<%BANNER%>

Record for a UF thesis. Title & abstract won't display until thesis is accessible after 2013-04-30.

Permanent Link: http://ufdc.ufl.edu/UFE0043029/00001

Material Information

Title: Record for a UF thesis. Title & abstract won't display until thesis is accessible after 2013-04-30.
Physical Description: Book
Language: english
Creator: TOLSON,DANIELLE
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: Sociology and Criminology & Law -- Dissertations, Academic -- UF
Genre: Criminology, Law, and Society thesis, M.A.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Statement of Responsibility: by DANIELLE TOLSON.
Thesis: Thesis (M.A.)--University of Florida, 2011.
Local: Adviser: Lanza-Kaduce, Lonn M.
Electronic Access: INACCESSIBLE UNTIL 2013-04-30

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0043029:00001

Permanent Link: http://ufdc.ufl.edu/UFE0043029/00001

Material Information

Title: Record for a UF thesis. Title & abstract won't display until thesis is accessible after 2013-04-30.
Physical Description: Book
Language: english
Creator: TOLSON,DANIELLE
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: Sociology and Criminology & Law -- Dissertations, Academic -- UF
Genre: Criminology, Law, and Society thesis, M.A.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Statement of Responsibility: by DANIELLE TOLSON.
Thesis: Thesis (M.A.)--University of Florida, 2011.
Local: Adviser: Lanza-Kaduce, Lonn M.
Electronic Access: INACCESSIBLE UNTIL 2013-04-30

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0043029:00001


This item has the following downloads:


Full Text

PAGE 1

1 MARIJUANA USE AND DIVERSITY IN DELINQUENCY By DANIELLE A. TOLSON A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFIL LMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ARTS UNIVERSITY OF FLORIDA 2011

PAGE 2

2 2011 Danielle Alexandria Tolson

PAGE 3

3 To my entire family, especially my mother, father, and sister, for their endless support, constant encourag ement, and honest uplift throughout this journey

PAGE 4

4 ACKNOWLEDGMENTS I would first like to acknowledge the members of my family. I thank my mother, Dr. Stephanie Tolson, for showing me how possible the impossible could be. I thank my father, Fred Tolson, for blessing me with his good humor and humble spirit that has allowed me to roll with the punches of graduate school. I thank my sister and brother in law, Tiffany and Ade Henry, for strengthening me with constant their support of my goals. I thank my nie ce, Bryanna, for being an inspiration to push forward and strive to be better. I would also like to acknowledge my MA committee members. I thank my chair, Dr. Lonn Lanza Kaduce, for working endless hours with me to prepare the best piece possible, I thank my co chair, Dr. Jodi Lane Wilson, for her open and honest advice that has helped make grad school life feel more manageable, and I think my committee member, Dr. Ronald Akers, for always lending an ear and offering suggestions that make me a better resea rcher and criminologist Last, but not least, I would like to thank my many colleagues and friends for their ongoing support. I thank my roommate, Taccara Williams, for being with me on this journey toward the MA, I thank member of my various cohorts, Jenn ifer Klein, Kathy Zambrana, Joe Ruckus, Cathy Collins, June Carrington, Lindsey Lauderdale and Adrianne Gilbert and I thank my extended family, LaTanya Mackey, LaShonda Mackey, Chanee Anderson, Joslyn Richardson, Jasmine Pampkin, Chant Jackson, Monica Gl eason, Marquitta Harmon and Do nell Copeland

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 ABSTRACT ................................ ................................ ................................ ..................... 8 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ ...... 9 2 SOCIAL STRUCTURE SOCIAL LEARNING THEORY ................................ .......... 11 Social Learning Theory ................................ ................................ ........................... 11 Social Structure Social Learning Theory ................................ ................................ 15 Testing of Social Learning Th eory ................................ ................................ .......... 18 Testing of SS SL Theory ................................ ................................ ......................... 21 Research on Juvenile Drug Use ................................ ................................ ............. 24 Research on Juvenile Drug Use and Delinquency ................................ .................. 29 3 THE CURRENT RESEARCH ................................ ................................ ................. 31 4 METHODS ................................ ................................ ................................ .............. 33 Participants and Design ................................ ................................ .......................... 33 Study Variables ................................ ................................ ................................ ....... 34 Dependent Variables ................................ ................................ ........................ 34 Marijuana use ................................ ................................ ............................ 34 Diversity of delinquency ................................ ................................ ............. 35 Independent Variables ................................ ................................ ..................... 36 Social structure variables ................................ ................................ ........... 36 Social learning process variables ................................ ............................... 38 Procedure and Analysis ................................ ................................ .................... 43 Missing Values ................................ ................................ ................................ 45 5 RESULTS ................................ ................................ ................................ ............... 50 Univariate Analysis ................................ ................................ ................................ 50 Bivariate Analysis ................................ ................................ ................................ ... 50 Multivariate Regression Analysis ................................ ................................ ............ 50 Predicting M arijuana Use ................................ ................................ ................. 51 Predicting Diversity of Delinquency ................................ ................................ .. 5 8

PAGE 6

6 6 DISCUSSION AND CONCLUSIONS ................................ ................................ ...... 69 Marijuana Use Models ................................ ................................ ............................ 69 Diversity of Delinquency Models ................................ ................................ ............. 71 Overview ................................ ................................ ................................ ................. 72 Research Limitations ................................ ................................ .............................. 74 Future Research ................................ ................................ ................................ ..... 74 APPENDIX: VARIABLE CONSTRUCTION ................................ ................................ ... 77 LIST OF REFERENCES ................................ ................................ ............................... 82 BIOGRAPHICAL SKETCH ................................ ................................ ............................ 88

