Deviant Behavior

Routine Activities and Individual Deviant Behavior

Author(s): D. Wayne Osgood, Janet K. Wilson, Patrick M. O'Malley, Jerald G. Bachman and Lloyd D. Johnston

Source: American Sociological Review , Aug., 1996, Vol. 61, No. 4 (Aug., 1996), pp. 635- 655

Published by: American Sociological Association

Stable URL: https://www.jstor.org/stable/2096397

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected]. Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at https://about.jstor.org/terms

American Sociological Association is collaborating with JSTOR to digitize, preserve and extend access to American Sociological Review

This content downloaded from ��������������66.77.17.54 on Sun, 11 Apr 2021 17:31:19 UTC��������������

All use subject to https://about.jstor.org/terms

ROUTINE ACTIVITIES AND INDIVIDUAL DEVIANT BEHAVIOR*

D. Wayne Osgood Janet K. Wilson The Pennsylvania State University University of Central Arkansas

Patrick M. O'Malley Jerald G. Bachman Lloyd D. Johnston University of Michigan University of Michigan University of Michigan

We extend the routine activity perspective's situational analysis of crime to

individual offending and to a broad range of deviant behaviors. In this view,

unstructured socializing with peers in the absence of authority figures pre-

sents opportunities for deviance: In the presence of peers, deviant acts will

be easier and more rewarding; the absence of authority figures reduces the

potential for social control responses to deviance; and the lack of structure

leaves time available for deviant behavior To determine whether individuals

who spend more time in unstructured socializing activities engage in deviant

behaviors more frequently, we analyzed within-individual changes in routine

activities and deviance across five waves of data for a national sample of

more than 1, 700 18- to 26-year-olds. Participation in these routine activities

was strongly associated with criminal behavior, heavy alcohol use, use of

marijuana and other illicit drugs, and dangerous driving. Furthermore, rou-

tine activities accounted for a substantial portion of the association between

these deviant behaviors and age, sex, and socioeconomic status.

he emergence of theories of crime that emphasize the influence of routine ac-

tivities (Cohen and Felson 1979) or lifestyle (Hindelang, Gottfredson, and Garofalo 1978) is one of the most significant developments in the study of deviance over the past two decades.1 This situational approach shifts at-

tention away from the personal histories of offenders toward the dependence of crime on opportunities presented by the routine activi- ties of everyday life. Birkbeck and LaFree (1993) note that this shift corresponds to Sutherland's (1947) distinction between his- torical explanations, which account for crime by past events, and situational explanations, which account for crime by the circum- stances in which it occurs. Routine activity theorists have applied this situational ap- proach to explain group differences in vic- timization (Hindelang et al. 1978) and trends in aggregate crime rates (Cohen and Felson 1979) in terms of the social structure's im- pact on routine activities.

According to Meier and Miethe (1993: 472-73), sociologists find the routine activ- ity perspective appealing because it identifies

* Direct correspondence to D. Wayne Osgood, 918 Oswald Tower, Program in Crime, Law, and Justice, Department of Sociology, The Pennsyl- vania State University, University Park, PA 16802-6214. While this research was being con- ducted, the first author was at the University of Nebraska-Lincoln, and the second author was at the University of Arkansas at Little Rock. This research was supported by the National Institute of Mental Health (Grant 42033) and the National Institute on Drug Abuse (Grant DA0141 1). The authors thank Robert Agnew, Jeffrey Arnett, Jack Gibbs, David Johnson, Barbara McMorris, Suzanne Ortega, and anonymous ASR reviewers for helpful suggestions, and Dawn Bare and Sun Yong-Min for assistance with preliminary data analysis. [Reviewers acknowledged by the au- thors include Marcus Felson and Robert F. Meier. -ED.]

1 Several authors agree that the lifestyle theory of Hindelang et al. (1978) and the routine activ-

ity theory of Cohen and Felson (1979) do not dif-

fer in substance (Birkbeck and LaFree 1993; Garofalo 1987; Meier and Miethe 1993). We treat these two theoretical positions, and other posi-

tions that are closely related to them (Miethe and Meier 1990, 1994), as versions of a single theo- retical perspective.

American Sociological Review, 1996, Vol. 61 (August:635-655) 635

This content downloaded from ��������������66.77.17.54 on Sun, 11 Apr 2021 17:31:19 UTC��������������

All use subject to https://about.jstor.org/terms

636 AMERICAN SOCIOLOGICAL REVIEW

a symbiotic relationship between conven- tional activities and illegal activities and points to fundamental ironies in links be- tween some otherwise constructive social changes and increasing crime (e.g., women's employment and daytime burglary). The rou- tine activity perspective challenges the com- monplace notion that crime must stem from other "bad" things, an idea that Felson

(1994) has labeled the "pestilence fallacy."' We extend the routine activity perspective

in several ways. First, rather than focusing on

victimization or aggregate crime rates, which studies from this perspective typically do, -we emphasize offending by individuals. This is unusual, as the routine activity perspective is often cited as redirecting the study of crime and deviance away from an exclusive con- cern with the offender. Indeed, the approach is often categorized as a theory of victimiza- tion because most routine activity studies rely on victimization data (Birkbeck and LaFree 1993). Yet the theory's basic predic- tion, that crime depends on routine activities, pertains to individual offending as well. In- deed, a convergence of the study of offend- ing and the study of victimization is implied by evidence that rates of victimization are especially high among offenders (Jensen and Brownfield 1986; Lauritsen, Laub, and Sam- pson 1992).

Thus far, relatively little attention has been given to the implications of the routine ac- tivity perspective for individual offending. Felson has often discussed such themes (1986, 1994; Felson and Gottfredson 1984), but Riley (1987) offers the only empirical study of individual deviance based on this perspective. Some theories that pertain to in- dividual offending include the situational emphasis of the routine activity perspective, such as Gottfredson and Hirschi's (1990) general theory of crime and Miethe and Meier's (1990, 1994) structural-choice theory of victimization. In these theories, however, historical factors (in Sutherland's sense of the term) are prominent in explain- ing individuals' rates of deviant behavior.

We also extend the routine activity per- spective to a wider range of deviant behav- iors-behaviors that are disapproved by con- ventional normative standards and that typi- cally provoke attempts at social control if detected by authority figures. Theoretical

statements defining the perspective are ex- plicitly limited to predatory crime, meaning incidents in which an offender does harm to or takes property from a victim (Cohen and Felson 1979; Hindelang et al. 1978; Miethe and Meier 1994). This sharp distinction be- tween offender and victim is not applicable to a large share of illegal or deviant behav- ior, such as the use of illicit drugs, reckless behavior, illegal services, and mutual vio- lence erupting from disputes. Even so, the relevance of routine activities to a wide range of deviant behaviors is illustrated in Felson's writings (Felscm 1.986, t994; Felson and Gottfredson 1984).

Building on his work and on several con- cepts from delinquency theory, we develop the rudiments of a routine activity theory of general deviance. We investigate these themes empirically through a study of the re- lationships between several types of deviant behaviors and a variety of routine activities. In accord with the routine activity perspec- tives' emphasis on connecting social struc- ture to crime, we also assess the degree to which routine activities can account for the relationship of deviance to some important dimensions of social differentiation.

PRIOR RESEARCH ON DEVIANCE AND ACTIVITIES

Several researchers have investigated the re- lationship between deviant behavior and the way that people spend their time, although only Riley (1987) has applied the routine ac- tivity perspective to the question. One poten- tial connection between routine activities and deviant behavior is captured by the old say- ing "Idle hands are the devil's workshop." This idea appears in Hirschi's (1969) social control theory as the bond of involvement: "The assumption, widely shared, is that a person may be simply too busy doing con- ventional things to find time to engage in de- viant behavior" (p. 22). Thus, it follows that the amount of time spent in virtually any nondeviant activity should be negatively as- sociated with rates of deviant behavior.

There are findings consistent with this pre- diction, such as Hundleby's (1987) results concerning home-centered activities, and Agnew and Peterson's (1989) findings on passive leisure and organized activities. Yet

This content downloaded from ��������������66.77.17.54 on Sun, 11 Apr 2021 17:31:19 UTC��������������

All use subject to https://about.jstor.org/terms

ROUTINE ACTIVITIES AND DEVIANT BEHAVIOR 637

these relationships are weak, and they are complemented by findings of weak positive relationships between other conventional ac- tivities and deviant behavior. Hirschi (1969: 190) found that adolescents who more fre- quently watch television, read comics, and play games exhibit higher levels of delin- quency, and Hundleby (1987) found that out- door activities (e.g., boating and camping) and athletic activities are positively associ- ated with substance use, sexual behavior, and delinquency. There is broad support for Hirschi's (1969:190) conclusion that delin- quent behavior simply is not sufficiently time consuming to support the logic underlying the bond of involvement. These weak find- ings are not surprising when viewed from the routine activity perspective, which focuses on activities that provide opportunities for deviant behavior.

In a different vein, it has been argued that certain activities are related to deviant be- havior because they are part of a deviant subculture (Agnew and Peterson 1989) or a deviant lifestyle (Jensen and Brownfield 1986). If hanging out in pool halls is popu- lar among delinquents, then people who do so "will be exposed to individuals who en- courage or provide opportunities for delin- quency, and/or foster values that approve of or are at least conducive to delinquency" (Agnew and Peterson 1989:334). Agnew and Peterson (1989:336) cite several studies establishing that spending time in such ac- tivities is positively associated with adoles- cent deviance.

Though such findings may be of descrip- tive value, their relevance to theoretical ex- planation is limited by the theoretical inde- terminacy discussed by Meier and Miethe (1993:484-87). The preceding quote from Agnew and Peterson suggests a connection with the routine activity perspective because activities characteristic of a delinquent sub- culture can provide opportunities for deviant behavior. At the same time, however, these activities are equally germane to theories that portray deviant behavior as arising through a process of social influence, such as differen- tial association theory (Sutherland 1947). The causal indeterminacy is compounded by the possibility of selection-one can choose activities that carry a reputation for deviance, and selecting such activities may simply re-

flect that one is already inclined toward de- viant behavior, while choosing conventional activities may indicate the opposite. For in- stance, Hirschi (1969:191) concluded that time spent on homework was negatively re- lated to delinquency because it indicated in- vestment in conventional goals. In addition to these theoretical pitfalls, focusing on ac- tivities that carry connotations of deviance or virtue is contrary to the broader aims and spirit of the routine activity perspective, which explicitly eschews explanation in terms of values and normative standards (Birkbeck and LaFree 1993; Felson 1994; Meier and Miethe 1993).

