9fio~ «v a mnn‘vrru .hdg. . I ? ; a.-- :r 5.51.: as"; '_ 4‘ 7:7 ‘1‘ ‘ ‘ ' ‘ 3 ' ,I. n‘. - ‘g""3"'/‘x§h‘5-<33¢m‘$’-‘~‘2 "' . V 5 v '..r -..«"l: ‘3‘? .... . w -:-i-;§§'.i"—‘5Z."~.mf’-. -7 u 3,.) ‘ I ‘6‘. v‘V‘JéJ; \ o q» 0"?4 .I -.-- .0.— >-O... ‘u v:a| Mafifili"‘3 ‘3.an up. {$72 “$2.: ? rawhg-u'xa- '“:.‘."".'V. V _ u ‘ O '. 3-... ' Cd-VA - ice-55743:.- w «.3 Aux. \‘ u‘ L? . - (THREE? 1 ll lllllll’ l 3 1293 00105 51mm 1 \C I ._..‘3—-"' - ’4’)" ’ CW’vm—rmm ‘1' m \I ‘ O. 2' r . E f’ 1. 12 2? .9». 2'34; E ... :- 53"? g. 1; 4 p '32< . ‘ a avaawaeet,zf££ 5.32:;4-‘35 p '0 :r‘ if '2, f " t. r: / : é":?>¢’/,‘-74~r F, ¢K4§Jg ( c’ a .213; h~<fiw This is to certify that the thesis entitled HOME, PEER, AND SCHOOL INFLUENCES ON FEMALE AND MALE SELF-REPORTED DELINQUENCY presented by Sara Margaret Wood-Kraft has been accepted towards fulfillment of the requirements for M. A. degree in Psychology MAW Major professor Date May 29 1985 Robin Redner 0.7539 MSU is an Attirmative Action/Equal Opportunity Institution MSU LIBRARIES RETURNING MATERIALS: Place in book drop to remove this checkout from your record. flfl§§ will be charged if book is returned after the date stamped below. «SEP_2;3199§7 cfgwe 3 tag; ..——.—.--—-_-_— . .. HOME, PEER, AND SCHOOL INFLUENCES ON FEMALE AND MALE SELF-REPORTED DELINQUENCY BY Sara Margaret WOod-Kraft A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF ARTS Department of Psychology 1985 ABSTRACT HOME, PEER, AND SCHOOL INFLUENCES ON FEMALE AND MALE SELF-REPORTED DELINQUENCY By Sara Margaret Wood-Kraft This study investigated the effects of positive home involvement, peer involvement, and school attendance on adolescent self-reported subcultural, minor and major property, and minor and major person offenses. The subjects were 229 male and 50 female delinquent offenders, average age 14.3 years, petitioned to the Ingham County court and referred for volunteer participation in the Michigan State University Adolescent Diversion Project between 1976 and 1980. Male and female offenders differed significantly (2 < 0.05) only in the frequency of subcultural offenses. In a stepwise regression using subscales of the ADP Life Domain Scales (LDS) and Self-Report Delinquency measure (SRD), the combined variables significantly predicted (p < .05) subcultural offenses for both males and females, and property offenses for males. The results imply that male and female offenders report similar offenses, but differ from the general adolescent population in susceptibility to the three social contexts. To John and Jennifer ii ACKNOWLEDGMENTS There are faculty, colleagues, friends, and family whose support and encouragement through this process were invaluable. I would like to thank Robin Redner, my Chair, for her assistance from conception to delivery, a labor longer than that required by Jennifer, who was conceived, delivered, and had her first birthday during it all. Anne Bogat also consistently encouraged me, and Bob Caldwell joined the process in time to provide fresh interest and focus during the final push. Mary Roberson and Jackie Williams were especially helpful and enduring, and Wes Novak, Mary Ann Reinhart, Lynn Snellman, Deb Bybee and Evette Williams were also supportive friends. John Jeppesen and Julia Parisian provided sustenance throughout, and my mother kept her faith over many trying years. John Kraft and Jennifer Wood-Kraft kept the home fires burning warmly through many make-shift meals and a long winter, and were still smiling when Spring came. Without all these folks the process and the product would have been very different. I thank them for enabling it to happen the way it did. iii TABLE OF CONTENTS Introduction Statement of the Problem The Current Research Review of the Literature Primary Social Contexts Family Peers School Summary of Primary Social Contexts Methodological Considerations Data Sources Official Self-Report Summary of Data Sources Delinquent Behaviors Subcultural Property Person Summary of Delinquent Behaviors Summary of Methodological Considerations Summary of the Problem Social Contexts Family Peers School Delinquent Behaviors Research Questions and Hypotheses Methods The Adolescent Diversion Project Subjects Measures Life Domain Survey Self-Report Delinquency Procedures Operationalized Hypotheses iv 5.; How-bk #- (NH Hid Results Frequency of Offenses Correlations among Variables Major Findings Subcultural offenses Minor property offenses Major property offenses Minor person offenses Major person offenses Summary of Major Findings Discussion Frequency of Offenses Methodology Subjects Measures Major Findings Male subcultural, minor property, and major property offenses Female subcultural offenses Major Disconfirmations Female property offenses Male and female person offenses Summary of Major Findings Implications for Further Research Subjects Frequencies of offense types General research findings Conclusion List of References Appendices Appendix A, Life Domain Survey Appendix B, Self-Report Delinquency (SRD) Questionnaire 39 39 39 43 43 49 50 52 53 53 56 56 57 57 58 62 62 65 66 68 68 69 69 70 70 73 74 77 Table Table Table Table Table Table Table Table LIST OF TABLES Offenses Included in Self-Report Subscales Frequencies of Offenses Correlations among Criterion Variables for Females Correlations among Criterion Variables for Males Simple Correlations between Predictor and Criterion Variables for Males and Females Comparison of Hypothesized Regression Equations with Results Stepwise Regression of Predictor Variables on Offense Types, Males Stepwise Regression of Predictor Variables on Offense Types, Females vi 18 40 42 42 44 45 48 Introduction In 1969, in a footnote to a book titled Causes of Delinquency, the author noted that in his study, "the girls disappear...Since girls have been neglected too long by students of delinquency, the exclusion of them is difficult to justify" (Hirschi, 1969, pp.35-36). Nine years later, in a 1978 book on juvenile justice, while commenting on juvenile crime rates Krisberg and Austin wrote, "so little is known about female delinquency that estimates have not been made" (1978, p.3). In 1982, in an article entitled "From Benign Neglect to Malign Attention," Chesney-Lind examined what she called "a veritable explosion of writings on the female delinquent," stating that "unfortunately, only some of this work can be said to have brought about a greater understanding of either the dynamics of female delinquency or the treatment of these young women who come into the criminal justice system" (1982, p.2). Statement of the Problem One reason for the continued obscurity of reliable information on female delinquency is that many of the research approaches have been gender-stereotyped. The study of female delinquency was viewed as having little relevance or meaning separate from the broader context of male delinquency and delinquent behavior as a whole. Figueira- McDonough, Barton, and Sarri (1981), in a review of the change in the research focus from gender stereotypes to gender differences and similarities, noted a current awareness of the relationship between gender differences in behavior and traditional role-stereotypes which are structured into society, commenting that as gender-specific roles decreased, there was a growing expectation that male and female behaviors would converge. In the mid-1970’s, because of an apparently dramatic increase in female delinquency rates and the coinciding rise of the women’s movement, several theorists (Adler, 1975; Adler & Simon, 1979; Simons, Miller, & Aigner, 1980) proposed that females’ increased opportunities for male- stereotyped delinquent behaviors were leading to increased criminality. Researchers investigating the reported change in the rate of female delinquency and the impact of the women’s movement on opportunities for delinquent behavior began to compile empirical data on juvenile female behaviors. Within the past decade, much information has been amassed regarding female delinquency, but the collection and compilation of data has not been systematic. Delinquency research has been affected by the variety of methodologies involved in data collection (for example, official and self- report data provide differing pictures of both crime rates and types of offenses committed by adolescent females and males), and by the variety of interpretations possible given the breadth of data. Figueira-McDonough et a1. (1981) summarize the focus of recent research as having gone from "offering separate explanations for male and female delinquency, most of them consistent with existing gender stereotypes" (p.18) to a "gender-integrated approach that investigates the conditions under which both genders will behave similarly" (p.19). The Current Research It is generally agreed that adolescents are affected, to varying degrees, by their involvements with their homes, their peers, and their schools (Bowker & Klein, 1983; Canter, 1982a;, Figueira-McDonough et a1. 1981; James & Thornton, 1980; Klemke, 1982). The following paper will examine theories and research regarding both male and female delinquency as a reaction to these three primary social contexts. It will also discuss some of the methodological complications in interpreting these data, specifically the differences between official and self-report delinquency data (SRD) and the issue of the types of offenses with which adolescents are typically involved. Male and female delinquent behaviors will be compared with regard to frequency and severity of subcultural, property, and person offenses, and a research project using data from the Michigan State University Adolescent Diversion Project will be described. Review of the Literature Primary Social Contexts In their discussion of adolescent deviance, Figueira- McDonough et a1. (1981) referred to the "primary social contexts" of adolescence as the family, the peer group, and the school. They proposed that the three contexts compete as sources of norms, suggesting that adolescent behavior is determined by the extent of the youth's attachment to each context. The following sections will review some of the theories and research which have.considered these aspects of delinquency. Family Researchers have tended to agree that familial relationships and interactions impact on the development of both male and female delinquency. Morash (1983) cited major assumptions regarding families and delinquency, stating that delinquent behavior is promoted by family conflict and parental rejection, and also by parental patterns of discipline. Simons et al. (1980) surveyed 3925 high school students in Iowa and found reported parental rejection to be a significant predictor of delinquency for both males and females. Klemke's (1982) "deviant vulnerability hypothesis" posited that youth experiencing problems within their major social involvements would be more vulnerable to delinquency involvements. His self-report investigation of shoplifting indicated that shoplifting activity was reported more often by youth who were having problems with their parents. Examining the impact of family relationships on male and female delinquency, Canter (1982a) studied 1725 adolescents and found modest but significant correlations between family variables and delinquent behavior. She found that family variables were significant predictors of delinquency for both sexes, and "stronger predictors of serious [italics added] offenses...for males than for females" (p.159). There are some methodological complications in Canter’s analyses which will be addressed in a later section. Because of the large size of her sample, however, it is useful to report her results as at least an example of the research that has been done in this area. Canter concluded that "while family bonds are generally stronger among females, their inhibitory effects on serious delinquency appear to be stronger for males" (p.163). Figueira-McDonough et a1. (1981) reported that in their analysis of the SRD responses of 1735 high school students, the attachment to parents was indirect and moderate and was similar for both males and females. Sarri (1983) reported that females who perceived that their parents disapproved of delinquency were less likely to be delinquent, while male concern with parental approval and disapproving norms did not result in lower delinquent behavior. Canter’s study suggested that serious female delinquency is less related to family variables than is serious male delinquency. Sarri’s seemingly contradictory results suggested that family norms have more impact on overall delinquent behavior for females than for males. The differences between the two studies may be explained by their methodologies. Sarri did not differentiate among types of offenses, whereas Canter’s conclusions were based on commission of felony theft and index offenses, which account for a relatively small percentage of all delinquent behaviors. The greater number of less serious offenses in Sarri's study may have counterbalanced any results regarding the impact of family variables on "serious" crimes. I Regarding specifically female delinquency, Morash (1983) proposed that girls with close familial relationships would develop a concern for others which would lead to an avoidance of delinquent behavior, but those who were not positively involved with their families because of tensions or parental rejections would exhibit a "tendency toward delinquency." James and Thornton (1980), in their study of 287 high school and junior high females, found that parental social control tended to inhibit social delinquency and suppress property delinquency but did not influence aggressive delinquency. Bowker and Klein (1983) studied data on 229 female gang members and controls, and found no relationship between delinquent behavior and family structure and family socioeconomic status. They suggested that "in general, a girl’s relations with her mother, father, and other adults appear to have little to do with whether she joins a gang or whether she commits delinquent acts" (p.744). The Bowker and Klein data, however, like those of Sarri and Figueira-McDonough et a1., do not include information regarding the seriousness of the offenses being considered. Because their study was based on official data, it is possible that the offenses reported were more serious than those which appear in SRD studies, and that the relationship between home involvement and serious offenses is different from the connection between home involvement and less serious but more commonly reported SRD offenses. Several researchers (Canter 1982a; Feyerherm, 1981) have suggested that the impact of the family variables may vary with the type of offense committed. Of the studies discussed above, only James and Thornton’s study of females and Canter’s study of family variables considered the variety of offenses which may be called delinquent. It is possible that, had Bowker and Klein, Figueira-McDonough et a1., and Sarri considered the severity and type of the reported offenses, the impact of the family variable might have been different. When type and severity of offense are considered, there are reports of impact for family variables. Generally, however, research does not report the type of offense, and the impact of family involvement on both females and males is unclear. This important issue will be discussed in a later section. 