PAGE 7

7 LIST OF TABLES Table page 4 1 Dependent Variable: Marijuana Use ................................ ................................ ... 47 4 2 Dependent Variable: Diversity of Delinquency Index ................................ .......... 47 4 3 Model 1: Can social structures explain juvenile marijuana use and diversity of delinquency? ................................ ................................ ................................ ...... 48 4 4 Model 2: Does social learning significantly predict marijuana use and diversity of del inquency? ................................ ................................ .................... 48 4 5 Model 3: Will the impact of the family structure variables be mediated by the social learning variables? ................................ ................................ ................... 49 5 1 Univariate Analysis for All Variables ................................ ................................ ... 64 5 2 ........... 65 5 3 Biva .... 66 5 4 Multivariate OLS Regression Predicting Juvenile Marijuana Use ....................... 67 5 5 Multivariate OLS Regression Predicting Juvenile Delinquency .......................... 68

PAGE 8

8 Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirement s for the Degree of Master of Arts FAMILIA L STRUCTURAL DIMENSIONS AND SOCIAL LEARNING OF MARIJUANA USE AND DIVERSITY OF DELINQUENCY By Danielle A. Tolson May 2011 Chair: Lonn Lanza Kaduce Major: Criminology, Law and Society This research exa mined the effects social structures and social learning processes have on juvenile marijuana use and diversity of delinquency. P revious research predicts that social processes will lessen the effect (mediate) that social structures have on juve nile drug us e and delinquency. T he researcher analyzed interview data from 146 juveniles housed in FCBDTI residential programs. The researchers found that social learning processes did not substantially mediate the relationship between social structure variables and j uvenile marijuana use and diversity of delinquency Although the broken home indicator variable was mediated it was the only one that did for either of the dependent variables and very few of the other structures significantly predicted juvenile marijuana use or diversity of delinquency. Differentia l association with peers who engage in deviant activities was the greatest predictor for both DVs, but imitation/modeling for parents drug use was significant in the marijuana models, suggesting that parents may have an effect on specific, instead of general, juvenile deviant behavior

PAGE 9

9 CHAPTER 1 INTRODUCTION Concern about juvenile behavior has been present at least since the movement to establish the first juvenile court in 1899 ( Platt 1977 ; Greenwood 1996 ; Sprott 1996 ) The transformation th at lead to the juvenile justice system was brought on by public recognition that more work needed to be done to help wayward youth and by people, to invest time and energy to see that change come to fruition ( Platt 1977 ) Even with the growth and changes tha t have occurred in the juvenile justice system, p eople still yearn to understand more about the deviant behaviors of juveniles ; the public is especially concerned with juvenile delinquency and drug use and how family and friends influence that behavior ( Triplett and Payne 2004 ; Brook, Brook et al. 2006 ) The s not a new phenomenon and many American crim inologists have sought to help with understanding juveniles through their research ( Johnston, O'Malley et al. 2004 ; Bahr, Hoffmann et al. 2005 ) Theoretical perspectives dealing with peers and juvenile groups reach as far back as the 1920s with the Chicago School of thought ( See Park, Burgess et al. 1984 ) In particular, ocused on the role of peers and friends ( Sutherland 1939 ) T his work has become a large body of research which has confirmed tha t peers influence delinquent behaviors ( Akers and Jensen 2006 ) Other ( 1969 ) social contr ol or bonding theory and Gottfredson and ( 1990 ) general theory place important emphasis on the role of families. The relationships of familial factors (broken homes, socialization, supervision, discipline, etc.) with delinquency have also been researched. Parents and siblings, as well as

PAGE 10

10 peers, are thought to affect delinquency (See Stormshak, Comeau et al. 2004 ; Bahr, Hoffm ann et al. 2005 ; Velleman, Templeton et al. 2005 ; Piko and Kovcs 2010 ) The purpose of this research is to examine some of the ways in which frien ds, parents, and siblings may contribute to drug use and delinquency among a sample of serious delinquents drawn from youth in residential facilities that were participating in a faith and community based treatment regimen a regimen which had minimal impac t according to an evaluation ( See Lane, Lanza Kaduce et al. 2009 ) The data are cross sectional responses to a survey administered to the youth during their residential ( 1998 ) Social Structure Social Learning Theory (SSSL) is used to provide a framework for the research. His theory is sele cted because it incorporates interaction and social psychological components dealing with family and friends (i.e. the social learning process involving differential association, differential reinforcement, imitation/modeling, and definitions), as well as attends to different structural arrangements (like broken homes) and social locations (like race and gender) that might affect delinquency. Akers posits that the effects of social structural and social location variables will be mediated by the social lear ning process ( Akers 1998 ; Akers 2009 ) The research addresses three basic questions: 1. Do family structures explain marijuana use and diversity of delinquency among this group of offenders? 2. Do variabl es reflecting the social learning process account for drug use and delinquency among this group of offenders? 3. Will the impact of the family structure variables be mediated by the social learning variable among this group of offenders?