Research on the relationship between rou- tine activities and individual deviance reveals a set of activities consistent with the routine activity perspective that is not as subject to alternative theoretical interpretations. Sev- eral studies suggest that individual offending is positively associated with time spent in unstructured socializing with peers in the ab- sence of authority figures. Rates of delin- quency are higher among adolescents who spend more time (1) talking with friends or riding in a car (Hirschi 1969:194-95), (2) in social activities, "hanging out," or with their peers (Agnew and Peterson 1989), and (3) away from home or with groups of friends (Riley 1987). Wallace and Bachman (1991) found that, among a large set of demographic and attitudinal measures, the frequency of spending evenings out for fun and recreation was the strongest predictor of substance use. The most comprehensive investigation of routine activities and deviance is by Hun- dleby (1987), who assessed the relationships of sexual behavior, several types of substance use, and delinquency to a wide variety of adolescent activities. Among these activities the only consistently strong correlate of de- viance was an index of informal socializing with friends.

Time spent in informal, unsupervised so- cializing with peers carries no direct conno- tation of deviance. Virtually everyone spends some time this way, and people can as easily use this time for conventionally valued pur- suits as for proscribed ones. Furthermore, this classification of activities is sufficiently general to be applicable across time and across social groups, providing the possibil- ity for assessing whether social change or

This content downloaded from ��������������66.77.17.54 on Sun, 11 Apr 2021 17:31:19 UTC��������������

All use subject to https://about.jstor.org/terms

638 AMERICAN SOCIOLOGICAL REVIEW

cross-cultural differences in time spent this way translates to differences in rates of devi- ance. Indeed, there is cross-cultural evidence of this sort. From their analysis of the Hu- man Relations Area Files for 50 cultures, Schlegel and Barry (1991:135-39) con- cluded that problems of adolescent antisocial behavior are more likely in cultures in which adolescents spend less time in the company of adults and more time in the company of peers. Interestingly, independent socializing with peers typically occurred through partici- pation in culturally valued religious or mili- tary activities.

The essence of the routine activity per- spective is that crime is dependent on oppor- tunity. The literature contains passing re- marks to the effect that opportunities for de- viance might be especially prevalent during informal, unsupervised socializing with peers (Agnew and Peterson 1989:334; Hun- dleby 1987:108), but the subject has not been developed theoretically. We now turn our at- tention to that task.

APPLYING THE ROUTINE ACTIVITY PERSPECTIVE TO INDIVIDUAL DEVIANT BEHAVIOR

Situational Motivation

Although Cohen and Felson (1979) specify the "motivated offender" as one of three nec- essary elements of predatory crime, they give little attention to the nature of this motiva- tion, noting that theories of crime and delin- quency offer many possibilities. We believe that a routine activity analysis of individual deviance is best built upon a conception of motivation in which situational factors are prominent. Fortunately, such motivational concepts can be found in theories of devi- ance, despite their overwhelming emphasis on historical rather than situational explana- tions.

A central concept for our analysis is Briar and Piliavin's (1965) idea of situational mo- tivation, which states that the motivation for delinquency is inherent in the situation rather than in the person.

[R]ather than considering delinquent acts as solely the product of long term motives deriv- ing from conflicts or frustrations whose gen- esis is far removed from the arenas in which

the illegal behavior occurs, we assume these acts are prompted by short-term situationally induced desires experienced by all boys. ... (P. 36)

Their conception meshes well with Mat- za's (1964) claim that delinquency arises from "drift"-a state of openness to deviant values but not a rejection of conventional values. Similarly, Gold's (1970:92-99) anal- ogy of delinquency to a "pickup game" of basketball or baseball emphasizes that devi- ance typically is casual and spontaneous. To participate, one needs "to be there when the opportunity arises and when others are will- ing" (p. 94). Yet the pickup game analogy is ambiguous about whether finding opportuni- ties for deviance stems from prior motiva- tion. One player comes to the court looking for a game, so her motivation is internal rather than situational. Another player joins the game only because a friend calls out as he passes by. This second image better matches the idea of situational motivation, in which the potential for deviance arises in the course of other pursuits.

Gottfredson and Hirschi (1990) include a situational conception of motivation in their general theory of crime (which they define to encompass a broad range of deviant be- haviors). "[O]ur theory suggests that the mo- tive to crime is inherent in or limited to im- mediate gains provided by the act itself" (p. 256). A situational conception of motivation also meshes well with the cost versus benefit analysis found in the rational-choice per- spective, which shares the routine activity perspective's emphasis on the contribution of opportunity to crime.

For Briar and Piliavin (1965), the concept of situational motivation is the basis of a so- cial control explanation of delinquency. Rea- soning that everyone encounters situations in which delinquent behavior would be reward- ing, they portray variation in delinquency as dependent on the "stakes in conformity" (Toby 1957) that induce an actor to forego those benefits. Similarly, Gottfredson and Hirschi's (1990) general theory of crime em- phasizes self control, which is an individual's capacity to resist temptations.

We depart from these theorists to focus on another implication of the concept of situ- ational motivation. If deviance arises from

This content downloaded from ��������������66.77.17.54 on Sun, 11 Apr 2021 17:31:19 UTC��������������

All use subject to https://about.jstor.org/terms

ROUTINE ACTIVITIES AND DEVIANT BEHAVIOR 639

conducive situations, then individuals who spend more time in these situations should have higher deviance rates. The routine ac- tivity perspective points to the ordinary ac- tivities of everyday life as a source of varia- tion in levels of exposure to such situations.

We do not assume that everyone is equally

receptive to the temptations of situations conducive to deviance, but neither do we as- sume that exposure to them is relevant only to a small group of "motivated offenders." "Subterranean values" supportive of devi-

ance (excitement, conspicuous consumption, and toughness) are part of the general cul- ture (Matza and Sykes 1961), so one need not reject conventional values in order to en- gage in deviance. Even youths with a history of delinquent behavior feel compelled to jus- tify their acts (Sykes and Matza 1957), rate conventional behavior positively and deviant behavior negatively (Short and Strodtbeck 1965), and erroneously claim that their friends engage in less deviant behavior than most other people (Gold 1970:96-97). Therefore, we reject a categorical distinction between offenders and nonoffenders. Instead, we assume that people vary widely in their susceptibility to deviance, that this variation is continuous and not discrete (Rowe, Os- good, and Nicewander 1990), and that most people have the potential for at least occa- sionally succumbing to an opportunity for deviant behavior.

We replace Cohen and Felson's (1979) "motivated offender" with an assumption that the motivation resides in the deviant be- havior itself. Their second element, the "suit- able target," provides a situational motivation appropriate to the domain of their analysis, namely, direct contact predatory crime (p. 589). To apply the routine activity perspec- tive to a broader range of deviant behavior, we substitute the more general notion of situ- ations in which a deviant act is possible and rewarding. Following Briar and Piliavin's (1965:38) ideas about variation among situa- tions and Gottfredson and Hirschi's (1990) portrayal of the inducements of crime, we state that the easier the deviant act and the greater the symbolic and tangible rewards, the greater the inducement to deviance.

The inducement to deviance of any specific situation in some respects depends on the de- viant act in question. For instance, income tax

fraud is not possible without earnings that are

subject to taxation, and it is made consider-

ably easier when a person has received earn- ings that were not reported to the government.

Rather than analyzing features idiosyncratic

to specific deviant behaviors, however, the present study is concerned with general

classes of situations that are relevant to many types of deviant behavior. No doubt addi-

tional situational contingencies apply to some deviant acts, such as being in stores for shop- lifting, being with a potential partner for pre- cocious sexuality, and being in a position of financial trust for embezzlement.

Time with Peers

Situations conducive to deviance are espe- cially prevalent in time spent with peers.

Gold's (1970) "pickup game" analogy em- phasizes the group nature of most deviance and fits with the abundant evidence that most

illegal behavior occurs in the company of others (Erickson and Jensen 1977). Research reviewed above reveals that individuals who

spend more time with friends engage in de- viant behavior more frequently.

Being with peers can increase the situ- ational potential for deviance by making de- viance easier and rewarding. Though deviant behavior is rarely difficult or complex (Gott- fredson and Hirschi 1990), companions can serve as useful resources. Friends are a com-

mon source of illicit drugs; being accompa-

nied by friends reduces the danger in chal- lenging a rival to a fight; and having a part- ner to serve as look-out can enhance the chances of success at theft.

The companionship of friends is even more central to the symbolic rewards of enhanced status and reputation. Deviant exploits bol- ster a social identity as brave, adventure- some, or tough only when they come to the attention of others. The presence of friends may not be required to garner status, but it enhances credibility. In this vein, Gold's (1970:98) "pickup game" analogy empha- sizes that deviance is often a performance, for which the peer group provides an appre- ciative audience.

This is not to say that the presence of peers is a necessary condition for deviant behavior (Erickson and Jensen 1977; Gold 1970:98). We simply claim that, other things being

This content downloaded from ��������������66.77.17.54 on Sun, 11 Apr 2021 17:31:19 UTC��������������

All use subject to https://about.jstor.org/terms

640 AMERICAN SOCIOLOGICAL REVIEW

equal, spending more time with peers ex- poses an individual to more situational in- ducements to deviance, and this leads to higher rates of deviance.

The Absence of Authority Figures

In Cohen and Felson's (1979) routine activ- ity theory, the last element necessary for predatory crime is the absence of capable

guardians. By guardianship, they do not mean special skills or security arrangements; they conclude, for example, that the simple presence of a person in a house greatly re- duces the chance of a burglary.