232$ Beyond the family perspective, researchers have also examined the impact of peer relations on the development of delinquency. Morash (1983) proposed that delinquent behavior is strongly influenced by association with delinquent peers. Figueira-McDonough et a1. (1981) reported that most adolescents disapprove of subcultural deviant behavior and have a generally low level of peer involvement. They concluded, however, that for both sexes high participation in social activities with peers was strongly related to subcultural deviance, and for youthSwho were very involved in both subcultural and criminal behavior, they found that high activity with peers was an important variable. Klemke’s (1982) "deviant socialization hypothesis" suggested that youth become involved in deviant behaviors similar to those of the youths’ significant others. Klemke’s shoplifting study found that more than 60 percent of the offenders had done their shoplifting with accomplices. Of the total group surveyed, 56.9 percent had close friends who shoplifted, and youth with significant others who had shoplifted reported much more frequent recent shoplifting activity. In a study of the court records of 255 juvenile males and females, Aultman (1980) found that 63 percent of the male and 57 percent of the female offenses involved group participation. She did not, however, report the proportions of participation of females and males in specific types of offenses. Focusing on female delinquency, Bowker and Klein (1983), having rejected the theories of family influence, embraced the notion that peers may determine delinquent female behavior. They hypothesized that gang membership and delinquent behavior might be determined much more by peer relationships than by parental or other adult relationships, and cited Giordano (1978) in her conclusion that "[girls’] participation in delinquent activities...was significantly related to approval of delinquent participation from one's girlfriends" (pp. 740-41). In their own research, Bowker and Klein found that for females, both number of offenses committed and gang membership were affected by relationships with girlfriends. Simons et a1. (1980) found that for their female subjects, "values of friends" correlated with self- reported measures of delinquent activity. James and Thornton (1980) found that social support for female 10 delinquency was related to social delinquency, property delinquency, and aggressive delinquency. Researchers agree that peers affect delinquency for both males and females. There remain, however, questions concerning the relative importance of peer activities for each gender, the different relationships with different types of delinquent behavior, and the relationships between gender, family involvement, and peer involvement. School In addition to family and peer involvements, levels of educational involvement and activity can impact on delinquency. Many theorists have proposed that delinquency is a response to a disruption in the academic arena. Bowker and Klein (1983), Chesney-Lind (1982), Rankin (cited in Chesney-Lind, 1982), Klemke (1982), Sarri (1983), and others have suggested that negative attitudes toward school and poor academic performance are predictive of delinquent behavior. Comparing the effects of school on male and female delinquency, Klemke (1982) found that academic performance and attitudes toward school related to shoplifting more strongly for females than for males, with shoplifting better predicted by school experiences than by economic or family variables. Chesney-Lind (1983) cited Rankin's study of school factors and delinquency, which concluded that "negative attitudes towards school and poor school 11 performance were both significant in predicting delinquency but, contrary to... expectations, this relationship was stronger for girls than for boys" (p.9). Sarri (1983) found attachment to school to be a strong inhibitor of delinquency for females, but not for males unless it was coupled with other variables. Phillips and Kelly (1979), discussing the impact of school on male delinquency, determined that "the contention that school failure produces at least some adolescent misconduct should be taken seriously" (p.204). They cited Elliot’s (1966) conclusion that male delinquency rates declined after the boys dropped out of school. Figueira-McDonough et a1. (1981), examining the same data as Sarri, concluded that school attachment was a strong influence on girls, showing a negative association with subcultural deviance. Bowker and Klein (1983), in their study of black juvenile females, found a tendency for gang members and serious delinquents to be less likely to plan to finish high school or to go to college. They also found that the delinquent girls did not anticipate positive effects from being involved in school. In summary, research suggests that there are relationships between school and delinquent behavior for both females and males, but the directions of the relationships and the types of offenses for which they occur are not clear. Summary 9; Primary Social Contexts In examining research regarding theories of 12 delinquency, theories which attribute delinquent behaviors to the individual’s negative involvement with the family, peers, and the school system have been considered. Research suggests that the family situation impacts strongly on delinquent behavior, although some argue that the impact is not as strong as other variables. There are indications that peer influences are primary, and also that educational functions are the best predictors of female delinquency. In general, there is support for the social-psychological theory that delinquency is related to disruptions in the home, peer, and school environments. What is not clear, however, is the extent to which these factors vary in significance with various types of delinquent behaviors, or the extent to which these variables have similar or different impacts on males and females. The next section will discuss in greater detail the methodology used in obtaining delinquency data and the complications which develop from inconsistencies in categorizing and interpreting these data. Methodological Considerations Two of the most significant variables to consider in any research on gender similarities and differences in delinquency are the sources of the recidivism data and the severity of offenses for which the youth have become involved with the court and judicial systems. The data 13 sources vary, but may be divided into two major categories; the official or archival data, that is, police and court records, and self-report delinquency (SRD) data. Data Sources Official. Official records may vary in the quantity and types of crimes recorded. Police and court records include statistics for arrests, convictions, and some warnings and other dispositions. They cover serious and moderate offenses, but do not include the types of minor crimes which may be reported in a self-report questionnaire (stealing items worth less than $2, or school-related offenses, for example). Official data, described by Selke (1982, p.398), "provide the most comprehensive estimates of general crime levels and police arrest practices." As Dunford and Elliott (1984) point out, however, official data represent "recorded official reactions to observed violations of the law rather than measures of actual behavior" (p.79). In their study of 1725 youth, they determined that "arrest data reflect only a small fraction of the illegal activity occurring in our communities. The relationship between official arrest and the frequency and/or seriousness of illegal acts reported by youth is extremely crude and appears to involve a threshold effect in which the likelihood of arrest is close to zero until one reports in excess of 100 total offenses..."(p.81). Williams and Gold (1978), in their discussion of the 14 difference between self-reported delinquent behavior and official delinquency, point out that "official delinquency is defined by official response to alleged delinquent behavior"(p.309). In their study of 847 adolescents, they discovered that in the three years prior to their interviews less than one percent of the self-reported offenses were recorded as official delinquency. They concluded that police records were not related to the frequency of delinquent behavior and were only slightly related to offense severity. Klemke’s (1982) study of shoplifting indicated that, of female adolescents who shoplifted, only 20% were apprehended. He cited Hindelang’s 1974 study which indicated that only 26% of the apprehended shoplifters were referred to the police. These studies and others suggest that although official data may be useful for examining the legal/judicial system, an accurate picture of delinquent behavior may best be derived by considering alternative data sources. Self-Report. Self-reported delinquency data are usually obtained through interviews or through questionnaires, which are usually completed voluntarily and are frequently anonymous. Self-reports of frequencies of crimes may include otherwise unreported and unwitnessed incidents. Statistically, the total number of self-reported delinquent behaviors is much greater than the number of officially recorded delinquent acts (Canter, 1982b; Dunford 15 & Elliott, 1984; Williams & Gold, 1978). Canter (1982b), in her discussion of self-report measures, concluded that while official records measure official reactions to delinquent behavior, self-report measures elicit information directly, eliminating "the confounding effects of differential and reporting practices with the actual volume and types of behavior" (p.375). There are, however, some methodological issues with the use of self-report data. Dunford and Elliott (1984) describe Elliott and Ageton’s 1980 criticisms of the usual self- report measures: Criticism has focused primarily upon the unrepresentativeness of items, that is, overrepresentation of trivial, nonserious offenses, including some that are not technically violations of the law, and normative response sets ("never", "once or twice", and ”three or more times") that lump offenders with high frequencies together with those reporting relatively low frequencies, a potentially serious problem when attempting to identify high-frequency offenders. (p.62) Many investigators have reported the finding that ratios reflecting gender differences are much lower in SRD than in official or archival data (Canter, 1982a, 1982b; Feyerherm, 1981; Williams & Gold, 1978). Explanations for these lower ratios are two-fold: first, the numbers of 16 trivial or subcultural offenses reported in SRD data greatly increase for both males and females, leading to a more equal distribution; second, the use of ’global' scales-may mask some of the differences occurring between male and female rates of commission of specific offenses. Feyerherm (1981) cited criticisms of the use of 'global’ scales in SRD, pointing out that single indices or total scores could totally confound sex differences. He quoted the conclusion of Hindelang et al. (1979) that "sex differences in self- report data are highly contingent on item content" (p.48). Canter (1982b) suggested that "it is essential...to consider findings at the item and more homogeneous scale levels as well as at the summary scale level...