PAGE 11

11 CHAPTER 2 SOCI AL STRUCTURE SOCIAL LEARNING THEO RY Social Learning Theory social learning theory is used to provide a framework for this research, it is the springboard for reviewing the literature. Social structure social learning theory is an effort to integrate our understanding of crime and delinquency across both structural and individual social psychological levels of analysis. Social structure social learning theory is a general theory of crime and deviance. It purports to ex plain all kinds of crime as well as specific crimes; it seeks to account for different kinds of involvement in crime such as frequency, duration, prevalence, and diversity across crimes ( Akers, Krohn et al. 1979 ) Akers ( 1998 ) has provided a summary of his theory: Its main proposition is that variations in the social structure, culture, and location of individual and groups in the social system explain variations in crime rates, principally t hrough their influence on difference among His integrated theory retains the four dimensions of the social learning theory that he developed over the years ( Burgess and Akers 1969 ; Akers, Krohn et al. 1979 ) Historically, t he theory that grew into social learning theory began as an advancement theory ( Burgess and Akers 1969 ) Burgess and Akers (1969) debuted differential association reinforcement theory. interaction with different people affected the differ ent activities that people participate in ( Sutherland 1939 ; Akers 1985 ; Akers 1998 ) Differential association reinforcement theory looked further into this co nnection and focused on the process of how interactions shaped behaviors particularly operant

PAGE 12

12 conditioning ( Burgess and Akers 1969 ) From this theory, Akers formulated social learning theory and explained the elements that he thought were a part of the process of influencing behavior ( Akers 1985 ) s key concept of differential association and contributes additional measurements to the process of learning: imitation, differential reinforcement, and definitions ( Akers 1985 ) While each measurement plays its own individual role in the learning process, it is also enhanced by the others ( Akers 1985 ) Differential reinforcement t, and probability of experienced and ( Akers 1998 ) This is an example of (1966b) explicitly identified the learning mechanisms as those found in modern behavioral theory. They reta ined the concepts of differential association and definitions ter ms and added concepts from behavioral learning theory. These concepts included differential conditioned or shaped by reward and punishments. They also contain classical and stimuli (the environmental and internal stimuli that provide cues or signals for behavior); schedules of reinforcemen t (the rate and ratio in which rewards and punishments follow ( Akers and Sellers 2004 )

PAGE 13

13 W hen people decid e to engage in an activity they receive responses that can be positive, negative, social or nonsocial ; the type of reinforcement received plays a role in determining whether or not the behavior will be continued ( Akers 1998 ) For example, students who are rewarded for good grades are more likely to continue to work for good grades. Also, students who are cheered on by their friends for having bad grades are more likely to continue to get bad grades to impress their friends. The consequences and rewards for their actions will be taken into consideration when deciding to contin ue or stop the behavior. Additionally, p eople may have different perceptions of the same consequence ( Akers 1990 ) Imitation ( Akers 1998 ) ; it also occurs when people observe a behavior from a person that they admire or respect and that they are interested in repeating ( Akers 1 985 ; Akers 1998 ) An example of this is when children first attend school and must get acclimate d to the rigorous schedule They may not initially know the daily routine, but they see how older children act and use that as a guide for what is expected of them and what is allowed. T hey are imitating the people around them. Although personal contact like the kind that takes place at school allows a venue for imitation, it can take place without physical in teraction ; i.e. people can imitate things that they see on televisio n ( Akers 1985 ) Definitions ( Akers 1998 ) Many of the differences in definitions are obtained through differential reinforcement and imitation ( Akers 1998 )

PAGE 14

14 whether or not to engage in certain activities. Definitions can b e formed about specific behaviors such as walking alone at night or about general behaviors such as maintaining personal hygiene; they are usually defined as favorable or unfavorable toward a general or specific belief ( Akers 1985 ) For example, students who have experienced praise for good grades at school are likely to form positive definiti ons about school that encourages them to continue to excel. Contrarily, students who are consistently in trouble with their teacher because of the behavior at school are likely to form negative definitions about school that discourage them from continuing. Decisions to engage in an activity are made on a continuum where multiple definitions shaped by differential association, differential reinforcement, and imitation/modeling may come into play; the more favorable definitions the more likely a person will e ngage in the activity ( Akers 1990 ) Given Akers behavioral reformulation of Sutherland ( 1939 ) definitions are conceptualized as being cue or discriminative stimuli to which behaviors are paired depending on reinfo rcement and punishment contingencies ( Akers 1998 ; Akers 2009 ) Differential association communication, interaction, and identification with others ( Akers 1998 ) This concept describes the variety of people with whom a person associates and provide s different opportunities to imitate behavior and learn the norms of society ( Akers 1985 ) Two of the most common groups that youths associate with include the family (specifically, parents and siblings) and peers; they are considered primary groups and each can have a strong influence on juveni les ( Akers 1985 ) importance of intimate groups and the frequency, primacy, and intensity of associations,

PAGE 15

15 which individuals associate, especially juveniles ( Akers 1998 ) Family and friends will more often have a greater influence than other groups such as a church, book club, or daycare ; although, each type of group provides the opportunity to association with groups of different people ( Akers 1998 ) It is through differential association that people are exposed to differential reinforcement and definitions ( Akers 1985 ) Altogether, differential association, imitation, differential reinforcement, and definitions constitute social learning theory ( Akers 1985 ) This theory may have blossomed out of various other theor ies but it has become one of the leading theories that research ers use to try and explain juvenile delinquency and drug use ( Akers and Jensen 2006 ) Since its inception, many researchers have used it to provide support for their assertions about behavior ( Krohn 1999 ; Akers and Jensen 2003 ) The following section looks at how social learning theory has been tested over the past 30 years. Social Structure Social Learning Theory In 1990, socia l learning theory was extended beyond its study of social process to include an analysis of social structure ( Akers 1998 ) Akers ( 19 98 ) discusses several dimensions of social structure that may be useful in predicting delinquent behaviors. These structures include: differential social organization, differential location in the social structure, theoretically defined social structure, and differential social location ( Akers 1998 ; Akers 2009 ) Only the theoretically defined social structure dimension incorporate s direction al relationships between structural variables and crime or devia nce. Fo r the other dimensions, the theory only states empirically established relationship between the variable and crime and deviance will be mediated by social learning variables ( the theory mediation hypothesis). Altogether these elements of social str ucture seek to add to the explanatory power of social process variables used in