The term "guardian" is apt for predatory crime, which involves a target, but it is less suitable for other forms of deviance. Felson (1986) addressed this issue by adding a fourth element to the earlier formulation: the absence of a "handler," a person capable of exercising social control over the potential offender. The handler role differs from the guardian role in that it concerns a relation- ship to the potential offender rather than to a valuable object or potential victim. Felson used his addition to integrate the routine ac- tivity perspective with Hirschi's (1969) so- cial control theory, reasoning that individu- als with strong social bonds will be more easily "handled."

We prefer to develop a strictly situational explanation of individual deviance that does not invoke individual characteristics, such as social bonds (e.g., relationships with conven- tional individuals and institutions). Gibbs's (1981) conceptual analysis of social control is useful in this regard. He defined social control as the use of social means to manipu- late the behavior of others (p. 78). Because the functions of "guardian" and "handler" re- flect the impact of the presence of others on the likelihood of deviance, they constitute social control in Gibbs's sense. In accor- dance with the routine activity perspective, these roles characterize the situation, not the potential offender. Specifying that the pres- ence of others serves a social control func- tion need not presume that the potential of- fender has strong social bonds. Indeed, Gibbs (1981:146-47) holds that theories of social bonding (e.g., Hirschi's social control theory) do not concern social control as he defines it.

Generalizing the handler and guardian roles, we state that a situation is more con- ducive to deviance if no authority figure is present. By authority figure, we mean someone whose role in a situation carries a responsibility for attempting to exert social control in response to deviance. Though people without this role obligation (e.g., peers and passersby) may attempt such so- cial control, they are less likely to do so. The authority figure's obligation to inter- vene may stem from a role in the setting, as in sales clerks who would be expected to take action when they observe shoplifting or to intervene in a fight on the premises. This corresponds to the "place manager" in Eck's (forthcoming) analysis of crime and places. Relationships with the potential of- fender, such as parent, teacher, or supervi- sor, may also bring obligations to exert so- cial control. Note that the social control function resides in the authority figure's role obligations, not in the actor's bonds to the authority figure. Whether you like or dislike your father, it will be more conve- nient to smoke marijuana when he isn't around.

In industrial society, role relationships sub- ordinate to authority figures are ubiquitous in the settings of work, school, and family of origin. This implies that situations conducive to deviance are most prevalent during leisure activities away from senior family members. Accordingly, prior research shows that ac- tivities most highly associated with deviance reflect either peer-centered leisure activities or activities that take place away from home. Felson has treated the balance of activities in the company of parents versus friends in ado- lescents' lives as especially pertinent to the routine activity perspective (Felson 1994; Felson and Gottfredson 1984).

Structured Versus Unstructured Activities

Unstructured activities that carry no agenda for how time is to be spent should be more conducive to deviance for two reasons. First, activities that are organized are likely to place some individuals in roles that make them responsible for social control. For ex- ample, athletic contests usually involve coaches, organized clubs have officers, and at restaurants and theaters employees are

This content downloaded from ��������������66.77.17.54 on Sun, 11 Apr 2021 17:31:19 UTC��������������

All use subject to https://about.jstor.org/terms

ROUTINE ACTIVITIES AND DEVIANT BEHAVIOR 641

charged with maintaining order. Second, structured activities offerfewer opportunities for deviance. Obviously, a person cannot en- gage in a deviant act without at least one op- portunity to do so. Yet, the minimal opportu- nities needed to make deviant activities pos- sible are so widespread that this sort of ab- solute exposure to opportunity is of little em- pirical or theoretical interest. As Gottfredson (1981) made clear, the routine activity per- spective directs our attention to relative ex- posure to opportunities, as reflected in how much time a person spends in situations con- ducive to deviance. The amount of structure in an activity is relevant here because greater structure means that more time will be spent in designated ways, and this time will not be available for deviance.

This is not to say, however, that spending time in structured activities reduces deviance (as in Hirschi's [1969] concept of involve- ment). As Felson (1994:108) notes, partici- pation in organized activities may as easily increase as decrease time spent in other ac- tivities that are conducive to deviance. Time in organized activities could take away from low-risk pursuits, such as watching television or doing household chores. Organized leisure activities, such as participating in clubs or sports, can potentially provide resources that enable deviance (Agnew 1990), indirectly leading to more extensive unstructured, un- supervised socializing by expanding friend- ship networks, taking one farther from home at later hours, and so forth. The previously mentioned inconsistent empirical support for the bond of involvement is understandable in this light.

DESCRIPTION OF THE PRESENT STUDY

This study tests our version of the routine

activity perspective by investigating the lon- gitudinal relationship between routine activi-

ties and individual offending. The cross-sec- tional designs of previous studies leave open the possibility that the observed relationship between routine activities and deviance is spuriously generated by other factors related to both, such as sensation-seeking, school failure, or attachment to parents. We use a longitudinal design to control for such stable individual differences.

We attempt to distinguish which routine activities are most related to deviant behav- ior. Previous studies either have investigated only a few activities (Hirschi 1969; Wallace and Bachman 1991) or have constructed indi- ces that combine unstructured, unsupervised socializing with less relevant activities, such as going to a school dance (where one finds authority figures; Hundleby 1987), or talking on the telephone (which typically occurs at home; Agnew and Peterson 1989).

Our national sample of 18- to 26-year-olds picks up where the available research on jun- ior and senior high school students leaves off. Data on this older age span is valuable for establishing that past evidence about the relevance of routine activities is not merely a by-product of adolescents' precocious in- volvement in adult activities or of the domi- nance of school and family of origin in ado- lescents' lives.

We also address the connection between the social structure and individual offending. The routine activity perspective is a theory of the social embeddedness of crime and deviance. In extending the perspective to individual of- fending, we examine the potential role of rou- tine activities as a mediator between struc- tural variables and deviance. Our analysis in- cludes three primary dimensions of social dif- ferentiation: age, sex, and social status.

We are especially interested in the poten- tial of routine activities for explaining the relationship of age to deviant behavior. Hirschi and Gottfredson (1983) documented that crime rates vary greatly with age, and similar age trends have been observed for various types of substance use (Johnston, O'Malley, and Bachman 1992). Hirschi and Gottfredson (1983) argue that available theories of crime and deviance are not able to explain these age trends. Indeed, although delinquency and illicit drug use during ado- lescence have been the major focus of theo- ries of deviance, these theories have offered little insight into how changes experienced during and after adolescence could produce the age trends that have been observed. The routine activity perspective directs our at- tention to age-related changes in the activi- ties of everyday life. Indeed, there is evi- dence of striking age trends in many such activities (Larson and Bradney 1988; Larson and Richards 1991; Osgood and Lee 1993).

This content downloaded from ��������������66.77.17.54 on Sun, 11 Apr 2021 17:31:19 UTC��������������

All use subject to https://about.jstor.org/terms

642 AMERICAN SOCIOLOGICAL REVIEW

METHOD

Sample

Our data come from the Monitoring the Fu- ture study. This ongoing study began in 1975 and gathers a wide range of information an- nually from a nationally representative sample of high school seniors. Each year, a three-stage national probability sample is drawn, and questionnaires are administered in approximately 130 high schools (roughly 110 public and 20 private). This procedure yields 15,000 to 19,000 respondents annu- ally. A random one-fifth of each sample com- pletes the version of the questionnaire that we use here. For a detailed description of the sample design and data collection, see Bach- man, O'Malley, and Johnston (1991).

We use the follow-up portion of the study, which includes a subsample of one-fifth of each senior class sample.2 Half the partici- pants in the follow-up study complete mailed questionnaires in every odd-numbered year after graduation; the other half do so every even-numbered year. Response rates for the base year average 80 percent, and follow-up response rates are generally 75 percent or more of the original group. The follow-up study oversamples the more serious drug us- ers in high school to obtain more accurate estimates for this segment of the population; the oversampled individuals are then given smaller weights in analyses to yield a repre- sentative sample.

The analysis is based on high school se- nior classes of 1977 through 1981. Five waves of data are used, obtained at the ap- proximate ages of 18, 19, 21, 23, and 25 or 18, 20, 22, 24, and 26. We included only those cohorts that had progressed through at least four of the five data waves, and respon- dents were included in the analyses only if they had valid data for at least three of the five questionnaires. Sample sizes ranged from 1,782 to 1,840 across the five depen- dent variables.

Because Monitoring the Future does not sample individuals who leave high school before spring of their senior year, our find- ings are generalizable only to high school

graduates, a group that represents about 80 percent of the all the age cohorts. Although dropouts tend to have higher rates of deviant behaviors, such as drug use (SAMSHA 1993) and delinquency (Fagan and Pabon 1990), their relatively small proportion of the population reduces the potential for bias in

our parameter estimates. Moreover, bias will occur only if relationships among these vari- ables are different for dropouts than for graduates, and we have no reason to suspect this is so. Although a broader sample would

be desirable, the Monitoring the Future data- set remains one of the best available for studying deviance during late adolescence and early adulthood.

Measures

Routine activities. To minimize the problem of theoretical indeterminacy articulated by Meier and Miethe (1993), we restricted our

analysis to activities that would be least sub- ject to alternative interpretations under other theoretical perspectives. Thus, we eliminated items about time spent at work, in school, and in religious activities as reflecting com- mitment to conventional lines of action, and we excluded an item about frequenting tav- erns, bars, or nightclubs as being too closely associated with alcohol consumption, which is illegal for a portion of this age span. Ap- pendix A presents the questions and response categories for the 13 activities that were in- cluded in the analysis.3

Four of the 13 activities we include typi- cally entail unstructured socializing with peers in the absence of responsible authority figures, as is specified by our theoretical analysis: riding around in a car for fun, get- ting together with friends informally, going to parties, and spending evenings out for fun and recreation. Five of the remaining activi- ties typically occur outside the home: going on dates, going to movies, participating in

2 The raw data from the longitudinal study are not publicly available, but researchers may con- tact the authors to obtain covariance matrices.

3 To simplify the analyses, three items-re- flecting playing music or singing, doing creative writing, and doing arts or crafts-were elimi- nated. The eliminated items are included in Ap- pendix A. Preliminary analyses indicated that these relatively infrequent activities played no important role in the results. Preliminary results are available from the first author, as are the re- sults of all other analyses to which we refer.