[though] item-level findings are not as reliable as findings concerning groups of items" (pp.381-2). Summary of data sources. The differences between self- report and official rates of delinquency have been explained by biases in the criminal justice system and by issues of methodology, specifically, the construction and use of SRD scales. Because studies suggest that the proportional representation of females and males in official data may be very different from the reports of adolescents themselves, and because studies also suggest that a broader range of information is available from SRD data, for purposes of this paper delinquent behavior will be considered from the base of self-reported delinquency. l7 Delinquent Behaviors When examining the varying rates of reported delinquency it is necessary to consider not only the sources of the data but also the relative severity of the crimes which male and female juveniles are likely to report. Feyerherm (1981) discussed the need to examine male/female ratios for specific sets of behaviors. In this consideration it is important to remember the differences between "status" or "subcultural" and "non-status" offenses, as well as the non-arrestable behaviors which adolescents may report as a crime. Researchers have categorized delinquent offenses in a variety of ways. Some examples are: trivial, moderate, serious, very serious (Wright, 1983); major or minor property or aggression, major or minor subcultural (Sarri, 1983); violent or property (Aultman, 1980); subcultural, person, or property (Figueira-McDonough et al. 1981). For purposes of this paper, offenses will be divided into subcultural, minor and major property, and minor and major person offenses. For offenses included in each category, see Table 1. These five categories are rationally determined combinations of the Sarri and Figueira-McDonough subscales; the items were selected from the Self-Reported Delinquency scale (Appendix B) used by the Michigan State University Adolescent Diversion Project, to be described in greater detail in a later section. The categories encompass 18 Table l Offenses Included in Self-Report Subscales (IN THE LAST YEAR, HOW OFTEN HAVE YOU:) SUBCULTURAL OFFENSES smoked marijuana taken drugs or pills, other than marijuana skipped class when in school taken something worth less than $2 drunk beer or liquor, including sips run away from home skipped a full day of school PROPERTY OFFENSES MINOR MAJOR' gone onto someone's land when they didn't want you to be there, or without their permission gone into a house or building when you weren't supposed to be there taken something from a store without paying for it (regardless of price) taken things worth less than $50.00 (over $2.00) taken something not belonging to you worth over $50.00 set fire to someone else's property taken a car without the owner's permission (includes joyriding) broken into a place and stolen something 19 Table 1 (cont'd) PERSON OFFENSES MINOR O threatened to hurt someone 0 hit a member of family (in anger) O beaten up on somebody or fought someone (physically) MAJOR 0 hurt someone badly enough for him/her to need bandages or a doctor 0 taken something from a person by force (w or w/o a weapon) 0 used or threatened to use a weapon to get something from a person 1) Never 2) Once 3) Twice 4) More than twice 20 the broad range of offenses while allowing discrimination between categories which may differ in proportions of females and males. Subcultural. Subcultural offenses have been defined by Figueira-McDonough et al. (1981) as "behaviors that aim to challenge adult authority and are often imitations of adult roles that have been defined as illegitimate for adolescents" (p.21). Subcultural offenses may include the status offenses, and such "tangential" offenses as marijuana smoking, theft of less than $2, school vandalism, and drug offenses. SRD studies suggest that the proportions of male and female offenders are similar for subcultural offenses as a whole (Canter, 1982b; Figueira-McDonough et a1. 1981; Sarri, 1982). Property. Considering male and female differences in property offenses, researchers have drawn various divisions among the types of offenses considered. Steffensmeier and Steffensmeier (1980) considered "petty property" offenses to include larceny-theft, fraud, forgery, and embezzlement. They pointed out that larceny-theft, an index offense, is normally considered a serious crime, but that such data sources as the Uniform Crime Report do not distinguish between petty and grand thefts. Reorganizing the data considering that much of the female larceny-theft is shoplifting, they determined that serious property crime continues to be committed by a much higher proportion of 21 males than females. Feyerherm (1981) divided property offenses among categories of Malicious, Theft, and Serious Delinquency. He determined that females committed fewer offenses than males in all three areas, commenting that "the ratio of male to female involvement increases as the seriousness of the offense category increases" (p.48). Wright (1983), using categories of trivial, moderate, serious, and very serious, determined that male and female behavior were similar at comparable levels of shoplifting and burglary with no Breaking and Entering, but that males began to significantly outnumber the females as the degree of offense became more serious (forcible robbery, B & E, car larceny). Canter (1982b) found significant gender differences in her categories of Property Damage and Index Offenses, but not in the categories of Felony or Minor Theft. The lack of difference in these categories may, like the results of Steffensmeier and Steffensmeier (1980) and Wright (1983), reflect the inclusion of such larceny offenses as shoplifting, a behavior more probable for females than, for instance, auto theft. Person. Considering self-reported male and female adolescent offenses against persons, Sarri (1983) reported that the number of serious person crimes was extremely small (between 1 and 2 percent). In Canter’s (1982b) study, the categories of Crimes against Persons and Felony Assault 22 showed significant gender differences. Figueira-McDonough et a1. (1981) found that the ratio of males to females for aggressive assault was 2.97 to l, but in their overall category of Person offenses, they did not discriminate between major and minor offenses. Wright (1983) found that, although male and female behavior was similar for what she considered "trivial" or "moderate" offenses, the ratios for more serious offenses increased dramatically, with victims of males 4 times as likely as females to require major medical care. Summary 9f delinquent behaviors. In sum, males and females seem to vary in their participation in delinquent behaviors. Several studies have suggested that the gender difference is much less for subcultural behaviors than for more serious crimes (Canter, 1982b; Figueira-McDonough et a1. 1981; Sarri, 1983). Although there have been studies which focused on the subcultural, property, and person types of delinquent offenses, they have tended to focus on one or two of the three types, and have failed to discriminate between minor and major offenses within the types. There are indications that males commit more frequent and more serious property and person offenses than females (Wright, 1983; Sarri, 1983), but the proportions of these differences are not yet firmly established. Summary 9: Methodological Considerations This section has discussed two types of variables which 23 have continued to confound the study of gender differences in delinquent behavior. The differences between official and self-report data, and the possible complications with using either (bias in the legal system in the former; the use of trivial questions and ’global’ scales in the latter) have affected the accuracy and interpretability of the available data. Additionally, variability in the definitions and categories of offenses has obfuscated male- female comparisons for specific offense types. The literature suggests that while subcultural offenses are similar for males and females, there are consistent gender differences in both the frequency and the seriousness of crimes against property and persons. Summary 9; the Problem In considering gender differences and similarities in delinquent behavior, three sets of information have been discussed. These are: the primary social contexts of family, peers, and school; the research methodology involved in collecting and interpreting self-report data; and the categories and frequencies of the specific delinquent behaviors being measured. To understand delinquency as a social phenomenon, a great deal remains to be learned regarding each of these areas, as well as the interconnections among them. Much of the research discussed above has been specific and thorough, but many gaps remain to be filled before there is an adequate 24 empirical baseline for reliable studies of etiology and prevention and treatment programs for both males and females. In the discussion which follows, some of the informational gaps are highlighted, and certain hypotheses are proposed which may begin to answer some of the more immediate questions. Social Contexts Research indicates that family, peer, and school involvement impact on the delinquent behaviors of both males and females, but more information is needed to determine whether impacts are similar or different, and whether the three variables have differing impacts on different types of crime. Family. Canter (1982a) suggested that family variables are important to both males and females, but may have more impact in inhibiting serious offenses among males. Her research found that serious female delinquency was less related to family variables than was serious male delinquency, but it is possible that, as with many other studies of adolescent delinquency, Canter’s findings reflected her methodology rather than the situation. This is an issue which will be discussed later in this paper. Figueira-McDonough et a1. (1981) found no gender differences in their parent attachment category. In Sarri’s (1983) study of the same data, she determined that when parents 25 were reported to disapprove of delinquent behavior, girls who were close to their parents "seemed to shy away from delinquency" while boy’s concerns with parental disapproval "did not result in lower delinquent behavior" (p.389). Bowker and Klein (1983) did not find family variables related to delinquency in females, but their study used official data and did not report the types of offenses included. James and Thornton (1980) suggested that family variables impact for females on subcultural delinquency, less on property delinquency, and least of all for aggressive delinquency. In most cases where studies seem to be contradictory, the results reflect methodological issues, specifically, the SRD-official data differences, and the lack of discrimination between offense types. Further research is needed to clarify the relationships between the genders, the impact of family involvement on both males and females, and the relationship between family involvement and the severity of reported offenses. nggg. Studies have indicated that involvement with peers varies for both males and females with the type and degree of delinquency (Aultman, 1980; Figueira-McDonough et a1. 1981; Klemke, 1982; Sarri, 1983). Figueira-McDonough et a1. and Sarri suggested that involvements are very similar for subcultural offenses. Considering violent and property offenses, Aultman (1980) found that violent acts were "more likely to be committed by youths acting alone rather than in 26 groups" while "nonviolent offenses are more likely to be committed in groups" (p.188). She did not report male/female ratios for these behaviors, nor did she consider possible interactions between peer and other social involvements. Other studies have suggested the importance of peers in female delinquency (Bowker & Klein, 1983; James & Thornton, 1980; Simons et a1. 