PAGE 16

16 ( Akers 1998 ; Akers 2009 ) Social structure social learning theory also theory hypothe sizes that social structural variables will be mediated by the social process variables in social learning theory ( Akers 1998 ; Akers 2009 ) Differential social organization, one of the two structures tha t have to do with geographical differences across systems (urban versus rural communities, cross national comparisons, regional variations in rates, differences by ecolog ical areas of the ( Akers 1998 ) For example, when studying drinking behavior it may be good to understand if the participant is a white, male from America versus a black, woman, from Canada. Their differential social organization may explain some of th e differences in the social processes that predict their behavior. Differential location in the social structure, the second of the two structures that probable v ariations in crime rates by sociodemographic characteristics, groupings, aggregates, collectives, or categories, such as race/ethnicity, class, gender, marital status, religion, occupation, and other dimensions of social differentiation that exist in socie ( Akers 1998 ) For ex ample, black children and white children may enjoy the same movie, but identify differently with messages or themes in the story. Just by being a part of their perspective locations in the social structure, these two groups view the same situation from dif ferent perspectives. Akers ( 1998 ) elaborates that locations of crime (e.g., there is something inherent being male that makes men more prone to

PAGE 17

17 crime than women), but more frequently the relationship between and crime is viewed conceptually defined features known structural theories of crime and deviance that propose elevated rates in those societies, or segments of societies, that are hypothesized to have higher levels of some abstract defined condition like anomie, conf lict, social disorganization, patriarchy, or class ( Akers 1998 ) For example, some conflict theories focus on the dynamic of the authority and subordinate relationship. The theory suggests that we would expect different process to go on in these very different structural groups ( Akers 1998 ; Akers 2009 ) groups comprising o f the small group and personal networks that impinge directly on ( Akers 1998 ) For examples, social learning theory, as well as other theories, highlights the importance of social groups; families and friends are theoretically expected to have a different influence on peers. Depending on the theory being focused on for structural analysis different groups or characteris tics will be identified. For the purposes of the current research, the researcher will focus only on differential location in the social structure and differential social location. This study will

PAGE 18

18 not be a complete test of social structure social learnin g theory, but will instead focus on family structures from within the differential location in the social structure dimension and demographic information from within the differential social location dimension. Testing of Social Learning Theory Social lea rning theory has been tested in many ways and has stood the test of time by withstanding criticisms and receiving a lot of support from various researchers ( Akers and Jensen 2006 ) In its earlier years, the theory was mostly used to explain juvenile delinquency ( Krohn 1999 ) Krohn (1999) looked at how social learning measurements predicte d behaviors such as use of alcohol, cigarettes and marijuana and they produced support for the theory. He paved the way for social learning theory to be considered one of the leading theories explaining juvenile delinquency and drug use ( Bahr, Hoffmann et al. 2005 ) A kers, Krohn, Lanza Kaduce, & Radosevich ( 1979 ) focused on how differential association with peers affected alcohol and drug use. They found support for their hypotheses and provi ded validation of the theory. Although the study supported learning theory, the authors recognized the need for further testing and urged other researchers to delve deeper into the predictive relationship of drug use and delinquency ( Akers, Krohn et al. 1979 ) Despite extensive evaluation of the theory, there still exists a need for comprehensive testing of the theory that incorporate s the effect of parents siblings, and peers and that inquires about a drug use and delinquency of varying severity ( McDermott 1984 ; Akers and Lee 1999 ; Bahr, Hoffmann et al. 2005 ) Over time, social learning theory was used even more extensively researchers departed from focusing primarily on peers to focu sing on other predictors of juvenile delinquency that social learning theory might help explain ( Akers and Jensen 2006 ) Krohn, Lanza Kaduce, & Aker s ( 1984 ) examined how community contexts (suburban,

PAGE 19

19 rural, and urban) affected juvenile delinquency. The researchers found that social learning mea surements represented the differences in each of these places and were predictive of the level delinquency that juveniles in each area would participate in. This study added to the explanatory versatility of the theory and was support that social learning theory could explain delinquency across varying communities ( Krohn, Lanza Kaduce et al. 1984 ) Parents siblings, and peers deal with juveniles in different social contexts that intermingle, and further research is needed to see how other contexts can be explore d using social learning theory ( Krohn, Lanza Kaduce et al. 1984 ) The current research will investigate how parents siblings, and peers provide different social contexts for juveniles to commit delinquent act and use drugs. Further support for the theory came from its first longitudinal test. Krohn Skinner, Massey, & Akers ( 1985 ) tackled criticisms that the theory received after th e publication of Akers et al ( 1979 ) Some researchers questioned the validity of the theory and if the causal order suggested by Akers would hold up to a longitudinal test. Theorist in support of social learning theory argued back against the mischaracterizations of the research ( Lanza Kaduce, Akers et al 1982 ) and further support for social learning theory was provided over time ( Akers and Jensen 2006 ) Krohn, et al. ( 1985 ) found su pport for the casual order suggested by the theory. This means that social learning also found that both parent s p rovided additional support that it is necessary to investigate the effects that parents and peers have on juvenile delinquency and drug use. At different stages (initiation and maintenance) they are shown to have different effects ( Krohn, Skinner et al. 1985 )