This content downloaded from ��������������66.77.17.54 on Sun, 11 Apr 2021 17:31:19 UTC��������������

All use subject to https://about.jstor.org/terms

ROUTINE ACTIVITIES AND DEVIANT BEHAVIOR 643

community affairs or volunteer work, partici- pating in active sports, and going shopping. In this age range, we would expect that most of these activities occur away from parents, and dating and going to movies imply social- izing with others. Unlike the first four activi- ties, however, these activities are more struc- tured-they entail a somewhat definite and partly constrained agenda. The remaining four activities are more likely to occur in the home and are less likely to involve compan- ionship: working around the house, watching television, relaxing alone for an hour or more, and reading.

Other explanatory variables. Our analysis included four variables reflecting structural differentiation. The first of these is age. Be- cause all respondents were high school se- niors at the start of the study, timing of the waves of data collection is the principle source of variation in age. Therefore, the base year wave of data was defined as age 18, and age for each subsequent wave was defined as 18 plus the number of years since the base year. We allowed for a curvilinear relation- ship between age and deviance by including both age and age squared in our models.

The initial questionnaire assessed the re- maining structural variables. Sex was coded as 0 for males and 1 for females. The only available indicator of socioeconomic status for the family of origin was parents' average education, on a scale of 1 (grade school or less) to 6 (graduate or professional school). Respondents were asked to answer the ques- tions about education with regard to what- ever parent figures were "most important in raising" them. More pertinent to respon- dents' future socioeconomic prospects, we also included respondents' self-reported av- erage high school grades (coded 1 for D or below, 2 for C-, 3 for C, 4 for C+, and so on through 9 for A).

Control variables. Traffic tickets and acci- dents, our measures of dangerous driving, are necessarily a function of both individual driving practices and time at risk (i.e., time spent driving). Therefore, in analyses of dan- gerous driving we controlled for a measure of how far respondents drive in an average week, ranging from 1, for not at all, to 6 for more than 200 miles.

A limited portion of our analysis concerns stable individual differences in deviance, and

some of these models include a number of additional control variables: urbanicity (I to 5 scale for population size of area of resi- dence), plans to attend a 4-year college (1 to

4 scale), two dummy variables for race (Af- rican American and White versus other),

three for region of the country (Northeast, North Central, and West versus South), five for year of high school senior class, five in- dicating whether or not a respondent had valid data for each data wave, and four indi- cating whether a respondent had missing data for sex, parent's education, high school grades, and college plans. Mean values were substituted for missing values on those vari- ables, and including these dummy variables allowed for the possibility that individuals with missing values systematically differed in their deviant behavior.

Deviant behaviors. Our analysis includes self-report measures of five types of deviant behavior: criminal behavior, heavy alcohol use, marijuana use, use of other illicit drugs, and dangerous driving. These behaviors rep- resent a broad range of conventionally pro- scribed activities that are common in late adolescence. Factor analytic studies of sub- stance abuse have shown that use of alcohol, marijuana, and hard drugs are relatively dis- tinct phenomena (Hays et al. 1986), so we consider them separately.

Our 10-item measure of criminal behavior was used in the Youth in Transition study (Bachman, O'Malley, and Johnston 1978); it is adapted from Gold's (1970) well-known measure. Three items concern violent of- fenses, such as serious fights and robbery, while the remaining 7 items concern property offenses of theft, trespassing, and arson. We excluded 4 items pertaining to offenses at work or school because they are tied to age- specific role statuses. Responses to each item ranged from 0, for not engaging in the behav- ior at all during the past year, to 4, for com- mitting the offense five or more times during that period. The index for criminal behavior was the sum of scores across the 10 items.

We measured heavy alcohol use by the number of occasions in the preceding two weeks a respondent had 5 or more drinks in a row. Scores ranged from 0 through 5 (10 or more times). The scale for marijuana use ranged from 0 (for no use in the past 12 months) through 9 (40 or more times in the

This content downloaded from ��������������66.77.17.54 on Sun, 11 Apr 2021 17:31:19 UTC��������������

All use subject to https://about.jstor.org/terms

644 AMERICAN SOCIOLOGICAL REVIEW

last 30 days). Use of other illicit drugs was measured as a sum across eight drugs, each one scored on the same scale as marijuana use: LSD, other psychedelic drugs, cocaine, quaaludes, barbiturates, tranquilizers, heroin, and other narcotics. Our measure of danger- ous driving was the sum of traffic tickets and traffic accidents reported by the respondent for the past 12 months. Possible scores for each of these were 0 through 4 (4 or more).

Osgood et al. (1988) obtained estimates of the reliability of these measures as part of a longitudinal causal model based on Heise's (1969) approach. They reported reliabilities of .70 for criminal behavior, .70 for heavy alcohol use, .90 for marijuana use, .76 for use of other illicit drugs, and .49 for danger- ous driving. Alpha reliabilities for the mul- tiple-item measures were .78 for criminal be- havior, .81 for use of illicit drugs other than marijuana, and .48 for dangerous driving.

RESULTS

The first phase of our analysis tests our hy- pothesis that activities involving unsuper- vised and unstructured socializing with peers will be closely associated with deviance; the second assesses the role of routine activities as a mediator between structural variables and deviance.

Analysis Strategy

A primary goal of our analytic strategy is to focus on the utility of routine activities for explaining within-individual change in devi- ant behavior. By using a "fixed-effects" panel model, which limits the analysis to within-individual change, we ensure that our findings cannot be due to any stable indi- vidual differences, whether measured or not (Petersen 1993:447). Thus, we capitalize on the strengths of our longitudinal data to con- trol for selection factors, which has not been possible for previous studies of activities and deviance. Although we still cannot rule out the possibility that results are due to chang- ing but unmeasured variables or to time- varying effects of stable variables, our ap- proach is a more stringent test than previous cross-sectional analyses.

Our primary analysis takes the form, of a pooled time series and cross-sections design,

meaning that each wave of data for each in-

dividual is treated as a separate case. This is

an appropriate strategy for assessing rela-

tionships in panel data when changes of mean levels over time are of interest (Hsiao 1986; Petersen 1993). Because the multiple

observations for a single individual are not statistically independent, it would be inap- propriate to apply ordinary least squares to

data in this form. Though unbiased, estimates of standard errors would be incorrect.

We used the "fixed-effects" estimator, which is one of the two common approaches

to correcting for violations of the assumption independence that are due to stable individual differences (Petersen 1993). This approach has the advantage of restricting the analysis to within-individual changes for both the in- dependent and dependent variables. The al- ternative approach, the "random-effects" es- timator, is somewhat more efficient, but a modest loss in statistical power was not prob- lematic with our sample size. Furthermore, unlike the fixed-effects estimator, the ran- dom-effects estimator assumes a normal dis- tribution of the stable individual differences (Petersen 1993:447-48), which is a poor match to our skewed dependent variables. We implemented the fixed-effects model by con- verting all variables to deviations from each individual's mean across time and conduct-

ing ordinary least squares regression on those within-individual deviations.4

This fixed-effects model does not correct

for serially correlated error, which is a sec- ond potential source of violations of the as- sumption of independence. Stimson (1985) concluded that serially correlated error is relatively unimportant in cases such as ours, where the sample is large and there are rela- tively few waves of data. Our results support his view in that the correlations between re- siduals, although higher for adjacent waves,

4 Data in this form are constrained to sum to 0 across waves, reflecting a loss of one degree of freedom per individual (used to calculate the in- dividual mean). A modified sample weight cor- rects the degrees of freedom:

/ ti -1 Wi = Wi t 1

1 t. where wi is the original sample weight for indi- vidual i (based on the over-sampling of more seri- ous drug users), w' is the modified sample weight,

and ti is the number of waves of valid data.

This content downloaded from ��������������66.77.17.54 on Sun, 11 Apr 2021 17:31:19 UTC��������������

All use subject to https://about.jstor.org/terms

ROUTINE ACTIVITIES AND DEVIANT BEHAVIOR 645

were still substantial over longer intervals. Furthermore, we obtained virtually identical results when we replicated our analyses us- ing hierarchical linear models (Bryk and Raudenbush 1992) that allowed for serial correlation through random effects in stable individual differences and age trends.

Examination of residuals from preliminary analyses indicated extremely skewed distri- butions for criminal behavior, use of illicit drugs other than marijuana, and dangerous driving. To improve these distributions, we transformed the data before subtracting the individual means. The natural logarithm was taken for criminal behavior and the use of il- licit drugs other than marijuana, and the square root of dangerous driving was taken (in all cases after adding 1 to the original score). Also, prior to making the transforma- tions, we recoded scores for criminal behav- ior to a maximum of 20 and illicit drug use to a maximum of 25. Less than 1 percent of scores fell above those levels, so we did not think it was meaningful to distinguish among them. Finally, the three transformed variables were multiplied by 10 to compensate for their reduced ranges.

Activities and Deviance

Table 1 presents coefficients from the regres- sions of change over time in five deviant be- haviors on change in routine activities. For each deviant behavior, routine activities ex- plained significant amounts of variance not accounted for by age: from 1.2 percent to 10.9 percent. In accord with prior research and our theoretical analysis, there were con- sistent positive associations between the four unstructured socializing activities and the five deviant behaviors: Riding in a car for fun, visiting with friends, going to parties, and spending evenings out coincided with crimi- nal behavior, heavy alcohol use, marijuana use, use of other illicit drugs, and dangerous driving. These activities typically accounted for the largest share of the variance explained by the set of 13 activities. All but one of the twenty relevant coefficients were positive, and each of the four unstructured socializing activities was significantly associated with at least three of the five deviant behaviors.

Results for the other five activities typi- cally occurring outside the home were in

marked contrast to findings for unstructured

socializing: Deviant behavior was not posi-

tively associated with going on dates, going to movies, being involved in community af-

fairs, engaging in active sports, or going shopping. Thus, it is not merely spending time outside the home or socializing that

leads to deviant behavior. The only nomi- nally significant positive association between

these routine activities and deviance was the

relationship between dangerous driving and going to movies, and this relationship would not be judged statistically significant under a Bonferroni correction for testing the associa- tion with five deviant behaviors (i.e., alter-

ing the alpha level to .01).

