calif.: Sage.

1956

1982

Social class and occupational aspiration: A comparison of absolute and relative measurement. American Sociological Review 21 :703 -709. American Delinquency: Its Meaning and Construction. Homewood. Ill.: Dorsey.

Strain theory revisited: Economic gods, educational means, and delin- quency. American Sociological Review 54263 -274.

Occupational status-assignment systems: The effect of status on self-esteem. American Journal of Sociology 95:378-40.

Emerging issues in the social psychology of justice. In Robert Folger (ed.), The Sense of Injustice: Social Psychological Perspectives. New York: Plenum. Rethinking equity theory: A referent cognitions model. In Hans Werner Bierhoff, Ronald L. Cohen, and Jerald Greenberg (eds.), Justice in Social Relations. New York: Plenum.

Farnworth, Margaret and Michael J. Leiber 1989

Faunce, William A. 1989

Folger, Robert 1984

1986

A GENERAL STRAIN THEORY 81

Folkman, Susan 1991 Coping across the life-span: Theoretical issues. In E. Mark Cummings,

Anita L. Greene, and Katherine H. Karraker (eds.), Life-Span Developmen- tal Psychology: Perspectives on Stress and Coping. Hillsdale, N.J.: Lawrence Erlbaum.

Garrett, James and William L. Libby, Jr. 1973 Role of intentionality in mediating responses to inequity in the dyad.

Journal of Personality and Social Psychology 28:21-27.

Gersten, Joanne C., Thomas S. Langer, Jeanne G. Eisenberg, and Lida Ozek Child behavior and life events: Undesirable change or change per se. In Barbara Snell Dohrenwend and Bruce P. Dohrenwend (eds.), Stressful Life Events: Their Nature and Effects. New York: John Wiley & Sons.

An evaluation of the etiological role of stressful lifethange events in psychological disorders. Journal of Health and Social Behavior 183228-24.

1974

Gersten, Joanne C., Thomas S. Langer, Jeanne G. Eisenberg, and Ora Smith-Fagon. 1977

Greenberg, David F. 1977 Delinquency and the age structure of society. Contemporary Crises

1 : 189-223.

Gruder, Charles L. 1977 Choice of comparison persons in evaluating oneself. In Jerry M. Suls and

Richard L. Miller (eds.), Social Comparison Processes. New York: Hemisphere.

Schooling and delinquency. In Elmer H. Johnson (ed.), Handbook on Crime and Delinquency Prevention. New York: Greenwood.

Healy, William and Augusta F. Bonner New Light on Delinquency and Its Treatment. New Haven, Conn.: Yale University Press.

When rewards are scarce: Equal or equitable distributions. Social Forces

The effects of relationship structure on emotional responses to inequity. Social Psychology Quarterly 53:214-228. Justice processes. In Martha Foschi and Edward J. Lawler (eds.), Group Processes: Sociological Analyses. Chicago: Nelson-Hall. Social comparison processes. In Edgar F. Borgotta and Marie E. Borgotta (eds.), Encyclopedia of Sociology. New York: Macmillan.

Causes of Delinquency. Berkeley: University of California Press. Separate and unequal is better. Journal of Research in Crime and Delinquency 16:34-38.

The distinction between crime and criminality. In Timothy F. Hartnagel and Robert A. Silverman (eds.), Critique and Explanation. New Brunswick, N.J.: Transaction Books.

Hawkins, J. David and Denise M. Lishner 1987

1969

Hegtvedt, Karen A. 1987

1990

1991a

1991b

66: 183 -207.

Hirschi, Travis 1969 1979

Hirschi, Travis and Michael Gottfredson 1986

AGNEW

Hochschild, Jennifer L. 198 1 What’s Fair: American Beliefs about Distributive Justice. Cambridge,

Mass.: Harvard University Press. Homans, George C.

1961 Social Behavior: Its Elementary Forms. New York: Harcourt, Brace and World.

House, James S.

Hyman, Herbert 1981 Work Stress and Social Support. Reading, Mass.: Addison-Wesley.

The value systems of the different classes: A social-psychological contribu- tion to the analysis of stratification. In Reinhard Bendix and Seymour Martin Lipset (eds.), Class, Status, and Power. New York: Free Press.

A new theory of distributive justice. American Sociological Review 453-32.

Distributive justice and earned income. American Sociological Review 42f539-651.