PAGE 20

20 Although it is clear that a complex relationship exists between parents and peers, siblings are another group that deserve consideration ( Akers 1990 ) P eers have a greater than parents and siblings. I t is unclear how interrelations of parent s, siblings, and peer s can impact how the y explain juvenile behavior ( Brook, Brook et al. 2006 ) Another study that looked into factors that affect juvenile drug use and delinquency was Akers and Lee ( 1999 ) This research addressed one of the serious criticisms of social learning theory proposed in Krohn ( 1999 ) ; social learning theory was claiming to subsume the elements of other theories, but had yet to directly test if it was true. In seeking to address the issue that no theory could explain the age crime curve, Akers and Lee ( 1999 ) looked at adolescent smoking from social bonding and social learning theory perspectives. Their research support s the assertion that social learning theory makes about common measurements between the two theories, e.g. how the relationship with peers and family is measured through various interactions and situations; they also found support that social learning expl ains the variance in how age is mediated by social learning processes when predicting marijuana use in juveniles more than social bonding theory ( Akers and Lee 1999 ) This supports the SS SL mediation hypothesis as well since age can be a structural variable ( Akers 2009 ) The current research will further test this theory by using family structure measures, as well as age. Akers ( 1990 ) further highlights the versatility of the theory by providing examples of how features of other theories such as deterrence and rational choice can also be found in social learning theory. This article shows how many researchers have

PAGE 21

21 disregarded the social learni ng variables that they are measuring when they look at other theories or attempt to integrate theories. Some studies do not even acknowledge that they are using social learning variables (Akers, 1990). This emphasizes the great reach that social learning t heory has to explain criminal behavior, specifically juvenile delinquency and drug use ( Akers and Cochran 1985 ) Several general themes can be fou nd in the research discussed above. First, social learning theory has made strides as a theory that explains how behavior is caused through the learning process ( Ben da 1994 ; Bahr, Marcos et al. 1995 ; Triplett and Payne 2004 ; Hwang and Akers 2006 ) Each of the a rticles looked at one or more of the social learning measurements and found it predicted behavior. Second, the articles urged for more studies to test the limits of the theory ( Krohn, Lanza Kaduce et al. 1984 ; Akers, La Greca et al. 1989 ; Hwang and Akers 2006 ) ; this has created a lineage of support for social lear ning theory which is especially strong in explaining the behavior of juvenile offenders ( Akers, La Gr eca et al. 1989 ; Hwang and Akers 2006 ) Lastly, the articles that focused on juvenile delinquency and/or drug use found support for claims that parents and peers are amongst the most influential groups ( Triplett and Payne 2004 ; Bahr, Hoffmann et al. 2005 ; Hwang and Akers 2006 ) Tests of social learning theory have rarely focused on parents, siblings, and peers in one study to explain juvenile drug use and delinquency ( Akers and Jensen 2006 ) ; the current research seeks to bridge this gap in the lit erature. Testing of SS SL Theory Akers ( 2009 ) goes into gr eat detail on the progress that has been made in researching SS SL theory over the past 12 years; although the body of literature is not large, it is increasing with time. Throughout the years, his students, friends, colleagues,

PAGE 22

22 and critics have performed analyses of the theory and if it performs as predicted: social processes mediating the social structure measures to predict behavior (Akers, 2009). In Akers ( 1998 ) SS SL theory was first posed to address the gap in macro and micro level theories. This attempt to create a structural and process analysis of crime and deviance has not drawn a lot of attention since then, but has been supported by several researchers ( Akers 2009 ) In Page ( 1998 ) the author found support for SS SL theory and its mediation hypothesis in a study where family structure predicts juvenile delinquency. Family str ucture consisted of whether the juvenile had 1 or 2 parents and if either or both were birth or step parents; this was a measure of differential social location ( Page 1998 ) Despite providing support for the mediation h ypothesis, Page ( 1998 ) found that not all structural measures (specifically gender) were mediated. This issue has been found in other research as well. In 2004, another attempt to test SS SL theory was made using multiple structural measures to predict juvenile alcohol and marijuana use ( Lee, Akers et al. 2004 ) Lee, et al. ( 2004 ) use structural equation model to test their hypotheses about SS SL theory. They found support for the mediation hypothesis of SS SL theory, but like other tests found that not all variable s mediated by the social process measures; once again, gender/sex was not mediated ( Lee, Akers et al. 2004 ) They address this issue by suggesting that mode ration may be affecting the model and express the potential need to extend SS SL theory to include consideration for moderation, as well as for mediation.