It is particularly interesting that once we controlled for other activities there was little

indication that going on dates or going to movies leads to-deviant behavior. These forms of socializing take place out of the home, and they had been included in com-

posite measures of socializing that were as- sociated with deviance in prior research (Agnew and Peterson 1989; Hirschi 1969:

168; Hundleby 1987). We did find positive zero-order correlations between these two

activities and most of the deviant behaviors (results not shown), but those correlations appear to result from higher rates of dating and movie attendance among individuals who more frequently engage in the unstruc- tured activities.

Conversely, controlling for dating and go- ing to movies helps clarify the meaning of some of the other socializing activities. Go- ing to parties and evenings out for fun are broad categories that may encompass any- thing from a formal dinner party to hanging out on a street corner. Because our regression models include the entire set of 13 activities, however, the coefficients for parties and eve- nings out are adjusted for rates of dating, go- ing to movies, and the other more structured activities. This gives us more confidence that findings for the first four unstructured social- izing activities are largely limited to infor- mal, unstructured and unsupervised socializ- ing with peers, in accord with our theoretical analysis.

For the nine activities other than unstruc-

tured socializing, the statistical significance of several of the coefficients surpasses the p < .05 level, but these results must be inter-

This content downloaded from ��������������66.77.17.54 on Sun, 11 Apr 2021 17:31:19 UTC��������������

All use subject to https://about.jstor.org/terms

646 AMERICAN SOCIOLOGICAL REVIEW

Table 1. Unstandardized Regression Coefficients and Variance Explained from Within-Individual Regressions of Five Deviant Behaviors on Routine Activities: Ages 18 to 26, Monitoring the Future Study

Criminal Heavy Marijuana Other Drug Dangerous Behavior Alcohol Use Use Use Driving

Routine Activity b (S.E.) b (S.E.) b (S.E.) b (S.E.) b (S.E.)

Unstructured Socializing

Ride for fun .359* (.084) .033* (.014) .113* (.022) .105 (.068) .176* (.047)

Visit with friends .177 (.114) .083* (.019) .068* (.030) .289* (.092) -.002 (.063)

Go to parties .927* (.122) .329* (.020) .350* (.032) .438* (.098) .117 (.068)

Evenings out .195 * (.077) .138* (.013) .117* (.021) .355* (.062) .121 * (.043)

Other Activities Outside the Home

Go on dates .015 (.067) -.030* (.011) -.028 (.018) -.091 (.054) .063 (.037)

Go to movies .068 (.141) -.037 (.023) -.025 (.038) -.085 (.114) .175* (.078)

Community affairs -.127 (.104) -.009 (.017) -.082*(.028) -.252*(.084) -.048 (.058)

Active sports .100 (.085) -.024 (.014) -.024 (.023) .038 (.068) .038 (.047)

Go Shopping -.184 (.118) -.015 (.020) -.053 (.032) -.137 (.096) -.001 (.066)

At-Home Activities

Work around house -.162 (.095) -.045 * (.016) -.077* (.025) -.078 (.076) -.124* (.053)

Watch TV -.136 (.136) -.003 (.022) -.097*(.036) -.212 (.109) .015 (.075)

Relax alone .077 (.085) .030* (.014) .069* (.023) .201* (.068) .043 (.047)

Read book or magazine -.053 (.116) -.023 (.019) -.040 (.031) -.092 (.093) -.026 (.064)

Unique Variance Explained R2 d.f. R2 d.f. R2 d.f. R2 d.f. R2 d.f.

All activities .0225 13 .1089* 13 .0545* 13 .0235* 13 .0121* 13

Unstructured activities .0188* 4 .1012* 4 .0457* 4 .0184* 4 .0060* 4

Age .0461* 2 .0047* 2 .0067* 2 .0129* 2 .0051* 2

Total .1505* 15 .1194* 15 .0825* 15 .0297* 15 .0550* 16

S.D. . 5.334 .837 1.341 3.990 2.709 N (weighted) 5,986 5,712 5,817 5,930 5,715

Note: The bs are unstandardized regression weights; their standard errors are in parentheses.

* p < .05

preted cautiously. None would be judged sta- tistically significant under a Bonferroni cor- rection for the 45 significance tests involved (i.e., nominal alpha level of .0011 and t-value of 3.6). Of the 20 coefficients for unstruc- tured socializing, 11 would remain statisti- cally significant under this criterion.

One of the more consistent trends among these nine routine activities was that spend- ing more time relaxing alone was associated with higher levels of deviance. This is inter- esting in that relaxing alone would consti- tute unstructured solitary activity, rather

than socializing activity. Conversely, partici- pation in community affairs and working

around the house both were consistently as- sociated with lower rates of deviance. Thus, these three activities seem to merit attention in future research.

Routine Activities, Social Structure, and Deviant Behavior

Much of the reason for sociological interest in routine activities is to explore a possible link between broad social structural catego- ries and important social outcomes by exam- ining the content of everyday life. We now examine (1) whether location in the social structure shapes people's lives in terms of

This content downloaded from ��������������66.77.17.54 on Sun, 11 Apr 2021 17:31:19 UTC��������������

All use subject to https://about.jstor.org/terms

ROUTINE ACTIVITIES AND DEVIANT BEHAVIOR 647

these routine activities, and (2) the degree to

which these activities account for relation- ships between dimensions of social stratifi-

cation and deviant behavior.

We focus on four variables relevant to so- cial structural differentiation: age, sex, high

school grades, and parents' education. Age and sex are the two dimensions of social stratification most related to a broad set of measures of criminal behavior (Jensen and

Rojek 1992). High school grades and par- ents' education reflect two important aspects of social class in the transitional world of 18- to 26-year-olds. Parents' education is an in- dicator of the social class of the family of

origin, and high school grades are an indica- tor of adolescents' trajectories as they seek their own position in the social hierarchy. Neither race, region, urbanicity, nor college plans was strongly or consistently related to deviant behavior, after controlling for the other structural variables.

Age. We are especially interested in whe- ther age-related changes in routine activities can explain some portion of the age trends observed in deviant behaviors. This could occur only if there were similarly shaped age trends for routine activities and deviant be- havior-but that alone is not enough. As Hirschi and Gottfredson (1985) point out, there are many variables that show similar

age trends. The difficulty here lies in demon- strating that an explanatory variable has in- dividual level relationships to both age and deviance that are strong enough to account for the relationship between them.

Table 2 presents the results of within-indi- viduals (fixed-effects) analyses of age trends in the activity measures and shows signifi- cant age-related changes for all of the activi- ties except shopping. Computing fitted val- ues from these coefficients yields a pattern of declining frequency for almost all of the routine activities over this age span, with the fastest decline at age 18 and the slowest at age 26.5 There were large age trends for all

four of the unstructured socializing activities, which were closely associated with deviance; coefficients correspond to a decrease of more than one full standard deviation in each ac- tivity over this age span.

To assess the degree to which routine ac- tivities explain the relationship between age

and deviant behavior, we examined the change in that relationship with activities controlled. One normally accomplishes this by comparing a coefficient before controls (the total effect) to the same coefficient after controls (the direct effect); the difference be-

tween these coefficients indicates the extent of mediation (indirect effect). The present problem is somewhat more complex, how- ever, because the age trend is a quadratic re- lationship that involves two coefficients. Ex- plained variance (R2) is a poor substitute for the regression coefficient because it will be reduced by any correlated explanatory vari- ables, even when they do not decrease the di- rect effect. To resolve this, we created a co- efficient b that summarizes the magnitude of the age trend in the same metric as the usual

unstandardized regression coefficient. We defined b as the standard deviation across

age levels of the fitted values implied by the two age coefficients (holding other variables constant), divided by the standard deviation of those age values. (For a linear relationship this calculation yields the original regression coefficient.)

Table 3 summarizes the relationship of age to within-individual change in deviant behav- ior, based on fixed-effects models with and without controls for routine activities. Rou- tine activities explain a substantial portion of the age-related change in criminal behavior, heavy alcohol use, marijuana use, and dan- gerous driving; controlling for routine activi- ties reduced the age trends in these behav- iors by 27 percent to 48 percent. Age trends had been most dramatic for criminal behav- ior (see Table 1). Here, routine activities ex- plained a smaller proportion of the age trend, but a similar absolute amount of the trend.

For use of illicit drugs other than marijuana the results are rather different: Controlling for routine activities increased rather than decreased the age trend (thus, the negative proportion explained in Table 3).

Figure 1 (see page 650) illustrates these findings. It shows age trends for three of the

5 The terms for both linear age and age squared are based on a transformation of the age variable

to a value of 0 at age 22. Thus, the coefficients for linear age in Table 2 reflect the slope at age 22, and the pattern of coefficients observed in

Table 2 (negative for linear age and positive for age squared) is consistent with our description of

the age trends.

This content downloaded from ��������������66.77.17.54 on Sun, 11 Apr 2021 17:31:19 UTC��������������

All use subject to https://about.jstor.org/terms

648 AMERICAN SOCIOLOGICAL REVIEW

Table 2. Unstandardized Coefficients from Separate Regressions of Routine Activities on Age, Sex,

High School Grades, and Parents' Education: Ages 18 to 26, Monitoring the Future Study

Age Sex High School Parents'

Age (Age )2 (Female = 1) Grades Education

Routine Activity b b R2 b (S.E.) b (S.E.) b (S.E.)