1980) but did not complete the picture with comparable information for males. School. Adding the third social context variable, school involvement, leads to still broader questions regarding the relative impacts of these variables on each gender for different types of crimes. Research has suggested that school involvement is related to some types of delinquency, and that it has more importance to females than to males (Rankin, cited in Chesney-Lind, 1983; Phillips & Kelly, 1979; Sarri, 1983). Sarri reported that "attachment to school was a strong inhibitor of delinquency for females, but had relatively little effect on males unless it was coupled with other variables” (p.389). The types of behaviors affected are not clear, nor is it clear how the family and peer variables related to school involvement. Delinquent Behaviors Unfortunately, researchers studying delinquency have not agreed on specific behaviors to study. The range of behaviors from the "trivial" subcultural and status offenses 27 through the criminal offenses, including the Index crimes, has rarely been considered within a single study. As no empirical scales have been derived which could cover and discriminate among this range of offenses (in part because the most serious offenses are extremely rare), investigators have continued to create rational subscales. Most categories seem to be included within a scheme of subcultural, minor and major prOperty offenses, and minor and major person offenses. When the data are divided into these categories, the relationship between self-reported offenses for males and females may become clearer. The current research was designed to address these various issues in the study of gender and delinquency. It was intended to provide further information regarding the relative impacts of the primary social contexts on the types of offenses reported by male and female adolescents. Research Questions and Hypotheses In an effort to bridge some of the information gaps discussed above, the present study attempted to answer the following research questions: Is there a difference between male and female adolescent offenders’ self-reports of delinquent activity? Are the self-reported delinquent behaviors of female offenders affected by the three social contexts, positive home involvement, peer involvement, and school attendance, and in what ways? 28 Do the effects of these contexts vary with the types of offenses reported by female offenders? Are the self-reported delinquent behaviors of male offenders affected by the three social context variables, and in what ways? Do the effects of these contexts vary for males with the types of offenses they report? In light of the research discussed above, the following results were expected: There would be no statistically significant difference between females and males in the reported frequencies of subcultural offenses, and in the property and person offense categories the rates would be similar for minor offenses, while males would exceed females in major offenses. For males, a lack of positive home involvement would be the primary predictor of all but the minor person offenses, for which peer involvements were expected to be primary. With the exception of minor person offenses, peer involvement would be the second most effective predictor of offenses, and the relationships would be positive for the subcultural, property, and minor person offenses, but negative for major person offenses. School attendance was expected to be the third strongest variable in the prediction of all offenses, and the relationships were expected to be positive for the subcultural and minor property and person offenses, but negative for the major 29 offenses, both person and property. For females, it was expected that a lack of positive home involvement would be the best predictor of subcultural, major property, and major person offenses, while peer involvement would be the best predictor of minor property and person offenses. It was expected that a lack of positive family involvement would be the second strongest predictor for minor property and person offenses, and that peer involvement would be the second predictor of the subcultural and major offenses. With the exception of major person offenses, the relationships for females between delinquent offenses and peer involvement were expected to be positive. School attendance was expected to be the third strongest predictor, and with the exception of minor person offenses, for which the direction of relationship was not predicted, the relationships were expected to be negative. The following section will describe the methods used in investigating the research hypotheses. 30 Methods The Adolescent Diversion Project At the Michigan State University Department of Psychology between 1976 and 1980, a research project called the Adolescent Diversion Project (ADP) investigated the use of college students as change agents in the lives of juveniles who had been referred to the court for a violation of the juvenile code. The Project was funded by a federal grant from the National Institute of Mental Health. The youths were diverted from the intake Division of the Ingham County Juvenile Court. Only youths who admitted committing the charged offense were referred to the Project, and court officers reported that they did not refer juveniles who without the diversion project would have been warned and released. Because the Adolescent Diversion Project was a study of the effects of treatment, extensive records were kept regarding the backgrounds and crimes of the offenders. Using these data, it was possible to derive a clear description of the youth involved in delinquent behavior in the Lansing, Michigan area during the years 1976 to 1980. Significantly for the purposes of this investigation, records were available for both male and female offenders. Data which were available from the ADP included self-report data and sociological and demographic information. With 31 these types of records, it was possible to examine and compare several aspects of male and female delinquency. Subjects The subjects in this investigation were the 50 female and 229 male juvenile offenders who participated in the ADP between 1976 and 1980. For the purposes of this study, the subjects were the two groups, male and female offenders, whose average ages were 14.3 years. Because the number of female offenders was small, subjects were not further divided into sub-groups according to race, socio-economic status, or other variables. Williams and Gold (1978), in their study of delinquent behavior and official delinquency, determined that these demographic variables had much more impact on official data than on the type of self-report data being considered in this study. The ADP youth were originally referred to the project by the Intake Department of the Ingham County Juvenile Court, following a preliminary hearing at which they had admitted committing the offense with which they were charged. The Intake Referee described the project to the parents and the youth, and if the family was interested, the referee scheduled an initial interview. Youth who agreed to participate understood that involvement was voluntary, that the ’treatment’ would last for 18 weeks, and that they would be required to release all school, police, and court records as well as participating in a series of interviews. 32 The data used in this investigation were gathered by undergraduate and graduate students in the Michigan State University Department of Psychology. These students were participants in the Adolescent Diversion Project, and collected the data for other research purposes. They were trained within the Project in specific interviewing techniques, test administration, data collection, and coding, and were supervised regularly by the administrators of the Project. Measures Two of the Project Measurements were used in this investigation: the Life Domain Survey (LDS) which includes information about the youth’s family, school, and social environments, and the Self-Reported Delinquency measure (SRD). Life Domain Survey. This measure was devised to assess the impact of intervention on various aspects of the lives of juvenile offenders. It includes questions which measure the youths’ perceptions of their relationships with their families, peers, and school systems. It is an interview procedure, developed in a sequence of stages which included generating rational items, rejecting those with over 90% endorsement in one response category, creating rational subscales, and testing internal consistencies and inter- interviewer reliabilities. The scales of interest to this study were Scale 1, Positive Home Involvement (PHI); Scale 33 4, Peer Involvement (PI), and one item from Scale 9, School Involvement, which asks how regularly within each week the youth attends school (SA). These first two scales are listed in Appendix A. Scale 1, Positive Home Involvement (PHI), measures the extent to which the youth spends time at home and participating in activities with parents and siblings. Previous research had demonstrated an internal consistency alpha for this scale of 0.78, and an inter- interviewer correlation of .81 (Parisian, 1982). Scale 4, Peer Involvement (PI), measures the amount of time and types of activities which the youth shares with peers. The internal consistency alpha for this scale was .75, and the between-interviewers correlation was .47 (Parisian, 1982). Items in these scales ask how often specific events have occurred in the last six weeks, with the five possible answers ranging from "never" to "more than five times." The LDS was conducted with each youth four times during involvement with the Project. Data for the current research were drawn from the first of the interviews, covering the time period six weeks prior to beginning the Project. Self-Report Delinquenqy measure. The measure which was used is included in Appendix B. It includes frequently committed delinquent behaviors, as well as several positive items, and requires two responses to each item: the frequency of the behavior within the last six weeks and within the past year. Possible frequencies are, "never," 34 "once", "twice", and "more than twice." According to Parisian (1982), the reliability of the scale was established via internal consistency analyses and test- retest methods. Information from the SRD which was used in the current study included frequency of offenses, frequency of arrestable status offenses and certain subculturally related items, and frequency of specific arrestable non- status offenses committed within the 12 months prior to participation in the Project. For purposes of this study, the scale was divided into five rational subscales (listed in Table 1, p.18), allowing greater information about specific offenses. Each subscale was computed by adding the total frequencies of SRD offenses within that category. Procedures The LDS and SRD measures were administered to the subjects by student interviewers trained in the ADP Project. The measures were considered parts of a process interview, which began with open-ended questions and continued with increasingly specific probes (Parisian, 1982). The interviews were conducted at the subjects’ homes, in private and with assured confidentiality, and the participants were paid $5.00 to complete the interviews, which lasted between one and one and one-half hours. In all cases, to prevent error variance due to reading skills, the items were read to the subjects and coded by the interviewers. Interrater reliability was established by independent rating by two 35 interviewers once in every four to six interviews. Operationalized Hypotheses In order to compare the frequencies of offenses committed by males and females, two-tailed E-tests were run on the five types of offenses; subcultural, minor and major property, and minor and major person. Given the current literature on these offenses, it was expected that for subcultural offenses there would be no statistically significant difference between females and males, and that in the property and person offense categories the rates would be similar for minor offenses, while males would exceed females in major offenses. To examine the relative impacts of peer and positive home involvement and school attendance on delinquency, the following hypotheses were tested using a stepwise regression with variables entered in order of their hypothesized importance: 1. For males, subcultural offenses, measured by the SRD Subcultural subscale, will be impacted most by Positive Family Involvement (PHI), somewhat by Peer Involvement (PI), and least by School Attendance (SA). The relationships between the SRD subscale and the Life Domain scales will be negative for PHI, positive for PI, and positive for SA. 36 For females, subcultural offenses, measured by the SRD Subcultural subscale, will be impacted most by Positive Family Involvement (PHI), somewhat by Peer Involvement (PI), and least by School Attendance (SA). These relationships between the SRD subscale and the Life Domain scales will be negative for PHI, positive for PI, and negative for SA. For males, minor property offenses, measured by the SRD Minor Property Offense subscale, will be affected similarly to subcultural offenses, described above. For females, minor property offenses, measured by the SRD Minor Property Offense subscale, will be impacted most by Peer Involvement, somewhat by Positive Family Involvement, and least by School Attendance. The relationships between the subscale and the Life Domain scales will be positive for PI, negative for PHI, and negative for SA. For males, major property offenses, measured by the SRD Major Property Offense subscale, will be impacted most by Positive Family Involvement, somewhat by Peer Involvement, and least by School Attendance. The relationships between the subscale and the Life Domain scales will be negative for PFI, positive for PI, and negative for SA. 37 For females, major property offenses, measured by the SRD Major Property Offense subscale, will be impacted most by Positive Family Involvement, somewhat by School Attendance and least by Peer Involvement. The relationships between the subscale and the Life Domain scales will be negative for PFI, negative for SA, and positive for PI. For males, minor person offenses, measured by the SRD Minor Person Offense subscale, will be related most to Peer Involvement, somewhat to Positive Family Involvement, and least to School Attendance. The relationships between the subscale and the Life Domain scales will be positive for PI, negative for PHI, and positive for SA. For females, minor person offenses, measured by the SRD Minor Person Offense subscale, will be related most to Peer Involvement, somewhat to Positive Family Involvement, and least to School Attendance. The relationships between the subscale and the Life Domain Scales will be positive for PI, negative for PFI, and for SA there is no prediction. 10. 38 For males, major person offenses, measured by the SRD Major Person Offense subscale, will be related most to Positive Family Involvement, somewhat to Peer Involvement, and least to School Attendance. The relationships between the subscale and the Life domain scales will be negative for PHI, negative for PI, and negative for SA. For females, the predictions for major person offenses are similar to those for males. 