PAGE 23

23 The comprehensive analysis of SS SL theory provided by Akers ( 2009 ) suggest that there is much to still be understood about how different structural dimensions contribute to the overall understanding of crime and deviant behaviors (Akers, 2009). Another major study focused on SS SL theory took a feminist approach to analyzing drinking behavior of college students ( Lanza Kaduce, Capece et al. 2006 ) Association with a fraternity or sorority, age, sex, and race were used as structural measures and tested for mediati on effect with social learning processes. The study found support for the SS SL theory mediation hypothesis, but reported that gender was not mediated by the social learning processes ( Lanza Kaduce, Capece et al. 2006 ) In accord with the previously discussed research, the authors in this study suggest further consideration of which structural measures are n ot mediated by structural measures and that moderation effects should be taken into consideration ( Lanza Ka duce, Capece et al. 2006 ) The first full test of SS SL learning theory was used to predict juvenile alcohol, marijuana, and illicit drug use ( Holland Dav is 2006 ) This study included measures for each of the four structural dimensions as well as a full social learning model; the research supported the SS SL mediation hypothesis ( Holland Davis 2006 ) Although many of the variables were mediated, there was still no mediation effect found for gender suggesting, as did the previous studies, that the learning process affect boys differently than it affects gi rls and that moderation effects need to be considered in the SS SL model ( Holland Davis 2006 )

PAGE 24

24 Building off the previous research, the current research plans to address the gap in understanding how different family structures can assist in predicting juvenile delinquency and marijuana use. Research on Juvenile Drug Use The next section will discuss the most current research on juvenile drug use, particu larly identification, prevention, and intervention. These studies come from a variety of theoretical perspectives and approaches. Although some do not identify themselves as testing social learning theory, they have measurements that can be used to examine social learning theory ( Akers 1990 ) The research discussed in this section not only highl ights the pervasive need to address juvenile delinquency and drug use in research, but it also shows that a great deal of the policy solutions being suggested are grounded in social learning theory. Most of the research discussed here occurred since 2000, but it can be beneficial to look at earlier studies because they help show how research on juvenile drug use and delinquency has evolved over time. McDermott ( 1984 ) examined the eff ects of perceived parental drug use and parental attitudes regarding drug use. The results showed that what parents say can be more important than what they do. The characterization of the measure speaks to the effect that parents behavior has on juveniles definitions and on their imitation/ modeling. Newcomb et al (1986) looked at the current state of the research at the time arguing that people wanted to find one came to use drugs ( Newcomb, Maddahan et al. 1986 ) In their study, Newcomb and his colleagues tried to understand what made teens susceptible to going down any pathway that led to a life of drugs ( Newcomb, Maddahan et al. 1986 ) Their research borrowed

PAGE 25

25 concepts already being ap plied to infectious diseases at the time and the risk factors that they used in their study such as poor relationship with parents, lack of social conformity, perceived peer and adult drug use can be considered social learning measurements of differential association and imitation/modeling ( Newcomb, Maddahan et al. 1986 ) Their results supported the idea that multiple pathways lead to drug use. This study was a major contribution to research because it took a new approach to explaining the phenomenon, in addition to collecting longitudinal data so that the interpersonal process could be better understood. At this time public concern was rising about juvenile drug use, and more and more research began to indicate that school and family factors were predictive of ju venile drug use ( Barnes and Welte 1986 ) Studies varied in whether they focused on the parents, sibling s or friends of the juvenile, but they continuously found that these groups were influential to juvenile behavior ( Kandel 1980 ; Kaplan, Martin et al. 1984 ; Brook, Whiteman et al. 1986 ; Simcha Fagan, Gersten et al. 1986 ; Krohn, Skinner et al. 1989 ) Other researchers have followed this research trend by looking at different potential predictors in hopes of helping provide solutions to juvenile drug use. More recently, Botvin et al. ( 2000 ) recog nized the need for improved strategies in prevention and the need for program evaluation. Although previous research had reported support for prevention programs, those studies were limited in their scope because they only studied the effects on use of one or a few drugs ( Botvin, Griffin et al. 2000 ) Understanding prevention beyond marijuana usage was needed to ac tively treat the vast array of drugs juveniles may use ( Penning and Barnes 1982 ; Botvin, Griffin et al. 2000 ; Svensson 2000 ) Their research extended the previous understanding about

PAGE 26

26 prevention program effects on use of gateway drugs, showing that prevention methods also have a deterrent effect later in lif e against illicit drugs. Additionally, they showed that researchers and practitioners need to pay careful attention to how prevention programs are structured. By better understanding the influence that parents and peers have on drug use programs, better us e can be made of them as a preventative measure ( Brook, Brook et al. 2006 ; Piko and Kovcs 2010 ) Researchers have also examined juvenile behaviors or characteristics that might affect the impact of parents and peers on their own drug use ( Moon, Hecht et al. 1999 ; Rohde, Kahler et al. 2004 ; T riplett and Payne 2004 ) Triplett and Payne ( 2004 ) highlight problem solving skills as an important factor that may prevent juveniles from getting involved in drug use. Problem solving skills are develop ed through socialization with family and friends ( Triplett and Payne 2004 ) Youths who do not have good problem solving skill s may view drug use as a release from the problems and may be more likely to use and use to more frequently ; youth with good problem solving skills are more likely to seek out other options to solve their problems besides drugs ( Triplett and Payne 2004 ) The results supported social learning theory by reporting that parents and peers affect the definitions and reinforcement that the juvenile receives before and after making their decision about drug use ( Triplett and Payne 2004 ) Although many juveniles encounter potential drug offers, it is necessary to remember juveniles are both similar and different from each other; research has shown that different characteristics such as race, gender or even psychiatric disorders can affect whether they are more or less susceptible to offers to use drug and drug use in general ( Moon, Hecht et al. 1999 ; Rohde, Kahler et al. 2004 )