Unstructured Socializing

Ride for fun -.146 .0236 .235* -.383* (.043) -.148* (.011) -.153* (.018)

Visit with friends -.107 .0024 .161 * -.193* (.027) -.004 (.007) .095* (.011)

Go to parties -.085 .0024 .124* -.183* (.030) -.012 (.008) .117* (.013)

Evenings out -.134 .0038 .117* -.411 * (.040) -.061 * (.011) .073* (.017)

Other Activities Outside the Home

Go on dates -.042 -.0124 .008* .007 (.047) -.001 (.013) .033 (.020)

Go to movies -.062 .0014 .097* .069* (.021) .018* (.006) .043* (.009)

Community affairs -.031 .0097 .035* -.015 (.031) .066* (.008) .079* (.013)

Active sports -.052 .0026 .030* -.405* (.043) .067* (.012) .188* (.018)

Go shopping -.002 .0008 .000 .435* (.023) .004 (.007) -.023* (.011)

At-Home Activities

Work around house .004 .0134 .013* .314* (.032) -.032* (.009) -.090* (.014)

Watch TV .029 .0092 .019* -.015 (.024) -.052* (.006) -.077* (.010)

Relax alone -.023 -.0081 .005* .005 (.033) .007 (.009) .060* (.014)

Read book or magazine .007 -.0029 .003* -.005 (.029) .057* (.008) .074* (.012)

Note: The bs are unstandardized regression weights (their standard errors are in parentheses).

*p < .05

deviant behaviors (as fit by the regression models), before and after adjusting for rou- tine activities. For four of the five deviant behaviors, a downward age trend was re-

duced or reversed by controlling for routine activities (in Figure 1, criminal behavior and heavy alcohol use show this effect). Unlike

the other deviant behaviors, use of other il- licit drugs varied little by age, showing only a slight peak at age 22. Illicit drug use was associated with unstructured socializing, however, which declined with age. Control- ling for routine activities increased this mod- est age trend because use of these drugs at later ages was higher than would be ex- pected, relative to the decline in unstructured socializing. This pattern would be consistent with a change in the context of use for these "harder" drugs, with use at older ages being

more solitary and persistent, rather than largely limited to social contexts, such as in- formal parties and hanging out with friends.

Sex. The analyses reported above are lim- ited to within-individual changes in deviant behaviors, using the fixed-effects correction for individual differences. This approach is

applicable only to independent variables that vary over time for at least some individuals (Petersen 1993:448), and our measures of sex, high school grades, and parents' educa-

tion do not vary over time. Thus, we con- ducted a separate analysis based on indi- vidual means over time on independent and

dependent variables. These means corre- spond to the between-individuals variance not used in the within-individuals, fixed-ef-

fects analysis. This separation of the analy- ses of within-individual variation and be- tween-individual variation is a standard fea- ture of analysis of variance, and it also ap- plies within a multiple regression framework (Judd and McClelland 1989, chap. 14-15).

Table 2 shows the relationships of sex, high school grades, and parents' education to the 13 routine activity measures, which are

assessed by bivariate between-individuals re- gressions (i.e., not controlling for any other variables). These results indicate that there

are substantial differences between the sexes in the routine activities of everyday life dur- ing late adolescence and early adulthood. Fe- males participated in the unstructured social-

This content downloaded from ��������������66.77.17.54 on Sun, 11 Apr 2021 17:31:19 UTC��������������

All use subject to https://about.jstor.org/terms

ROUTINE ACTIVITIES AND DEVIANT BEHAVIOR 649

Table 3. Within-Individual and Between-Individual Regression Models Showing the Impact on Se- lected Group Differences in Deviant Behavior of Controlling for Routine Activities: Ages 18

to 26, Monitoring the Future Study

Without Controls With Controls for Activities for Activities

Proportion

Deviant Behaviors b R2 b R2 Explained

Age Differences

Criminal behavior .690 .128* .505 .046* .268

Heavy alcohol use .041 .011* .023 .005* .444

Marijuana use .112 .028* .058 .007* .482

Other illicit drug use .107 .006* .184 .013* -.717

Dangerous driving .174 .023* .104 .005* .405

Proportion b (S.E.) b (S.E.) Explained

Sex Differences

Criminal behavior -3.300* (.242) -2.790* (.276) .155

Heavy alcohol use -.610* (.046) -.336* (.044) .449

Marijuana use -.519* (.109) -.007 (.114) .986

Other illicit drug use -.552* (.259) .360 (.284) 1.653

Dangerous driving -1.065* (.108) -.886* (.122) .168

High School Grades

Criminal behavior -.271 (.071) -.109 (.071) .598

Heavy alcohol use -.105* (.014) -.043* (.011) .595

Marijuana use -.215* (.032) -.087* (.030) .594

Other illicit drug use -.472* (.076) -.221* (.073) .531

Dangerous driving -.145* (.030) -.102* (.030) .298

Parents' Education

Criminal behavior .219 (.113) .058 (.111) .734

Heavy alcohol use .054* (.021) .007 (.018) .869

Marijuana use .146* (.051) .076 (.046) .477

Other illicit drug use .457* (.121) .319* (.115) .300

Dangerous driving .155* (.048) .114* (.047) .262

Note: The b's are unstandardized regression weights (their standard errors appear in parentheses); b is an approximation to b for a curvilinear relationship. R2 reflects variance solely attributable to age. Proportion explained refers to the proportionate reduction in b or b produced by controlling for the 13 routine activi- ties.

*p < .05

izing activities much less often than did males, as would be consistent with sex dif- ferences in deviant behavior. The other large sex differences in activities matched com- mon gender stereotypes: Males more fre- quently participated in active sports while fe- males more frequently went shopping and worked around the house.

Table 3 indicates the magnitude of sex dif- ferences in the five deviant behaviors, with and without controlling for routine activities.

Before controlling for activities, males en-

gaged in all of the deviant behaviors signifi- cantly more often than females. The differ- ence was largest for criminal behavior and heavy alcohol use (over half a standard de-

viation in both cases) and the smallest dif- ference was for use of drugs other than mari- juana (about one tenth of a standard devia-

tion). Controlling for routine activities re- duced all of the sex differences, but the de- gree of reduction varied considerably across

This content downloaded from ��������������66.77.17.54 on Sun, 11 Apr 2021 17:31:19 UTC��������������

All use subject to https://about.jstor.org/terms

650 AMERICAN SOCIOLOGICAL REVIEW

12

10

4

2

0 l l l l l l I 18 19 20 21 22 23 24 25 26

Age

1.4 -

1.2

10 0

1.0 92 2 2 2 2 5 2

.5-

~ 8 0 0

;R .6

@ .4

.2

18 19 20 21 22 23 24 25 26

Age

5

4

03

2

0

18 19 20 21 22 23 24 25 26

Age

- Females * Males

-K>--- Females, controlled Males, controlled

Figure 1. Fitted Relationships of Age and Sex to Deviant Behaviors, Before and After Controlling for Routine Activities: Ages 18 to 26, Monitoring the Future Study

Note: See the text on page 643 for a description of the scales measuring criminal behavior, heavy al- cohol use, and other drug use, and page 645 for transformations applied to the scales.

the behaviors. Routine activities accounted for virtually all of the sex differences in marijuana use and use of other illicit drugs.

About half of the sex difference in heavy al-

cohol use was attributable to routine activi- ties, but only 16 percent for criminal behav-

ior and 17 percent for dangerous driving.

Figure 1 gives a visual representation of the extent to which routine activities mediate the relationships of sex and age to the deviant behaviors. The distance between the lines for females and males is smaller after control- ling for routine activities, which illustrates

that routine activities account for much of the sex difference in deviance. The figure

shows that once we control for routine activi- ties, females' use of illicit drugs other than marijuana slightly exceeds that of males.6

High school grades. Table 2 also indicates that 18- to 26-year-olds who differed in their high school grades differed substantially in their everyday activities as well. Better high school grades were strongly associated with less frequently riding in a car for fun and spending evenings out, two of the four un- structured socializing activities. Respondents with better grades also had especially high rates for the more structured activities of

community affairs and active sports. Accord- ingly, controlling for routine activities con-

siderably reduced the relationship between high school grades and deviant behavior. As Table 3 reveals, rates for all five deviant be-

haviors were higher for respondents who had lower high school grades, and at least half of the relationship between grades and deviant behavior was explained by routine activities, except in the case of dangerous driving (30 percent).

These findings indicate that the lower rates of deviance among people who succeed at school is not simply a reflection of commit- ment to conventional avenues of success, as

6 We also examined sex differences in the rela- tionship of deviant behavior to age and to routine

activities. Though many of these interactions were statistically significant, the substantive re- lationships of both age and routine activities to deviant behavior were the same for both males and females. Typically, relationships were some- what stronger for males than for females, which is a common result of the combination of higher rates of deviance for males and skewed measures

of deviance.

This content downloaded from ��������������66.77.17.54 on Sun, 11 Apr 2021 17:31:19 UTC��������������

All use subject to https://about.jstor.org/terms

ROUTINE ACTIVITIES AND DEVIANT BEHAVIOR 651

is portrayed by social control theory (Hirschi 1969). Students who succeed at school spend their time in ways that present them with fewer opportunities for deviance, and this

explains much of the relationship. Parents' education. Because parents' edu-

cation is a major component of socioeco-

nomic status, one might expect this variable to be negatively associated with deviant be- havior. Yet research over the past two de-

cades shows that links between social class and deviance are elusive (Tittle, Villemez, and Smith 1978). Indeed, we found that for

these respondents, having parents with more education was associated with higher levels of deviant behavior (significantly so for all

except criminal behavior), and this was true both before and after controlling for factors such as sex and high school grades (see Table 3). Table 2 indicates that this finding is con- sistent with most (though not all) relation- ships between parents' education and respon- dents' routine activities.

Table 3 shows that controlling for routine

activities explains much of the observed re- lationship between parents' education and re- spondents' deviant behavior. This was most true for criminal behavior and heavy alcohol use, where routine activities accounted for 73 percent and 87 percent of the relationships, respectively. Routine activities explained 26

percent to 48 percent of the association be- tween deviance and parents' education for the remaining three deviant behaviors. Con- sistent with Agnew's (1990) notion of re- sources for deviance, it appears that higher

levels of parental social class offer youth greater freedom of movement and more time for socializing, which enable higher rates of deviant behavior.

An Alternative Model

We developed a structural equation model to address some limitations of our primary analysis. Although space permits only a brief discussion of this model, the results provide valuable corroboration for the fixed-effects

analysis we have presented. The purpose of the structural equation model was: (1) to cor- rect for error in the measures of routine ac- tivities, (2) to adjust for serially correlated error, (3) to treat the skewed measures of de- viance as ordinal rather than interval, and (4)

to maintain the separation of within-indi- vidual and between-individual relationships. Full details of this model are available from the first author on request.