39 Results Frequenqy of Offenses To examine the differences between frequencies of types of offenses committed by male and female subjects, a two- tailed E—test was run on these groups (229 males and 50 females). It was expected that there would be no differences in subcultural, minor property, and minor person offense frequencies for males and females, and that males would commit significantly more major property and person offenses than females. However, the only significant difference between the two genders was in the subcultural offense group, where the average number of offenses committed by females significantly exceeded those committed by males, 3(277) = 2.58, p < .05 (see Table 2). While not significantly different, males did commit more minor property, major property, and major person offenses than females. Correlations among Variables A correlation matrix was established to examine the relationships between the variables. Considering the predictor variables of home and peer involvement and school attendance, for females there were no significant intercorrelations. For males, there was a significant positive relationship between positive home involvement and school attendance, and a significant negative relationship 40 Table 2 Frequencies of Offense Females Males (n=50) (n=229) Offenses _ _" 3 (two- x S.D. x 5.0. g (277) tailed) Subcultural 10.78 5.23 8.68 5.20 -2.58 .01 Minor PrOperty 3.98 2.79 4.14 3.17 .32 .75 Major Property 1.08 1.43 1.48 1.84 1.44 .15 Minor Person 4.72 3.00 4.68 2.85 -.09 .93 Major Person .88 1.59 1.18 1.67 .92 .36 41 between peer involvement and school attendance, Pearson’s g = .19 and -.32, respectively, p < .05. Considering the criterion variables (the five offense categories), for females there were significant positive relationships (9 < .05) between subcultural and minor property offenses; between minor and major property offenses; and between major person offenses and minor property, major property, and minor person offenses (see Table 3). For males, the correlations among all the offense types were statistically significant (see Table 4). Subcultural offenses were highly correlated with minor property, g = .49, p < .05; the minor prOperty and major property offenses and the minor person and major person offenses were highly positively correlated, 5’s = .48 and 0.46 respectively, p < .05; and minor person and minor property offenses were also strongly positively related, g = 0.41, p < .05. Considering the simple correlations between predictor and criterion variables, for females, only the positive relationship between peer involvement and subcultural offenses was significant, 3 = .39, p < .05 (see Table 5). For males, there was a positive correlation between positive home involvement and minor property offenses, g = .13, p < 0.05; peer involvement was significantly and positively 42 Table 3 Correlations among Criterion Variables for Females (n=50) Sub- Minor Major Minor Major cultural Property Property Person Person Subcultural 1.00 *.50 .06 .04 .08 Minor Property 1.00 *.30 .05 *.29 Major Property 1.00 .13 *.26 Minor Person 1.00 *.42 Major Person 1.00 * p<:.05 Table 4 Correlations among Criterion Variables for Males (n=229) Sub- Minor Major Minor Major cultural Property Property Person Person Subcultural 1.00 .49 .38 .27 .26 Minor PrOperty 1.00 .48 .41 .29 Major Property 1.00 .23 .25 Minor Person 1.00 .46 Major Person 1.00 all p (. 05 43 related to all offense types; and school attendance was significantly and negatively related to all but major person offenses (see Table 5). Major Findings To refine the statistical analysis and increase the predictability of the variables, a stepwise regression was performed with variables entered in order of their hypothesized importance. See Table 6 for a summary of hypothesized regressions and results. Subcultural Offenses Males. It was hypothesized that for males, subcultural offenses would be predicted by the variables positive home involvement (PHI), peer involvement (PI), and school attendance (SA), in that order, and that the relationships would be negative for PHI, positive for PI, and positive for SA. Results of the stepwise regression indicated that subcultural offenses were significantly predicted by the combined effects of the home, peer, and school predictor variables, overall F(3,225) = 32.59, p < .05. Thirty percent of the variance in subcultural offenses was attributable to these three variables (R1 =.30). The impact of home involvement alone, however, was not significant. Peer involvement was a major contributor (r2 = .25) with school attendance adding somewhat to the prediction (r1 change = .04). For both peer involvement and school 44 Table 5 Simple Correlations Between Predictor and Criterion Variables for Females and Males Predictor Variables Criterion Variables Home Peer School age Subcultural -.03 .50* -.37* Minor Property .13* .23* -.21* Major PrOperty -.05 .16* -.30* Minor Person .05 .11* -.l4* Major Person .09 .13* -.09 Females Subcultural —.08 .39* -.12 Minor Property -.16 .16 .02 Major Property .13 .15 -.21 Minor Person .08 .16 -.07 Major Person .08 .15 .05 *p (.05 45 Table 6 . Comparison of Hypothesized Regression Equations with Results Predictor Variables Positive Peer School Home Involve- Involve- _ Attendance ment (PHI) ment (PI) (SA) Criterion Variables pa A p A p A Subcultural Male 19 0‘3 2+ 1+ 3+ 2- Female 1- 0 2+ ‘ 1+ 3- a Minor Property Male 1- 0 2+ 1+ 3+ 2- Female 2- 9 1+ 8 3- g Major Property . Male ' l- 0 . 2+ 2+ 3- 1- Female 1- 9 2+ 0 3- 8 Minor Person . ” .Male 2- 9 1+ 0 3+ ° 8 Female 2- 0 1+ 0 3 a Major Person ' Male 1- fl 2- fl 3- g Female 1- fl 2- fl 3- fl aP = Predicted A = Actual b = The numbers indicate the order of entry of the variable in the equation; the signs indicate the direction of the relationship. C0 = No significant relationship 46 attendance the F-to-enters were significant, with the Beta- weights positive for peer involvement and negative for school attendance (see Table 7). This paralleled the signs of the relationships in the zero-order correlations (g = .50 and 0.37, respectively, see Table 5). In sum, for males, subcultural offenSes were not significantly influenced by positive home involvement, but were related positively to involvement with peers and related negatively to school attendance. These results did not support the hypothesis that the order of the predictor variables would be home, peer, and school, nor did they confirm the hypothesized directions of the relationships. Femalest It was hypothesized that for females, subcultural offenses would be predicted best by positive home involvement, somewhat by peer involvement, and least by school attendance, and that the relationships would be negative for PHI, positive for PI, and negative for SA. Results of the stepwise regression indicated that subcultural offenses as measured by the SRD Subcultural subscale were significantly predicted by the combined effects of the home, peer, and school variables, overall {(3,46) = 2.86, p (.05 (see Table 8). Sixteen percent of the variance in subcultural offenses was attributable to these variables (R:'=.16). The impact of individual predictors varied. Positive home involvement and school attendance were not individually significant. Peer 47 Table'7 Stepwise Regression of Predictor Variables on Offense Types Males (n=229) Mult. .9. R2 F-to-enter 2 Beta Subcultural Home .03 .00 .26 .61 .17 Peer .50 .25. 76.40 .00 .18 School .55 .30 15.95 .00 .19 Overall §_= 32.59, pp(.05 Minor Property Home .13 .02 3.72 .06 .17 Peer .27 .07 14-14 .00 .18 School .32 .10 7-53 .00 -.19 Overall F = 8.7, p (.059 Major Property Home .15 .00 .61 .44 .00 Peer.‘ .16 .03 5.48 .02 .07 School .31 .09 16.84 .00 -.28 Overall §_= 7.79, P (.05 Minor Person Peer .11 .01 2.75 .10 .07 Home .12 .12 .76 .38 .08 School .11 .03 3-76 .05 -.14 Overall 3 = 2.44, p).05 Major Person ' Home .09 .01 1-82 .18 .11 Peer .16 .03 3-92 .05 .11 School .17 .03 1-13 .28 -.08 Overall F = 2.31, E)-05 48 Table 8 Stepwise Regression of Predictor Variables on Offense Types Females (n=50) Offense Types Mult . R R2 F-to-enter 2 Beta Subcultural Home .08 .02 ~29 .60 .07 peer .40 .16 8-35 .01 .38 School .40 .16 ~07 .80 .04 Overall _F_‘ = 2.86, p(.05 Minor Property Home .16 .03 1.23 .27 .17 Peer .23 .05 1-31 .26 -.18 School .24 .06 .38 ' .54 .09 Overall F = .97, p).05 Major Property ‘ . Home .13 .02 -78 .38 .18 School .27 .08 3-01 .09 -.23 Peer ‘ .30 .09 . ~52 .44 .11 Overall F = 1.47, p) .05 Minor Person peer .16 .02 1.19 .28 .15 Home .17 .03 -29 .59 .09 School .18 .03 .19 .67 .07 OVerall F = .54, p') .05 Major Person Peer .15 .02 1.06 .31 .16 Home .17 .03 -32 .57 . .06 School .18 .03 -16 .69 .06 Overall F = .50, p) .05 49 involvement accounted for most of the variance, E-to-enter (1,46) = 8.35, p < .05, and the Beta was positive. In sum, for females, subcultural offenses were not impacted significantly by positive home involvement or school attendance, but were predicted by peer involvement. These results did not confirm the hypothesized importance of the predictor variables, though they did confirm the expected direction of relationship among the variables. Minor Property Offenses Males. It was hypothesized that for males, minor property offenses would be predicted similarly to the subcultural offenses, described above (see Table 6). Results of the stepwise regression indicated that minor property offenses were significantly predicted by the combined effects of the home, peer, and school variables, £(3,225) = 8.71, p < .05. Ten percent of the variance in minor property offenses was attributable to these three variables (R:‘=.10). Individually, the impact of positive home involvement was not significant, but both peer involvement and school attendance were predictors of minor property offenses for males, F’s-to-enter(l,225) = 14.12, p < .05 and 7.53, p < .05, respectively, with Betas positive for peers and negative for attendance (see Table 7). Again, the Betas paralleled the signs of the relationships in the zero-order correlations (.23 and -0.21, respectively, see Table 5). In sum, for males, minor property offenses were 50 not impacted significantly by positive home involvement, but were related positively to peer involvement and negatively to school attendance. These results did not confirm the hypothesis regarding the order of importance of the predictor variables, nor did they confirm the hypothesized direction of relationship for positive home involvement or school attendance. Females. It was hypothesized that for females, minor property offenses would be predicted best by peer involvement, somewhat by positive home involvement, and least by school attendance. The relationships were expected to be positive for PI, negative for PHI, and negative for SA (see Table 6). Results of the stepwise regression indicated that minor property offenses were not significantly related to any of the three predictor variables, nor were the effects of the variables cumulatively significant, overall F(3,46) = .97, p > .05. In sum, positive home involvement, peer involvement, and school attendance were not predictors of minor property offenses committed by the adolescent females in this study. Major Property Offenses Males. For males, it was hypothesized that major property offenses would be predicted best by positive home involvement, somewhat by peer involvement, and least by school attendance. The relationships were expected to be negative for PHI, positive for PI, and negative for 51 SA (Table 6). Results of the stepwise regression indicated that major property offenses were significantly predicted by the combined effects of the home, peer, and school variables, overall F(3,225) = 7.79, p < .05. Nine percent of the variance in major property offenses was attributable to these variables (Ra'=.09). The separate impact of positive home involvement was not significant; peer involvement and school attendance (including the negligible effect of positive home involvement) were both significant. The E-to-enter(l,225) for peer involvement was 5.48, p < .05; for attendance, the statistic was 16.84, p < 0.05. The Betas were positive for peer involvement and negative for school attendance (see Table 7). These relationships matched those found in the zero-order correlations (.16 and -.30, respectively, see Table 5). In sum, for males, major property offenses were not predicted by positive home involvement but were related positively to involvement with peers and negatively to school attendance. These results did not confirm the hypothesized order of importance among the predictor variables. They did confirm the predicted direction of relationships between the independent variables and major property offenses. Females. It was hypothesized that for females, major property offenses would be predicted best by positive home involvement, somewhat by school attendance, and least by peer involvement. The directions were expected to be 52 negative for PHI, negative for SA, and positive for PI (Table 6). Results of the stepwise regression indicated that major property offenses were not significantly related to any of the three predictor variables. The overall F for the equation was also not significant, F(3,46) = 1.47, p > 0.05. In sum, for female adolescents, major prOperty offenses were not predicted by positive involvement with family, involvement with peers, or regular school attendance. Minor Person Offenses Males. For males, minor person offenses were hypothesized to be predicted in the order of peer involvement, positive home involvement, and school attendance. The relationships were expected to be positive for PI, negative for PHI, and positive for SA (Table 6). The results of the stepwise