PAGE 27

27 As time progressed, research advanced in recognizing that peers were not the only primary group worth focusing on to address juven ile drug use; eventually researchers realized the need to focus more on parents and the combination of parents and peers as both preventative and risk factors ( Brook Whiteman et al. 1990 ; Andrews, Hops et al. 1993 ; Aseltine 1995 ) Research has shown that both groups have an effect on whether or not juveniles use drugs and that peers have a more significant effect than parents do but their overall and interactive relationship is still not completely understood ( Huiz inga, Loeber et al. 1995 ; Stice and Barrera 1995 ; Hoffman and Johnson 1998 ; Dorius, Bahr et al. 2004 ; Stormshak, Comeau et al. 2004 ; Velleman, Templeton et al. 2005 ; Piko and Kovcs 2010 ) .Velleman et al ( 2005 ) discussed many factors that influence the relat ionship such as parental supervision, parent/peer influences, age of first use, and social coping skills. Bahr et al ( 2005 ) took a theoretica l approach to understanding how family and friends affect juvenile drug use using social learning and social bonding theories. Furthermore, research has also showed that juveniles are more likely to be introduced to drugs through family, friends, and other peers than through drug dealers as previously suggested ( Khavari 1993 ) Additionally, consideration is needed that parents and peers may affect how each other are able to influence juveniles. Mediator or moderator effects may be present, but undetected; research has drawn mi xed conclusions about the presence of mediation and moderation effects in regard to juvenile drug use ( Hoffman and Su 1998 ; Akers and Lee 1999 ; Hwang and Akers 2006 ) Bahr et al ( 2005 ) concluded that their data were more consistent with social learning theory than social control theory measures; again providing confidence that

PAGE 28

28 social learning theory has a role to play in uncovering the mysteries of the parent/peer effect of juvenile drug use. Vell eman et al ( 2005 ) includes a review of the literature that led the author s to suggest more integrated prevention and intervention methods involving parents siblings and peers. The effects of the program can benefit the family as well as the juvenile by allowing them to have greater knowledge of how their support can help the juvenile and their drug use problem. Together these articles show that just like there is no one pathway that leads people to drug use, there is no group that single handedly can cause or solve the juvenile drug use issue ( Simcha Fagan, Gersten et al. 1986 ; Petraitis, Flay et al. 1995 ; Reed and Wilcox Roun tree 1997 ; Brook, Brook et al. 2006 ) As more focus is paid to the problem of adolescent drug use, adequate solutions need to continue being discussed ( Terry, Va nderwaal et al. 2000 ) Parent and peer influences need to be considered regarding all aspects of juvenile drug use ( Piko and Kovcs 2010 ) Researchers have stepped up to the call for improved understanding of how parents and peers affect juvenile drug use. They have even begun to have an understanding, although limited, of how they interact. Brook et al ( 2006 ) had results support the idea that the influences of use. They found that a conflict free parent child attachment will in crease the likelihood that a child will identify with the parent, that attachment and close identification with the traits, attitudes, and behaviors, and that thos e attained characteristics are common in children who affiliate with peers who do not use drugs ( Brook, Brook et al. 2006 )

PAGE 29

2 9 This review of the current research has shown that parents and peers do have an effect on the types of approaches used in the prevention and intervention of juvenile drug use. The previously disc ussed studies offer different protective factors that may help juveniles not use drugs, as well as risk factors that may contribute to their drug use ( Trost, Langan et al. 1999 ; Windle 2000 ) Some research even found that juveniles with drug prone behavior may not be affected by protective factors ( Brook, Brook et al. 2006 ) U nfortunately, the need juvenile drug use Although our knowledge has improved about juvenile drug use, understanding delinquen cy more generally and how it is affected by juvenile, parent, and peer drug use is also vital to understanding youth and their problems. Research indicates delinquency more generally is related often to juvenile drug use ( Hwang and Akers 2006 ) A reason for this is that juveniles do not tend to specialize in a particular kind of crime, but to be generalist (i.e. to commit various kinds of delinquent acts; Visher 1995 ; Akers and Jensen 2003 ) Research on Juvenile Drug Use and Delinquency A lot more focus has been given to adult drug use and the criminal justi ce system than juvenile drug use and the juvenile justice system. Unfortunately, the problem exists and many youth are living delinquent lives that not only lead them to drug use, but also get them involved in the justice system ( Chassin 2008 ) Research has shown that the system is not equipped to deal with juvenile drug use in the same way that it can deal with adult drug use ( Terry, Vanderwaal et al. 2000 ) Terr y et al ( 2000 ) indicate s that an increasing number of juveniles who enter the juvenile justice system have drug problems. Despite the support for drug programs throughout the justice system,

PAGE 30

30 programs for juveniles are few and far between in most jurisdictions, and they are not well integrated to help the juvenile as they progress through the various stages of the system ( Terry, Vanderwaal et al. 2000 ) Since no one treatment has been found to work in all cases ( Terry, Vanderwaal et al. 2000 ) it is essential that programs are made to work with different types of offenders at different stages of their drug use ( Chassin 2008 ) Without programs to help juveniles build skills to resist drug us e, youths are likely to return to drug use once they are released from the system and return to their family and friends that facilitated the drug use in the first place ( Khavari 1993 ; Trost, Langan et al. 1999 ; Terry, Vanderwaal et al. 2000 ) In trying to prevent juveniles from entering the juvenile justice system it is equally as important to address drug use as it is to address delinquency ( Chassin 2008 ) The two have been shown to have a strong relationship ( Botvin, Griffin et al. 2000 ; Terry, Vanderwaal et al. 2000 ; Akers and Jensen 2003 ) and improving upon t he status of one problem will help alleviate the other problem ( Stice and Barrera 1995 ; Catalano, Kosterman et al. 1996 ) Without prevention and intervention methods, juveniles who are involved in delinquency and drug use are likely to end up in the justice system and have difficulties growing up to live a productive life ( Bishop 2000 ) Unfortunately, drug and delinquency lead to more problems like homelessness ( Mallett, Rosenthal et al. 2005 ) or incarceration for juveniles ( Akers 1985 )