The structural equation analysis confirmed that there were strong relationships between unstructured socializing activities and all five forms of deviance. Also, the within-indivi- dual and between-individual relationships be- tween activities and deviance were of compa- rable magnitude, which had not been true of the fixed-effects analysis. Further, the struc- tural equation results indicated that routine activities mediate much of the relationship between the structural variables and deviant behaviors. The structural equation and fixed- effects analyses yielded very similar esti- mates of the extent of mediation for sex, high school grades, and parents' education, but the two analyses diverged for age. Where routine activities accounted for 27 percent of the re- lationship between age and criminal behav- ior in the fixed-effects model, this figure rose to 88 percent in the structural equation model.

For the other four deviant behaviors, how- ever, the structural equation model "over-ex- plained" the relationship of age to deviance- after controlling for activities, the relation- ship changed to the opposite direction.

DISCUSSION

Our theoretical analysis extends the situ- ational explanation of crime found in the routine activities perspective to explaining individual offending and a broader range of deviant behaviors. Specifically, we have ar- gued that situations conducive to deviance are especially prevalent in unstructured so- cializing activities with peers that occur in the absence of authority figures. The lack of structure leaves time available for deviance; the presence of peers makes it easier to par- ticipate in deviant acts and makes them more rewarding; and the absence of authority fig- ures reduces the potential for social control responses to deviance.

Our results provide strong support for this hypothesis. We found consistent evidence

that socializing with peers away from home and authority figures is closely related to de- viant behavior, but only in the absence of a structuring agenda such as going on a date or participating in sports. The magnitude of

This content downloaded from ��������������66.77.17.54 on Sun, 11 Apr 2021 17:31:19 UTC��������������

All use subject to https://about.jstor.org/terms

652 AMERICAN SOCIOLOGICAL REVIEW

these relationships between routine activities

and deviance is exceeded only for measures

of other deviant behaviors, attitudes about deviance, and the deviant behavior of one's peers. Unlike those measures, our routine ac- tivity measures carry no direct connotations

of deviance, so they are more clearly inde-

pendent of the phenomena to be explained.

We found that the routine activities of ev-

eryday life are heavily dependent on struc-

tural variables, which supports a central theme of the routine activity perspective. For instance, in accord with Matza and Sykes's

(1961) portrayal of adolescents as a leisure

class, our results showed a consistent decline in virtually all leisure activities as respon-

dents entered adulthood. Most pertinent to our concerns, there were dramatic age, sex, and class differences in the unstructured so- cializing activities most closely associated with deviance. The regression coefficients of Table 2 imply that individuals in the most deviant structural position- 18-year-old males with D grade-point averages whose parents have graduate or professional de- grees-typically go riding in a car for fun 110 times per year, visit informally with friends 200 times, go to 40 parties, and spend

170 evenings out for fun. In contrast, 26- year-old females who had A grades in high school and whose parents had grade school educations typically go riding in a car for fun 9 times, visit with friends informally 25

times, go to 6 parties, and spend 53 evenings out for fun. Such differences suggest that routine activities are a key intersection be- tween the macro-level of social structure and the micro-level of individual lives. Accord- ingly, we found that routine activities ac- count for much of the relationship between deviance and the structural variables of age, sex, and social status. This is not to say, how-

ever, that structural differences were entirely attributable to routine activities, as substan- tial portions of some of these relationships remain to be explained.

We must address a potential alternative in- terpretation of our findings as, instead, re- flecting a short-term influence of deviance on activities. This would arise if a decision to engage in deviance precedes the decision to participate in an activity-to use the pick- up game analogy, going to the basketball court in search of a game. We attempted to

minimize this possibility by excluding from our analysis any routine activities that carry connotations of deviance (e.g., going to bars) in favor of those that do not (e.g., spending evenings out for fun). Indeed, this alternative explanation does not appear plausible for most of the relationships between particular types of unstructured socializing and specific deviant acts (see Table 1). In some cases, the logic simply does not apply. For instance, there is little sense in going to parties if your intent is to commit crimes like theft, assault. and vandalism. In other cases, the routine ac- tivities are simply too frequent (e.g., visiting with friends informally and spending eve- nings out for fun) for the less common devi- ant behaviors to generate the kind of increase needed to produce these findings. Neverthe- less, for relationships such as going to par- ties and using marijuana, it remains conceiv- able that our findings are influenced by a process such as going to parties because the parties afford opportunities to smoke mari-

juana. The routine activity perspective represents

a radical departure from traditional concerns in the study of crime and deviance. We have attempted to apply the logic of the perspec- tive to the traditional concern with explain- ing variation in individuals' rates of offend-

ing. This contributes to the utility of the rou- tine activity perspective as a unifying ap- proach to crime and deviance.

We suggest three directions for future re- search. First, we should refine and elaborate

the measures of unstructured socializing ac- tivities. Better measures would more explic- itly distinguish when authority figures are present from when they are not. Also, our set of four unstructured activities is only a narrow sample of the relevant universe of activities.

Although we have interpreted routine ac- tivities as a source of situations conducive to

deviance, we cannot rule out the possibility that nonsituational factors might influence our results. Because of our longitudinal re- search design and within-individuals analy- sis, our findings cannot be attributed to any stable individual characteristic, such as self control. Nevertheless, we have not controlled for variation over time in other explanatory factors, such as social bonds or differential

associations. Thus, there remains the possi- bility that within-individual changes in such

This content downloaded from ��������������66.77.17.54 on Sun, 11 Apr 2021 17:31:19 UTC��������������

All use subject to https://about.jstor.org/terms

ROUTINE ACTIVITIES AND DEVIANT BEHAVIOR 653

variables influenced both the routine activi- ties and deviance to generate our findings. Future research should address this interplay of situational and nonsituational influences on individual deviance. A more sophisticated research design might also address the pros- pect that our findings reflect influences of deviance on routine activities, in addition to (or instead of) the influence of activities on deviance that we have assumed.

A third direction for future research is to investigate social roles as a link between so- cial structure and the routine activities of ev- eryday life, and correspondingly routine ac- tivities as a link between social roles and de- viance. Such research would bring together the routine activity perspective and a concep- tion of social stratification as organizing ev- eryday activities through roles differentiated by factors such as age, sex, class, and race. In this vein, other research has found changes in substance use and criminal behav- ior to be closely related to changing from adolescent to adult roles in the domains of work, family, and living arrangements (Bach- man, O'Malley, and Johnston 1984; Bach- man et al. 1992; Horney, Osgood, and Mar- shall 1995; Yamaguchi and Kandel 1985). Horney et al. (1995) concluded that social control (i.e., social bonding) and routine ac- tivities are the most plausible explanations for their findings. Furthermore, Osgood and Lee (1993) provided evidence that roles in these domains are, indeed, related to routine activities.

D. Wayne Osgood is Professor of Crime, Law, and Justice and Sociology at The Pennsylvania State University. His current research interests include routine activities, criminal careers, age and deviance, and the generality of deviance.

Janet K Wilson is Assistant Professor of Sociol- ogy at the University of Central Arkansas. While continuing research in routine activities, she is also investigating Supreme Court decisions that extend rights to victims of crime.

Patrick M. O'Malley is Research Scientist and Program Director at the Survey Research Center at the Institute for Social Research at the Univer- sity of Michigan. Since 1976 he has been a co- director of the Monitoring the Future project, an ongoing study of American youth. This study pro- vides annual reports on trends in the use of psy- choactive drugs, including alcohol, tobacco, and illicit drugs.

Jerald G. Bachman is a Program Director and Research Scientist at the Survey Research Cen- ter, Institute for Social Research, at the Univer- sity of Michigan. He has directed a program of research on youth and social issues for more than 30 years. His current research interests focus on drug use and its correlates, and on the views of physicians and the public on end-of-life issues.

Lloyd D. Johnston is a Program Director and Research Scientist at the Survey Research Cen- ter, Institute for Social Research, at the Univer- sity of Michigan. He is principal investigator on the ongoing national research project entitled Monitoring the Future. He has also been involved in the development of foreign and multinational studies of substance abuse.

Appendix A. Items Measuring Routine Activities: Monitoring the Future Questionnaire Form 2, Administered 1977 to 1986

The next questions ask about the kinds of things you might do. How often do you do each of the following?

(1) Never, (2) A few times a year, (3) Once or twice a month, (4) At least once a week, (5) Almost everyday

A02A. Watch TV

A02B. Go to movies

A02D. Ride around in a car (or motorcycle) just for fun

A02E. Participate in community affairs or volunteer work

A02F. Play a musical instrument or sing

A02G. Do creative writing

A02H. Actively participate in sports. athletics, or exercising

A021. Do art or craft work

A02J. Work around the house, yard, garden, car, etc.

A02K. Get together with friends, informally

A02L. Go shopping or window-shopping

A02M. Spend at least an hour of leisure time alone

A02N. Read books, magazines, or newspaper

A02P. Go to parties or other social affairs

C25: During a typical week, on how many evenings do you go out for fun and recre- ation?

(1) Less than one, (2) One, (3) Two, (4) Three, (5) Four or five, (6) Six or seven

C26: On the average, how often do you go out with a date (or your spouse, if you are married)?

(1) Never, (2) Once a month or less, (3) 2 or 3 times a month, (4) Once a week, (5) 2 or 3 times a week, (6) Over 3 times a week

This content downloaded from ��������������66.77.17.54 on Sun, 11 Apr 2021 17:31:19 UTC��������������

All use subject to https://about.jstor.org/terms

654 AMERICAN SOCIOLOGICAL REVIEW

REFERENCES

Agnew, Robert. 1990. "Adolescent Resources and Delinquency." Criminology 28: 535-66.

Agnew, Robert and David M. Peterson. 1989.

"Leisure and Delinquency." Social Problems 36:332-50.