regression indicated that minor person offenses were not significantly related to any of the three predictor variables. The overall F was also not significant, F(3,225) = 2.44, p > .05. In sum, for male adolescents, minor person offenses were not predictable by positive involvement with the home, involvement with peers, or by regular school attendance. Females. For females, minor person offenses were hypothesized to be predicted best by peer involvement, somewhat by positive home involvement, and least by school attendance. The relationships were predicted to be positive 53 for PI, negative for PHI, and for SA there was no prediction. Results of the stepwise regression indicated that minor person offenses were not significantly related to any of the three predictor variables. The overall F was also not significant, £(3,46) = .54, p > .05. In sum, for female adolescents, minor person offenses were not predicted by degree of positive home involvement, peer involvement, or school attendance. Major person offenses For both males and females, major person offenses were hypothesized to be predicted in the order of positive family involvement, peer involvement, and school attendance. The relationships were expected to be negative for all three variables. Results of the stepwise regression indicated that major person offenses were not significantly related to any of the three predictor variables. For both males and females the variables were unable to predict major person offenses. For males, the overall §(3,225) was 2.31, p > .05; for females, F(3,46) = 0.50, p > .05. Summary 9f Major Findings For males, three of the five offense types were found to be predictable from the three predictor variables. Subcultural, minor property, and major property offenses were predicted by the combined independent variables of peer and positive home involvement and school attendance. Peer involvement made a unique, significant contribution to 54 these predictions; the direction of the relationship was positive. After peer involvement was entered, the school attendance variable also made a unique and significant contribution; the direction of the relationship was negative. For females, subcultural delinquency was predicted by the combined effects of the three independent variables. In this equation, peer involvement made a unique and significant contribution; the direction of the relationship waS‘positive. For both males and females, person offenses were not related to the three hypothesized predictor variables, nor was there a significant relationship between the variables and minor or major property offenses for females. In no case was positive home involvement a significant predictor variable. With positive home involvement entered, peer involvement was a predictor of subcultural offenses for males and females, and property offenses for males. With both peer and positive home involvement entered, irregular school attendance predicted high subcultural and property offenses for males, but did not relate to any of the offenses reported by females. In sum, examining all the regression equations, of the three predictor variables, peer involvement most often had 55 the greatest impact, school attendance was of some importance, and positive home involvement, the variable predicted to be most significant, had a measured impact which, if present at all, was negligible. 56 Discussion The following section will include a discussion of the discrepancies between the anticipated and actual relative frequencies of offenses reported by male and female adolescents. Two aspects of the research methodology, subject selection and the validity of the measures, will be examined as possible contributors to these discrepancies. The major findings and major disconfirmations in the current research will be discussed and summarized, and implications and directions for future research will be considered. Frequency 9f Offenses It was expected that male and female adolescent offenders would show similar self-reported offense rates for subcultural, minor property, and minor person offenses, and that males would exceed females in the major property and major person categories. These expectations were not confirmed. The results of this investigation showed no significant differences between between males and females in the frequencies of any self-reported person and property offenses, while the anticipated gender similarity in subcultural offenses was absent, that is, females reported more subcultural offenses than males. In attempting to understand the differences between the hypothesized and the actual frequencies, two aspects of the research methodology must be considered; the subjects used in the study and the validity of the research measures. 57 Methodology Subjects. In considering the possible impact of the specific sample population on the results of this study, two points are important; the first concerns the total sample, the second concerns the male/female differences within the sample. The subjects in this study represented a sample of juvenile offenders petitioned by the court rather than a sample which had been randomly selected from a population of adolescents in general. Youth with arrest records may not provide similar data to youth tapped in general population samples. Among adolescents with ’official status’ (an official record of delinquent behavior), Hindelang et a1. (1981) found a significantly higher rate of self-reported delinquency, and Elliott et a1. (1983) found that in an official population serious delinquent activity was underreported by one out of five youths. According to several randomly sampled studies, males tend to exceed females in both severity and frequency of offenses (Feyerherm, 1981; Steffensmeier & Steffensmeier, 1980; Wright, 1983), but Canter (1982b), reporting on an offender population, found no differences between males and females on self-reported measures of theft and crimes against persons. The subjects in the current study may have more closely resembled Canter’s offenders than the general adolescent population. Canter’s findings, therefore, may 58 support those of the present investigation. The expected male/female differences in the current study may have been minimized by male arrestee underreporting of serious offenses. Additionally, a lack of clarity in specific research items may have elicited overreporting of trivial offenses by both males and females, a possibility which will be discussed further in the following section. In sum, both males and females may have reported higher frequencies of offenses than would be found in the general population, but if serious offenses were underreported by males and trivial offenses were overreported by females, the male/female ratios might have been lowered as a result. Other research reporting data from offenders (Canter, 1982b) may support this finding. The results of this study, therefore, may differ from those derived from general adolescent populations both because of the particular responses elicited by the nature of the items and because of certain characteristics of the sample population. Measures. As discussed earlier, the use of self-report scales has been criticized because of the difficulty in accurately assessing offense frequencies (Elliott et a1., 1983), and because of the tendency to trivialize offenses (Canter, 1982a, 1982b; Hood & Sparks, 1970). The Adolescent Diversion Project SRD scale as a whole, and the subscales in particular, may be subject to these criticisms. Although 59 the reliability of the scale as a whole had been tested (Parisian, 1982), the subscales were constructed rationally, to categorize offenses ranging from mildly to seriously delinquent. Because the maximum frequency reportable was "three or more times", any high-frequency offenses may have been obscured. As mentioned above, males generally tend to exceed females in both severity and frequency of offenses (Feyerherm, 1981; Steffensmeier & Steffensmeier, 1980; Wright, 1983). It is possible that the forced limit in this SRD measure hid a gender difference in offense frequencies. As discussed by Hindelang et al. (1981), increased specificity in scale items tends to increase similarity to official measures and to decrease trivial responses. In the current subscales, item generality may have been a particular problem in the description of serious offenses. Validity studies of SRD data have indicated that frequently juveniles report events'which would not be considered seriously delinquent or worthy of intervention by much of the general public. Elliott et al. (1983) found that while serious offenses were underreported by 20 percent, trivial items tended to be overreported. It is possible that some of the items in the current SRD subscales could have been answered in this way (for example, threatening to hurt someone is a minor person offense for which a juvenile might report much frequency but which could be of little or no severity.) Hindelang et a1. (1981) discovered that changing 60 items to specify offenses related to strangers rather than family (for example, "have you taken a car belonging to someone you didn’t know without the owner’s permission" as contrasted with "have you taken a car without the owner’s permission") elicited different responses and increased the self-reported male/female ratios, rendering them more similar to official data. It is possible that the lack of such specificity may account for the lack of significant male/female differences in the current study. It is a common criticism of self-reportdelinquency studies that the most serious offenses recorded as official delinquency are rarely included in self-report measures because of their anticipated infrequency (Canter, 1982b; Elliott et a1. 1983; Sarri, 1983). Offenses which are considered serious in most SRD questionnaires, then, are relatively more common events. An example of this type of item might be a question from the ADP SRD questionnaire asking how often the respondent has "taken something from a person by force (with or without a weapon)", the answer to which could include a range of behaviors from wrenching an object out of another’s hands to attacking with a knife or gun. Since there is not a specific question concerning the use of guns or other weapons, it is not possible to identify the most serious offenses, or to separate them from other relatively trivial events. As another example of the lack of differentiation among degrees of offense severity, 61 Hindelang et a1. (1981) cite Rossi’s 1974 study which ranked 140 items for seriousness and found "repeated running away from home", a regular item in SRD scales, ranked 137th. Running away has been considered an offense serious enough to investigate in SRD questionnaires, and indeed, it is an item which may seriously impact the results, a fact which will be discussed shortly. It does not, however, carry the weight of, for instance, homicide or forcible rape, offenses more typically committed by males, but not usually included in SRD scales because of the infrequency of responses. In sum, the lack of extremely serious items in the present research may contribute to the lack of significant gender differences in the major person and property offenses. In the current study, it appears that the item concerning running away could have had a significant effect on the findings regarding subcultural offenses. The greater proportion of subcultural offenses reported by females could have been attributable to the one item in the subscale which solicits information on running away. Research has suggested that a large proportion of young females who come to the attention of law enforcement agencies do so for running away (Chesney-Lind, 1978; Strouse, 1978). According to Hindelang et al. (1981), in the 1976 official arrest data "runaway" accounted for 23 percent of female and less than 5 percent of male arrests. This statistic raises the possibility that similar percentages of the subjects in the 62 current study might have been referred to the court for this particular offense, which alone could have determined the significantly greater frequency of subcultural offenses reported by females in this study. In sum, within the present study, the lack of gender differences may be attributable both to the particular composition of the sample, that is, offenders as opposed to general population, and to the general content of the measures employed, which Hindelang et a1. (1981) refer to as "the swamping effect of high-frequency low-seriousness offenses." Major Findiggs Male subcultural, minor property, and major property offenses. Three of the hypotheses predicted that for males, subcultural, minor property, and major property offenses would be impacted by positive home involvement (PHI), peer involvement (PI), and school attendance (SA), in that order. In all cases, the relationships were predicted to be negative for PHI and positive for PI. SA was expected to be related positively to subcultural and minor property and negatively to major property offenses. (See Table 6 for a summary of hypothesized and actual regressions). The bases for these hypotheses included studies citing the significance of family involvement in male delinquency (Simons et a1., 1980; Klemke, 1982; Canter, 1982a, Feyerherm, 1981); studies citing the relevance of peers in 63 male delinquency (Figueira-McDonough et a1. 