1953

Jasso, Guillermina

Jasso, Guillermina and Peter H. Rossi

1980

1977

Jensen, Gary 1986 Dis-integrating integrated theory: A critical analysis of attempts to save

strain theory. Paper presented at the annual meeting of the American Society of Criminology, Atlanta.

Juvenile Delinquency and Its Origins. London: Cambridge University Press. Johnson, Richard E.

Kaplan, Howard B.

Kaplan, Howard B., Cynthia Robbins, and Steven S. Martin

1979

1980 Deviant Behavior in Defense of Self. New York: Academic Press.

1983 Toward the testing of a general theory of deviant behavior in longitudinal perspective: Patterns of psychopathology. In James R. Greenley and Roberta G. Simmons (eds.), Research in Community and Mental Health. Greenwich, Conn.: Jai Press.

A Social Interactional Theory of Emotions. New York: John Wiley & Sons. Kemper, Theodore D.

Kluegel, James R. and Eliot R. Smith

Kornhauser, Ruth Rosner

Labouvie, Erich W.

1978

1986 Beliefs about Inequality. New York: Aldine De Gruyter.

1978 Social Sources of Delinquency. Chicago: University of Chicago Press.

Alcohol and marijuana use in relation to adolescent stress. International Journal of the Addictions 21:333-345. The coping function of adolescent alcohol and drug use. In Rainer K. Silbereisen, Klaus Eyfeth and Georg Rudinger (eds.), Development as Action in Context. New York: Springer.

1986a

1986b

Lauritsen, Janet L., Robert I. Sampson, and John Laub 199 1 The link between offending and victimization among adolescents. Criminol-

ogy 29~265-292.

A GENERAL STRAIN THEORY 83

Lerner, Melvin J. 1977 The justice motive: Some hypotheses as to its origins and forms. Journal of

Personality 45:l-52.

Leventhal, Gerald S. 1976 The distribution of rewards and resources in groups and organizations. In

Leonard Berkowitz and Elaine Walster (eds.), Advances in Experimental Social Psychology: Equity Theory: Toward a General Theory of Social Interaction. New York: Academic Press.

Leventhal, Gerald S., Jurgis Kanizajr, and William Rick Fry 1980 Beyond fairness: A theory of allocation preferences. In Gerald Mikula (ed.),

Justice and Social Interaction. New York: Springer-Verlag.

Lind, E. Allan and Tom R. Tyler

Linsky, Arnold S. and Murray A. Straus

Liska, Allen E.

1988 The Social Psychology of Procedural Justice. New York: Plenum.

1986 Social Stress in the United States. Dover, Mass.: Auburn House.

1987 Perspectives on Deviance. Englewood Cliffs, N.J.: Prentice-Hall

1990 The social management of ambition. Sociological Quarterly 3 1:225-251.

1987 Ain’t No Makin’ It. Boulder, Colo.: Westview Press.

McClelland, Katherine

MacLeod, Jay

Mark, Melvin M. and Robert Folger 1984 Responses to relative deprivation: A conceptual framework. In Philip

Shaver (ed.), Review of Personality and Social Psychology. Vol. 5. Beverly Hills, Calif.: Sage.

When expectations and justice do not coincide: Blue collar visions of a just world. In Hans Weiner Bierhoff, Ronald L. Cohen, and Jerald Greenberg (eds.), Justice in Social Relations. New York: Plenum.

Martin, Joanne 1986

Martin, Joanne and Alan Murray 1983 Distributive injustice and unfair exchange. In David M. Messick and Karen

S. Cook (eds.), Equity Theory: Psychological and Social Perspectives. New York: Praeger. Catalysts for collective violence: The importance of a psychological approach. In Robert Folger (ed.), The Sense of Injustice: Social Psychologi- cal Perspectives. New York: Plenum.

A longitudinal examination of an integrated social process model of deviant behavior. Social Forces 65: 106134.

1984

Massey, James L. and Marvin Krohn 1986

Mawson, Anthony R. 1987 Criminality: A Model of Stress-Induced Crime. New York: Praeger.

1982 Menaghan, Elizabeth

Measuring coping effectiveness: A panel analysis of marital problems and coping efforts. Journal of Health and Social Behavior 23:220-234.

AGNEW

1983 Individual coping efforts: Moderators of the relationship between life stress and mental health outcomes. In Howard B. Kaplan (ed.), Psychosocial Stress: Trends in Theory and Research. New York: Academic Press.

Social structure and anomie. American Sociological Review 3:672-682.