Bachman, Jerald G., Patrick M. O'Malley, and Jerome Johnston. 1978. Adolescence to Adult- hood: Change and Stability in the Lives of Young Men. Ann Arbor, MI: Institute for So- cial Research.

Bachman, Jerald G., Patrick M. O'Malley, and Lloyd D. Johnston. 1984. "Drug Use among Young Adults: The Impacts of Role Status and Social Environments." Journal of Personality and Social Psychology 47:629-45.

. 1991. The Monitoring the Future Project After Seventeen Years: Design and Procedures (Occasional Paper #33). Ann Arbor, MI: Insti- tute for Social Research.

Bachman, Jerald G., Patrick M. O'Malley, Lloyd D. Johnston, Willard L. Rodgers, and John E. Schulenberg. 1992. Changes in Drug Use Dur- ing the Post-High School Years (Occasional Paper #35). Ann Arbor, MI: Institute for So- cial Research.

Birkbeck, Christopher and Gary LaFree. 1993. "The Situational Analysis of Crime and Devi- ance." Annual Review of Sociology 19:113-37.

Briar, Scott and Irving Piliavin. 1965. "Delin- quency, Situational Inducements, and Commit- ment to Conformity." Social Problems 13:35- 45.

Bryk, Anthony S. and Stephen W. Raudenbush. 1992. Hierarchical Linear Models: Applica- tions and Data Analysis Methods. Newbury Park, CA: Sage.

Cohen, Lawrence E. and Marcus Felson. 1979. "Social Change and Crime Rate Trends: A Routine Activity Approach." American Socio- logical Review 44:588-608.

Eck, John E. Forthcoming. "A General Model of the Geography of Illicit Retail Market Places." In Crime and Place, edited by J. E. Eck and D. Weisburd. Monsey, NY: Criminal Justice Press.

Erickson, Maynard L. and Gary F. Jensen. 1977. "Delinquency is Still Group Behavior!: Toward Revitalizing the Group Premise in the Sociol- ogy of Deviance." Journal of Criminal Law and Criminology 68:262-73.

Fagan, Jeffrey and Edward Pabon. 1990. "Contri- butions of Delinquency and Substance Use to School Dropout among Inner-City Youths." Youth and Society 21:306-54.

Felson, Marcus. 1986. "Linking Criminal Choices, Routine Activities, Informal Control, and Criminal Outcomes." Pp. 119-28 in The Reasoning Criminal: Rational Choice Perspec-

tives on Offending, edited by D. B. Corrnish and R. V. Clarke. New York: Springer-Verlag.

. 1994. Crime and Everyday Life: Insights and Implications for Society. Thousand Oaks,

CA: Pine Forge Press. Felson, Marcus and Michael Gottfredson. 1984.

"Social Indicators of Adolescent Activities Near Peers and Parents." Journal of Marriage and the Family 46:709-14.

Garofalo, James. 1987. "Reassessing the Lifestyle Model of Criminal Victimization." Pp. 23-42 in Positive Criminology, edited by M. R. Gottfredson and T. Hirschi. Newbury Park, CA: Sage.

Gibbs, Jack P. 1981. Norms, Deviance, and So- cial Control: Conceptual Matters. New York: Elsevier.

Gold, Martin. 1970. Delinquent Behavior in an American City. Belmont, CA: Brooks/Cole.

Gottfredson, Michael R. 1981. "On the Etiology of Criminal Victimization." Journal of Crimi- nal Law and Criminology 72:711-26.

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

Heise, David R. 1969. "Separating Reliability and Stability in Test-Retest Correlation." American Sociological Review 34:93-101.

Hays, Ron D., Keith F. Widaman, M. Robin DiMatteo, and Alan W. Stacy. 1986. "Struc- tural-Equation Models of Current Drug Use: Are Appropriate Models so Simple(x)?" Journal of Personality and Social Psychology 52:134-44.

Hindelang, Michael J., Michael R. Gottfredson, and James Garofalo. 1978. Victims of Personal Crime: An Empirical Foundation for a Theory of Personal Victimization. Cambridge, MA: Ballinger.

Hirschi, Travis. 1969. Causes of Delinquency. Berkeley, CA: University of California Press.

Hirschi, Travis and Michael R. Gottfredson. 1983. "Age and the Explanation of Crime." American Journal of Sociology 89:552-84.

. 1985. "All Wise after the Fact Learning Theory, Again: Reply to Baldwin." American Journal of Sociology 90:1330-33.

Homey, Julie, D. Wayne Osgood, and Ineke Haen Marshall. 1995. "Criminal Careers in the Short-Term: Intra-Individual Variability in Crime and Its Relation to Local Life Circum- stances." American Sociological Review 60: 655-73.

Hsiao, Cheng. 1986. Analysis of Panel Data. Cambridge, England: Cambridge University Press.

Hundleby, John D. 1987. "Adolescent Drug Use in a Behavioral Matrix: A Confirmation and Comparison of the Sexes." Addictive Behaviors 12: 103-12.

This content downloaded from ��������������66.77.17.54 on Sun, 11 Apr 2021 17:31:19 UTC��������������

All use subject to https://about.jstor.org/terms

ROUTINE ACTIVITIES AND DEVIANT BEHAVIOR 655

Jensen, Gary F. and David Brownfield. 1986. "Gender, Lifestyles, and Victimization: Be- yond Routine Activity." Violence and Victims 1:85-99.

Jensen, Gary F. and Dean G. Rojek. 1992. Delin- quency and Youth Crime. 2d ed. Prospect Heights, IL: Waveland.

Johnston, Lloyd D., Patrick M. O'Malley, and Jerald G. Bachman. 1992. Smoking, Drinking, and Illicit Drug Use among Secondary School Students, College Students, and Young Adults, 1975-1991. 2 vols. Rockville, MD: National Institute on Drug Abuse.

Judd, Charles M. and Gary H. McClelland. 1989. Data Analysis: A Model-Comparison Ap- proach. San Diego, CA: Harcourt Brace Jovan- ovich.

Larson, Reed W. and Nancy Bradney. 1988. "Pre- cious Moments with Family Members and Friends." Pp. 107-26 in Families and Social Networks, edited by R. M. Milardo. Newbury Park, CA: Sage.

Larson, Reed and Maryse H. Richards. 1991. "Daily Companionship in Late Childhood and Early Adolescence: Changing Developmental Contexts." Child Development 62:283-300.

Lauritsen, Janet L., John H. Laub, and Robert J. Sampson. 1992. "Conventional and Delinquent Activities: Implications for the Prevention of Violent Victimization among Adolescents." Violence and Victims 7:91-108.

Matza, David. 1964. Delinquency and Drift. New York: John Wiley and Sons.

Matza, David and Gresham M. Sykes. 1961. "Ju- venile Delinquency and Subterranean Values." American Sociological Review 26:712-19.

Meier, Robert F. and Terance D. Miethe. 1993. "Understanding Theories of Criminal Victim- ization." Crime and Justice: A Review of Re- search 17:459-99.

Miethe, Terance D. and Robert F. Meier. 1990. "Criminal Opportunity and Victimization Rates: A Structural Choice Theory of Criminal Victimization." Journal of Research in Crime and Delinquency 27:243-66.

. 1994. Crime and Its Social Context: To- ward an Integrated Theory of Offenders, Vic- tims and Situations. Albany, NY: State Univer- sity of New York Press.

Osgood, D. Wayne, Lloyd D. Johnston, Patrick M. O'Malley, and Jerald G. Bachman. 1988. "The Generality of Deviance in Late Adoles- cence and Early Adulthood." American Socio- logical Review 53:81-94.

Osgood, D. Wayne and Hyunkee Lee. 1993. "Lei- sure Activities, Age, and Adult Roles Across the Lifespan." Society and Leisure 16:181-208.

Petersen, Trond. 1993. "Recent Advances in Lon- gitudinal Methodology." Annual Review of So- ciology 19:425-54.

Riley, David. 1987. "Time and Crime: The Link Between Teenager Lifestyle and Delinquency." Journal of Quantitative Criminology 3:339-54.

Rowe, David C., D. Wayne Osgood, and Alan W. Nicewander. 1990. "A Latent Trait Approach to Unifying Criminal Careers." Criminology 28:237-70.

Substance Abuse and Mental Health Services Ad- ministration (SAMSHA). 1993. National Household Survey on Drug Abuse: Main Find- ings 1991. (DHHS Publication No. [SMA] 93- 1980). Rockville, MD: SAMSHA.

Schlegel, Alice and Herbert Barry, III. 1991. Ado- lescence: An Anthropological Inquiry. New York: Free Press.

Short, James F. and Fred L. Strodtbeck. 1965. Group Process and Gang Delinquency. Chi- cago, IL: University of Chicago Press.

Stimson, James A. 1985. "Regression and Space and Time: A Statistical Essay." American Jour- nal of Political Science 29:914-47.

Sutherland, Edwin H. 1947. Principles of Crimi- nology. 4th ed. Philadelphia, PA: J. B. Lippin- cott.

Sykes, Gresham M. and David Matza. 1957. "Techniques of Neutralization: A Theory of Delinquency." American Sociological Review 22:664-70.

Tittle, Charles R., Wayne J. Villemez, and Dou- glas A. Smith. 1978. "The Myth of Social Class and Criminality: An Empirical Assessment of the Empirical Evidence." American Sociologi- cal Review 43:643-56.

Toby, Jackson. 1957. "Social Disorganization and Stake in Conformity: Complementary Factors in the Predatory Behavior of Hoodlums." Jour- nal of Criminal Law, Criminology, and Police Science 48:12-7.

Yamaguchi, Kazuo and Denise B. Kandel. 1985. "On the Resolution of Role Incompatibility: A Life Event History Analysis of Family Roles and Marijuana Use." American Journal of So- ciology 90:1284-1325.

Wallace, John M., Jr. and Jerald G. Bachman. 1991. "Explaining Racial/Ethnic Differences in Adolescent Drug Use: The Impact of Back- ground and Lifestyle." Social Problems 38: 333-57.

This content downloaded from ��������������66.77.17.54 on Sun, 11 Apr 2021 17:31:19 UTC��������������

All use subject to https://about.jstor.org/terms