1981, Klemke, 1982; Aultman, 1980) and studies regarding male delinquency and school attendance (reviewed in Phillips & Kelly, 1979). Results of this investigation confirmed that the combined variables were indeed statistically significant predictors for males for all three offense types. The results also, however, indicated that the three predictors varied considerably in their power, and that in fact positive home involvement was not, alone, a good predictor of any type of offense. The significance level of the results of these regressions depended largely on the effects of peer involvement, and (for major property offenses) school attendance. The two variables, peer involvement and school attendance, were related. Simply stated, it may be that boys who ’hang out’ with friends and attend school irregularly were at risk for subcultural and property offenses regardless of their relationships with their families. For major property offenses, irregular school attendance was an even better predictor of illegal activity than involvement with peers. The hypotheses regarding subcultural and minor property offenses had predicted, on the basis of several studies included in Phillips and Kelly’s 1979 review of the literature, that the direction of relationship between school attendance and delinquent behaviors would be positive. These theories were based on the idea that boys who were not attending school were 64 experiencing no dissatisfaction with school or frustration from school failure, and would therefore display fewer delinquent behaviors (Phillips & Kelly, 1979; Elliott, 1966, cited in Phillips & Kelly, 1979; Gold & Petronio, 1980). Boys who did attend school were predicted to be more delinquent. The results of the present investigation indicated that boys who attended school irregularly were more likely to commit subcultural, minor property, and especially major property offenses. It is of interest in examining these results that peer involvement and positive family involvement were not mutually exclusive categories; one did not increase as the other decreased. For subjects in this study, involvement with family was simply unrelated to involvement with peers, which was the important social context. This finding is supported by Bowker and Klein’s (1983) study using data from ’official’ offenders, which also found family involvement to be unrelated to delinquency. Of other studies cited earlier in this paper, Figueira-McDonough et a1. (1981) and Sarri (1983) found family involvement to be of only modest significance in determining delinquent offenses. Several studies have suggested that family involvement might have deterred delinquent behavior, in other words, a positive family involvement might be related to an absence of delinquency (James & Thornton, 1980; 6S Klemke, 1982). In the present study it may be the case that although positive family involvement is a deterrent for non- offenders, for offenders the family plays a different role. Female subcultural offenses. For females, the expected order of predictor variables affecting subcultural delinquency was similar to the prediction for males, with only a difference in the expected direction of relationship to school attendance. In other words, the subculturally delinquent female was not expected to have a strong positive relationship with her family, but was expected to have a positive relationship with her peers, and an irregular record of school attendance (see Table 6). The hypothesis regarding family involvement was founded on the research and theories of Morash (1983), Simons et a1. (1980) and Sarri (1983), which suggest that family involvement is an inhibitor of delinquency. Canter (1982a), Figueira- McDonough et al. (1981), and Bowker and Klein (1983), however, had found less significant impact for family variables on self-reported delinquency. Indeed, as discussed above, the Bowker and Klein study, which used official offenders as the data base, found no relationship between family variables and delinquency. Female subjects in the current study did not, in any offense type, report a level of positive or negative family involvement which correlated significantly with delinquent offenses. Peer involvement, on the other hand, which was hypothesized to rank second 66 among the contributions of the three variables, was the major and only significant contributor to subcultural delinquency. These results confirmed those of Bowker and Klein (1983), who reported the impact of peers to be far greater than that of adults. School attendance, which had been predicted to be negatively related to delinquency for females, was not related at all. It seems that, like positive family involvement, regular school attendance may function as a deterrent to those females who are not classified as offenders. For the females in this sample, however, other unidentified variables may have intervened. School attendance had no significant effect on any of the offense categories examined in the current study. This issue will be discussed in a later section. Major Disconfirmations In the above discussion of hypotheses and results, the combined predictor variables of peer and positive home involvement and school attendance were seen to account for a statistically significant portion of the variance in subcultural offenses for males and females and in male minor and major property offenses as well. In addition to those four offense categories, however, there remain the female property offenses and both male and female person offenses. In the current study, there were no significant relationships between these offense types and any of the 67 three predictor variables. Female property offenses. For females, it had been predicted that minor property offenses would be affected by peer involvement, lack of positive home involvement, and school attendance, in that order, and that the relationships would be positive for PI and negative for PHI and SA (see Table 6). This hypothesis was based on research suggesting, for females, that peers are of primary importance to delinquency (Figueira-McDonough et a1., 1981); that positive family involvement would inhibit and negative family involvement might foster delinquency (Klemke, 1982; James & Thornton, 1980); and that involvement in school might deter while lack of involvement might encourage delinquent behaviors (Sarri, 1983). Major female property offenses were expected to be rare, and it was anticipated that a lack of positive involvement with families would surpass peer involvement as a predictor of these behaviors (Canter, 1982a, James & Thornton, 1980). These hypotheses were not confirmed. The three combined variables accounted for less than six percent of the variance in minor property offenses and less than nine percent of the variance in the major offenses. It is clear that, despite the research literature which has determined that these variables do affect the delinquent behaviors of the general population of adolescents, the sample of female offenders in the present study was not affected by variables in the forms in which 68 they were considered in this investigation. Male and female person offenses. Considering the minor person offenses, it was hypothesized for both males and females that peer involvement would be primary and positive, home involvement would be secondary and negative, and school attendance would be third in the order of importance of the variables. For females there was no prediction of the direction of relationship for school attendance; for males the relationship was predicted to be positive, because of increased opportunities for illegal activity available at school. As with the hypotheses discussed above, it was the case that none of the three predictor variables was related to either minor or major person offenses for males or females. Indeed, the combined variables accounted for only three percent of the variance in all of these cases. Summary 9f Major Findings In general, it may be stated that in this study peer involvement was related to boys’ and girls’ subcultural activities, and also boys’ property offenses, though irregular school attendance was a stronger predictor of boys’ serious property offenses. Irregular school attendance also affected, to a lesser extent, male minor property and subcultural offenses. For both male and female offenders in this study, degree of family involvement was unrelated to all types of offenses. Positive family involvement was not, for males or females, an inhibitor of 69 delinquent behavior. For the female juvenile offenders in this study, delinquent behaviors were also unrelated to school attendance. Like family involvement, school attendance was not, for these girls, an inhibitor of delinquent behavior. Implications for Further Research Although the results of the current study provide some confirmation for the research hypotheses, some important questions remain unanswered. These questions may be grouped into three categories; they concern the specific sample population, the frequencies of offense types, and the implications of the general research findings. Subjects. It seems to be a significant aspect of the present study that the subjects were court-referred offenders. A consistent explanation for findings which differed from those hypothesized is that offenders are different in some significant ways from the general population of adolescents, on whom SRD findings are usually based. One important implication from these findings is that information regarding variables which predict delinquent offenses may best be obtained from offenders themselves. To determine appropriate preventive interventions, there is considerable justification for further investigation of the differences between these two groups (’official’ offenders and the general adolescent population). 70 Frequencies 9f offense types. The results of the present study indicate that while male offenders outnumber female offenders in the general population, male and female ’official’ offenders are not significantly different in the types of person and property offenses they report, including major offenses. Although this finding may be supported by other studies (Canter, 1982b), it is possible that the exclusion of extremely serious items and the generality in the included items may account for these results. With a more specific offense measure, the offense categories might show greater discriminant ability. Further research which provided greater item specificity and a broader range-of frequency responses could provide further information regarding the gender differences in offense frequencies. General research findings. Although there was some confirmation of the research hypotheses, the three predictor variables, positive home involvement, peer involvement, and school attendance were generally not related to the behaviors of these offenders in ways which were predicted based on their reported impacts on adolescents in general. For male offenders, the importance of peer involvements and the interactions of peer involvement and irregular school attendance, combined with the lack of importance of positive family involvement, suggest that these boys, as opposed to adolescents in general, have in some way transferred their allegiance from ’home base’ to a non- .71 academic ’street’ world. It appears that for these boys, the family situation need not be negative for delinquent behaviors to appear. Further research in this area should investigate the values of both the peers and families of male offenders, in an attempt to understand why involvement with peers becomes paramount, and why school attendance becomes irregular. Perhaps court-referred offenders experience an increased need to identify with peers who have shared similar experiences. The current study did not include information regarding negative family involvement, that is, family conflicts. Future studies should also investigate whether family dissension impacts on delinquent behaviors differently for different offense types, and whether this impact is similar or different for males and females. If it is indeed the case that adolescent offenders are involved with their families differently from adolescents in general (thus positive family involvement does not deter delinquent behavior), it is important to know in what ways their family involvements differ. Considering the impact of peers on adolescent offenders, it is interesting to note that Aultman’s (1980) finding that peer involvements did not affect person offenses was confirmed. For males, as discussed above, peer involvements were important determinants of subcultural and property offenses. For females, the lack of significance of 72 peers in property offenses as well as person offenses seems important. While the results regarding males suggest that relationship to a peer group is a significant aspect of their behaviors, for the female offenders the peer group, like the family, shows no impact. The results of this study suggest that these girls are different from the general population, where peers and family are both seen as powerful influences. Because school attendance also shows no impact on the behavior of these females, the image projected by these data is that of a loner, whose behavior is not motivated by positive involvements with family, peers, or school. Female ’offenders’ represent a much smaller group than do the males; perhaps an offender peer group is more difficult to find. Based on the results of the present research, it could be concluded that female adolescent offenders are different from both male offenders and adolescents in general. Further research with this subject population should include motivational and personality investigations, to