Fairness and preference. Journal of Experimental Social Psychology

Fairness, preference, and fairness biases. In David M. Messick and Karen S. Cook (eds.), Equity Theory: Psychological and Sociological Perspectives. New York: Praeger.

Positive and negative reinforcement, a distinction that is no longer necessary; or a better way to talk about bad things. In Eugene Ramp and George Semb (eds.), Behavior Analysis: Areas of Research and Application. Englewood Cliffs, N.J.: Prentice-Hall.

The attitude-achievement paradox among black adolescents. Sociology of Education 63:44-61.

Merton, Robert

Messick, David M. and Keith Sentis 1938

1979

1983 15:4 18-434.

Michael, Jack 1973

Mickelson, Roslyn Arlin 1990

Mikula, Gerold 1980 Justice and Social Interaction. New York: Springer-Verlag. 1986 The experience of injustice: Toward a better understanding of its phenome-

nology. In Hans Werner Bierhoff, Ronald L. Cohen, and Jerald Greenberg (eds.), Justice in Social Relations. New York: Plenum.

Mirowsky, John and Catherine E. Ross 1990 The consolation-prize theory of alienation. American Journal of Sociology

95: 1505-1 535.

Morgan, Rick L. and David Heise Structure of emotions. Social Psychology Quarterly 5 1: 19-3 1.

Mueller, Charles W.

1988

1983 Environmental stressors and aggressive behavior. In Russell G. Geen and Edward I. Donnerstein (eds.), Aggression: Theoretical and Empirical Reviews. Vol. 2. New York: Academic Press.

Newcomb, Michael D. and L.L. Harlow 1986 Life events and substance use among adolescents: Mediating effects of

perceived loss of control and meaninglessness in life. Journal of Personality and Social Psychology 5 1564-577.

The relationship between adolescent life stress events and delinquent conduct including conduct indicating a need for supervision. Adolescence

Novy, Diane M. and Stephen Donohue 1985

78:313-321.

Pearlin, Leonard I. 1982

1983

The social contexts of stress. In Leo Goldberger and Shlomo Berznitz (eds.), Handbook of Stress. New York: Free Press. Role strains and personal stress. In Howard Kaplan (ed.), Psychosocial Stress: Trends in Theory and Research. New York: Academic Press.

A GENERAL STRAIN THEORY 85

Pearlin, Leonard 1. and Carmi Schooler The structure of coping. Journal of Health and Social Behavior 19:2-21.

Pearlin, Leonard I. and Morton A. Lieberman Social sources of emotional distress. In Roberta G. Simmons (ed.), Research in Community and Mental Health. Vol. I. Greenwich, Conn.: Jai Press.

1978

1979

Pearlin, Leonard I., Elizabeth G. Menaghan, Morton A. Lieberman, and Joseph T.

The stress process. Journal of Health and Social Behavior 22:337-356.

The effect of goal discrepancy on delinquency. Social Problems 22:7&86.

Mullan 1981

Quicker, John

Rivera, Beverly and Cathy Spatz Widom

1974

1990 Childhood victimization and violent offending. Violence and Victims 5: 19-35.

Rosenberg, Morris 1979 Conceiving the Self. New York: Basic. 1990

1971

Reflexivity and emotions. Social Psychology Quarterly 53:3-12.

Some determinants of the intensity of social protest. Journal of Experimen- tal Social Psychology 7:401-418.

Just allocations of goods: Decisions among three principles. In Gerald Mikula (ed.), Justice and Social Interaction. New York: Springer-Verlag.

The role of self-explanations and self-evaluations in legitimating inequality. American Sociological Review 52:495-503.

Ross, Michael, John Thibaut, and Scott Evenback

Schwinger, Thomas 1980

Shepelak, Norma J. 1987

Shepelak, Norma J. and Duane Alwin 1986 Beliefs about inequality and perceptions of distributive justice. American

Sociological Review 5 1 :30-46. Short, James F. and Fred L. Strodtbeck

1965 Group Process and Gang Delinquency. Chicago: University of Chicago Press.

Slaby, Ronald G. and Nancy G. Guerra 1988 Cognitive mediators of aggression in adolescent offenders: 1. Developmental

Psychology 24: 5 80-5 8 8.

The relationship between inequity and emotions in close relationships. Social Psychology Quarterly 49:309-32 1.

An experimental study of some effects of injustice on children’s moral behavior. Journal of Experimental Social Psychology 4:460-469.

Discipline and deviance: Physical punishment of children and violence and other crimes in adulthood. Social Problems 38:133-154.