see whether there are essential offender/non-offender differences. Research should also include social histories, to determine at what points and in what ways these girls’ interactions with their social contexts ceased to restrict their behaviors to socially appropriate acts. . The final finding which seems to have some potential practical significance concerns the predictor variable of 73 school attendance, which was not an inhibitor of delinquent behavior for females, and which, when irregular, was a predictor of subcultural and property offenses for males. Future research concerning the impact of school on this type of pOpulation should include items regarding academic success and failure, and should also ask which of the reported offenses occurred at school. With this information it would be possible to pinpoint more closely which aspects of this significant component of adolescent life might be important in planning preventive interventions. Conclusion. The current study provides some significant information regarding the uses of positive home involvement, peer involvement, and school attendance as predictor variables in examining types of offenses reported by male and female offenders. It also suggests that, among male and female offenders, there are few significant differences in self-reported person and property offenses, as measured by subscales of the ADP SRD questionnaire, and discusses some of the possible methodological issues with the questionnaire. Finally, it raises the issue of the differences between ’official’ offenders and random samples of adolescents in the study of delinquency. To understand and effectively intervene in delinquent adolescent behaviors, further research should acknowledge and investigate the differences between the social environments experienced by these two reference groups. LIST OF REFERENCES 74 List of References Adler, F. (1975). Sisters in crime: The rise of the new female criminal. New York: McGraw-Hill Book Company. Adler, F., & Simon, R.J. (Eds.) (1979). The criminology 9f deviant women. Boston: Houghton Mifflin Co. Aultman, M. (1980). Group involvement in delinquent acts: A study of offense types and male-female participation. Criminal Justice and Behavior, 1, 185-192. Bowker, L. H., & Klein, M. W. (1983). The etiology of female juvenile delinquency and gang membership: a test of psychological and social structural explanations. Adolescence, 18, 739-751. Canter, R.J. (1982a). Family correlates of male and female delinquency. Criminology, 29, 149-167. Canter, R.J. (1982b). Sex differences in self-report delinquency. Criminology, 29, 373-393. Chesney-Lind, M. (1978). Judicial paternalism and the female status offender: Training women to know their place. In B. Krisberg and J. Austin (Eds.), Egg children pf Ishmael: Critical perspectives 93 juvenile justice (pp.376-387). Palo Alto, CA: Mayfield. Chesney-Lind, M. (1982). From benign neglect t9 maliqn attention: a critical review 9; recent research 9g female delinquency. (Report No. 275). Manoa: University of Hawaii-Manoa, Youth Development and Research Center. Dunford, F.W., & Elliott, D.S. (1984). Identifying career offenders using self-reported data. Journal 9f Research i3 Crime and Delinquency, 21(1), 57-86. Elliott, D.C., Ageton, S.S., Huizinga, D., Knowles, 8., & Canter, R. (1983). The prevalence and incidence 9f delinqgent behavior: 1976-1980 (National Youth Survey Project Report No.26). Boulder: Behavioral Research Institute. Elliott, D.S., & Huizinga, D. (1983). Social class and delinquent behavior in a national youth panel. Criminology, 21, 149-177. 75 Feyerherm, W. (1981). Measuring gender differences in delinquency: Self-reports versus police contact. In M. Warren (Ed.), Comparing female and male offenders (pp.46-54). Beverly Hills, CA: Sage Publications. Figueira-McDonough, J., Barton, W., & Sarri, R. (1981). Normal deviance: Gender Similarities in adolescent subcultures. In M. Warren (Ed.), Comparing female and male offenders (pp. 17-45). Beverly Hills, CA: Sage Publications. Giordano, H.C. (1978). Girls, guys, and gangs: The changing context of female delinquency. Journal 2: Criminal Law and Criminology, 69, 126-232. Gold, M. & Petronio, R. (1980). Delinquent behavior in adolescence. In J. Adelson (Ed.), Handbook 2; adolescent psychology (pp.495-535). New York: John Wiley & Sons. Hindelang, M., Hirschi, T., & Weis, J. (1981). Measuring delinquency. Beverly Hills, CA: Sage Publications. Hirschi, H. (1969). Causes pf delinquency. Berkeley: University of California Press. Hood, R. & Sparks, R. (1970). Key issues i2 criminology. New York: McGraw-Hill. James, J., & Thornton, W. (1980). Women’s liberation and the female delinquent. Journal 9f Research 1Q Crime and Delinquency, 11, 230-244. Klemke, L.W. (1982). Exploring juvenile shoplifting. Sociology and Social Research, 61, 59-75. Krisberg, B., & Austin, J. (Eds.) (1978) The Children 2: Ishmael: Critical perspectives 93 juvenile justice. Palo Alto, CA: Mayfield. Morash, M. (1983). An explanation of juvenile delinquency: the integration of moral-reasoning theory and sociological knowledge. In W. Laufer and J. Day (Eds.), Personality theory, moral development, and criminal behavior (pp.385-408). Lexington, MA: Lexington Books. 76 Parisian, J.A. (1982). 99 examination 92 the processes 92 change: Exploring the success 92 99 effective juvenile diversion project. Unpublished master’s thesis, Michigan State University, East Lansing. Phillips, J.C., & Kelly, D. H. (1979). School failure and delinquency: Which causes which? Criminology, 21, 194- 207. Sarri, R.C. (1983). Gender issues in juvenile justice. Crime and Delinquency, 29, 381-397. Selke, W. L. (1982). Diversion and crime prevention. Criminology, 29, 395-406. Simons, R.L., Miller, M.G., & Aigner, S.M. (1980). Contemporary theories of deviance and female delinquency: An empirical test. Journal 92 Research 29 Crime and Delinquency, 21, 42-53. Steffensmeier, D., & Steffensmeier, R.H. (1980). Trends in female delinquency: An examination of arrest, juvenile court, self-report, and field data. Criminology, 22, 62-85. Strouse, J. (1978). To be minor and female: The legal rights of women under 21. In B. Krisberg & J. Austin (Eds.), The children 92 Ishmael: Critical perspectives 99 juvenile justice (pp.368-376). Palo Alto, CA: Mayfield. Williams, J.R. & Gold, M. (1978). From delinquent behavior to official delinquency. In B. Krisberg and J. Austin (Eds.), The children 92 Ishmael: Critical perspectives 99 juvenile justice (Pp- 308-330). Palo Alto, CA: Mayfield. Wright, S.K. (1983). The relationship 92 gender 29 qpalitative and quantitative aspects 92 self-reported delinquenqy. Unpublished master’s thesis, Michigan State University, East Lansing. APPENDICES APPENDIX A 1015. 1016. 1017. 1018. 1019. 77 Life Domain Survey Positive Home Involvement (all questions refer to the last six weeks) How often does youth spend time with parents in athletics? l 2 3 4 ‘5 Never Once Twice 3 or 4 More than times 4 times How often do the youth and parent(s) go to movies together? 1 “2 3 4 5 Never Once Twice 3 or 4 More than times 4 times How often do the youth and parent(s) go camping/fishing/hunting, etc.? 1 2 3 4 5 Never Once Twice 3 or 4 More than times 4 times How often does youth visit relatives with parents? 1 ‘2 3 4 5 Never Once Twice 3 or 4 More than times 4 times How often does the parent(s) instruct the youth in some skill/activity? l \2 3 4 5 Never Once Twice 3 or 4 More than times 4 times 1020. 1021. 1022. 1023. 1024. 1025. 1026. 78 How often does the youth participate in purchased activities with parents? 1 2 3 4 5 Never Once Twice 3 or 4 More than times 4 times How often do the parent(s) talk with the youth about day- to-day things? 1 2 3 4 5 Never Once a Once a More than Daily month week once/week How often does the youth Spend time with siblings in athletics? *(no siblings) l 2 3 4 5 Never Once Twice 3-4 More than times 4 times How often does the youth spend time with siblings going to movies? *(no siblings) 1 2 3 4 5 Never Once Twice 3-4 More than times 4 times How often does the youth spend time with siblings camping/fishing/hunting, etc.? *(no siblings) l 2 3 4 5 Never Once Twice 3-4 More than times 4 times How often does the youth spend time with siblings going out of town? *(no siblings) l 2 3 4 5 Never Once Twice 3-4 More than times 4 times How often does the youth spend time with siblings at indoor activities (TV)? *(no siblings) l 2 3 4 ' 5 Never Once Twice 3-4 More than times 4 times 79 1027. How often does the source say the youth and siblings "hang around together? (*no siblings) l 2 3 4 5 Never Once 2-6 More than Daily times once/week 1028. How much is expected of youth in terms of household responsibilities? 1 2 3 4 5 Nothing Very Some A fair A lot little amount 1029. How often does the youth complete his/her household responsibilities? (*no responsibilities) 1 2 3 4 5 Never Seldom Half the Most of All time the time the time 1030. How often does the youth spend evenings at home? 1 2 3 4 5 Never Less than More than Almost Daily once/week once/week everyday 1031. How often does the youth engage in other spontaneous activities with his/her parents? 1 2 3 4 5 Never Once 2-6 More than Daily times once/week 1032. How often does the youth engage in other spontaneous activities with his/her siblings? (activities not covered by other items) (*no siblings) l 2 3 4 5 Never Once 2-6 More than Daily times once/week 1033. 1034. 80 How often does the youth engage in other purchased activities with siblings? (*no siblings) l 2 3 4 5 Never Once Twice 3-4 More than times 4 times How often does the youth sleep at home at night? 1 2 3 4 5 Never Less than More than Almost Daily once/week once/week daily 1114. 1115. 1116. 1117. 1118. 81 Life Domain Survey Peer Involvement (all questions refer to the last six weeks) How often does youth spend time with friends during school time? (*not in school last 6 weeks) 1 2 3 4 5 Never 2 times Almost Few times Major part a week everyday a day of the day How often does youth skip school with friends? (*not in school last 6 weeks) 1 2 3 4 5 Daily More than 2-6 Once Never once/week times How often does youth spend time with friends on weekends? I 2 3 4 5 Never Weekend Part of Part of Most of per month most every every every or less weekend weekend weekend How often does youth participate in purchased activities with friends? 1 2 3 4 5 Never Once 2-6 More than Daily times once/week How often does youth participate in other spontaneous activities with friends? 1 2 3 4 5 Never Once/week More than Almost ' Daily or less once/week daily 1119. How often does youth 1120. 1121. 1122. 1123. 1124. 1125. afternoons? l 2 Never Once/week or less How often does youth 1 2 Never Once How often does youth 1 2 Never Once How often does youth 1 2 Never Once How often does youth 1 2 Never Not very often How often does youth 1 2 Never Once 82 spend time 3 More than once/week spend time 3 2-6 times drink with 3 2-6 times with friends in the 4 Almost daily 5 Every afternoon with friends evenings? 4 More than once/week friends? 4 More than once/week 5 Every evening 5 Daily go to parties with friends? 3 Twice 4 3-4 times 5 More than 4 times smoke dope with friends? 3 Sometimes 4 Most of the time 5 All the time spend time at a friend’s home? 3 2-6 times 4 More than once/week 5 Daily How many close friends does youth associate with? 1 None (no set of friends) 2 One 3 Two 4 Three 5 More than three APPENDIX B 83 Appendix B Self-Report Delinquency (SRD) In the last year, how often have you: 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 Skipped class when you were in school? Gone onto someone's land when they didn't want you to be there, or without their permission? Gone into a house or building when you weren't supposed to be there? Played on a school atheltic team? Threatened to hurt someone? Been told to bring your parents to school for something you did wrong? Damaged or messed up something not belonging to you? Hurt someone badly enough for him/her to need bandages or a doctor? ' Gotten on the honor roll for good grades in school? Taken some part of a car or some gasoline? Hit a member of your family (in anger)? Have not been allowed to go to school until the superintendant or principal told you that you could go again? Taken something not belonging to you worth less than $2.00? Earned some money at a jOb? Drunk beer or liquor (includes sips)? Run away from home? Skipped a full day of school? Been sent to the school principal's office for bad behavior in class? Carried a gun or a knife? 84 (In the last year, how often have you ;.. ) 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 l) werked on a school newspaper or yearbook? Taken something not belonging to you worth over $50.00? Done something around the house or for the family that really pleased your family? Set fire to someone else's property? Used or threatened to use a weapon to get something from a person? Taken something from a store without paying for it (regardless of price)? Smoked without your parents knowing about it or without permission (regular cigs.)? Worked free for a charity organization? Taken a car without the owner's permission? (includes joyriding) Smoked marijuana? Taken something from a person by force? (with or without a weapon) Beaten up on somebody or fought someone physically? Taken drugs or pills, other than marijuana? Bought or gotten something that was stolen by someone else? Broken into a place and stolen something? Taken things worth less than $50.00 (over $2.00)? Never 2) Once 3) Twice 4) More than twice