Sprecher, Susan 1986

Stephenson, G.M. and J.H. White 1968

Straw, Murray 1991

Sullivan, Mercer L. 1989 Getting Paid. Ithaca, N.Y.: Cornell University Press.

86 AGNEW

Suls, Jerry M. 1977 Social comparison theory and research: An overview from 1954. In Jerry

M. Suls and Richard L. Miller (eds.), Social Comparison Processes. New York: Hemisphere.

Suls, Jerry M. and Thomas Ashby Wills 1991 Social Comparison: Contemporary Theory and Research. Hillsdale, N.J.:

Lawrence Erlbaum.

Tavris, Carol 1984 On the wisdom of counting to ten. In Philip Shaver (ed.), Review of

Personality and Social Psychology: 5. Beverly Hills, Calif.: Sage.

The Social Psychology of Groups. New York: John Wiley & Sons. Thibaut, John W. and Harold H. Kelley

Thoits, Peggy

1959

1983

1984

1989

1990

1991a

1991b

Dimensions of life events that influence psychological distress: An evalua- tion and synthesis of the literature. In Howard B. Kaplan (ed.), Psychosocial Stress: Trends in Theory and Research. New York: Academic Press. Coping, social support, and psychological outcomes: The central role of emotion. In Philip Shaver (ed.), Review of Personality and Social Psychology: 5. Beverly Hills, Calif.: Sage. The sociology of emotions. In W. Richard Scott and Judith Blake (eds.), Annual Review of Sociology. Vol. 15. Palo Alto, Calif.: Annual Reviews. Emotional deviance research. In Theodore D. Kemper (ed.), Research Agendas in the Sociology of Emotions. Albany: State University of New York Press. On merging identity theory and stress research. Social Psychology Quarterly

Patterns of coping with controllable and uncontrollable events. In E. Mark Cummings, Anita L. Greene, and Katherine H. Karraker (eds.), Life-Span Developmental Psychology: Perspectives on Stress and Coping. Hillsdale, N.J.: Lawrence Erlbaum.

54:101-112.

Thornberry, Terence P.

Tonry, Michael, Lloyd E. Ohlin, and David P. Farrington

Tornblum, Kjell Y.

Utne, Mary Kristine and Robert Kidd

1987 Toward an Interactional Theory of Delinquency. Criminology 25:863 -89 1.

1991 Human Development and Criminal Behavior. New York: Springer-Verlag.

1977 Distributive justice: Typology and propositions. Human Relations 30: 1-24.

1980 Equity and attribution. In Gerald Mikula (ed.), Justice and Social Interaction. New York: Springer-Verlag.

Van Houten, Ron 1983 Punishment: From the animal laboratory to the applied setting. In Saul

Axelrod and Jack Apsche (eds.), The Effects of Punishment on Human Behavior. New York: Academic Press.

Vaux, Alan 1988 Social support: Theory, Research, and Intervention. New York: Praeger.

A GENERAL STRAIN THEORY

Vaux, Alan and Mary Ruggiero 1983 Stressful life change and delinquent behavior. American Journal of

Community Psychology 11:169-183.

New directions in equity research. Journal of Personality and Social Walster, Elaine, Ellen Berscheid, and G. William Walster

1973 Psychology 25:151-176.

Walster, Elaine, G. William Walster, and Ellen Berscheid

Wang, Alvin Y. and R. Stephen Richarde 1978 Equity: Theory and Research. Boston: Allyn & Bacon.

1988 Global versus task-specific measures of self-efficacy. Psychological Record 38:533-541.

Weiner, Bernard 1982 The emotional consequences of causal attributions. In Margaret S. Clark

and Susan T. Fiske (eds.), Affect and Cognition: The Seventeenth Annual Carnegie Symposium on Cognition. Hillsdale, N.J.: Lawrence Erlbaum.

Assessment of life events during adolescence: The use of self-report inventories. Adolescence 24:95-118.

Williams, Carolyn L. and Craige Uchiyama 1989

Wylie, Ruth

Zillman, Dolf 1979 The Self-concept. Vol. 2. Lincoln: University of Nebraska Press.

1979 Hostility and Aggression. Hillsdale, N.J.: Lawrence Erlbaum.

87

Robert Agnew is Associate Professor of Sociology at Emory University in Atlanta, Geor- gia. His research focuses on the causes of delinquency, particularly strain and social con- trol theories. He is currently conducting an empirical test of the general strain theory with Helene Raskin White.