PAGE 31

31 CHAPTER 3 THE CURRENT RESEARCH The current research uses cross sectional data. The data are self report data from a population of youth committed to several residential facilities. The outcomes of interest are marijuana use (a fre quency of use ) and diversity of delinquency (the count of different types of delinquency engaged in). There are three hypotheses relevant to social structure social learning theory The direction of these hypothes e s reflect s empirical findings and is not d erived from the theory itself. The i r importance to the theory is that any significant relationship between structural dimensions and crime or deviance should be substantially mediated (See Hypothesis 3 below). Hypothesis # 1: Social Structure variables refl ecting differential social location in family structures (e.g., differential physical lodging, broken home indicator, and residential mobility) and social location variables (e.g., sex, age, and race) will be significantly related to marijuana use and dive rsity of delinquency respectively a) Juveniles who do not live at the dwelling of at least one parent (differential physical lodging) will be significantly more likely than those who do to have also participated in juvenile marijuana use and delinque ncy. b) Juveniles whose parents are not married (broken home indicator) will be significantly more likely than those whose parents are married to have also participated in juvenile marijuana use and delinquency. c) Juveniles who have changed residences (resident ial stability) will be significantly more likely than juveniles who have not changed residence to have also participated in juvenile marijuana use and delinquency d) Juveniles who are white (race), older (age), or male (sex) will be significantly more lik ely than juveniles who are of othe r races younger or female to have also participated in juvenile marijuana use and delinquency

PAGE 32

32 Hypothesis # 2 : Social l earning process variables (definitions, differential reinforcement, differential association, and imit ation) will be significantly related to marijuana use and diversity of delinquency respectively. a) Juveniles with definitions un favorable to marijuana use and delinquency will be significantly less likely than those with definitions favorable to marijuana use and for delinquency to have also participated in juvenile marijuana use and delinquency. b) Juveniles who are differentially reinforce d for delinquency and marijuana use will be significantly more likely than those who are not to have also participated in juvenile marijuana use and delinquency. c) Juveniles who associate with more peers who participate in delinquency and marijuana use will be significantly more likely than those who associate with fewer peers to have also participated in juvenile marijuana us e and delinquency d) Juveniles who associate with peers who mor e often participate in religious activities will be significantly less likely to have participated in juvenile marijuana use and delinquency. e) Juveniles whose parents have use d illegal drugs or ha ve been arrested will be significantly more likely to have also participated in juvenile marijuana use and delinquency. f) Juveniles whose siblings use drugs or have been arrested significantly more likely to have also participated in juvenile marijuana use a nd delinquency Hypothesis # 3 : Social structure measures will be substantially mediated by social learning process in models significa ntly predicting juvenile marijuana use and diversity of delinquency respectively. a) The effects of social structure measure s on juvenile marijuana use and diversity of delinquency will be substantially reduced in models that include social process measures.

PAGE 33

33 CHAPTER 4 METHODS Participants and Design The current research conducts secondary analysis us ing cross sectional data f rom a n evaluation that was conducted from within the Florida Department of Juvenile Justice, the Florida Faith and Community Based Delinquency and Treatment Initiative ( FCBDTI; Lane, Lanza Kaduce et al. 2009 ) Part of that initiative included cognitive behavioral programming, in particular the Thinking for a Change component, specifically designed to alter thinking patter ns to which risky behaviors have been conditioned. The researcher will not collect data from new participants for this study. Partici pants (n=14 6 ) were youths committed to one of five juvenile residential incarceration facilities ( Broward Intensive Halfway House, Britt, San Anto nio Boys Village or SABV, First Step Residential Facility in Orange County or First Step Orange Intensive Halfway House ) who volunteered, with consent of their parents, to participate in a study evaluating a pilot program with faith ba sed components in juvenile institutions S elf report data w ere collected from these participants in an interviewer administered questionnaire 1 Data were collected as part of a process and impact evaluation conducted on the program. Specifically, this project will use data collected from interviews conducted with the participants while housed in five facilities throughout Florida (three male facilities and two female facilities). Researchers from the University of Florida conducted the interviews which asked the participants both closed and open ended survey questions. Collection of the data took place from April 1, 2005 June 30, 2006. 1 This group of participants was a portion of the total number of participants from the evaluation study; an additional 214 participants who official records were gathered for were included in the original study as a control group. Those individual are not included as participants in this study.

PAGE 34

34 The original study included measures from multiple theories and asked the participants to self report on a variety of activities including drug use and delinquency. Youths also answered questions about their family and friends. Other measures collected in the evaluation but not used in the proposed study are official records data, contact data and surveys from mentors and staff at the facility ( Lane, Lanza Kaduce et al. 2009 ) T here was no evidence that these youth were different from offenders committed to other residen tial programs ; results from the evaluation of the intervention program show that the intervention was not very effective ( Lane, Lanza Kaduce et al. 2009 ) The interviews were conducted while the participants were incarcerat ed in residential centers. The questions were worded to reflect what happened before the juvenile was placed in the residential incarceration facility. P articipants completed a self report questionnaire; participants received cards with the questions on them, as well as had the interviewer read the question s aloud. Study Variables Dependent Variables The dependent variables in this study are juvenile marijuana use and a diversity of juvenile delinquency index The marijuana use and diversity of delinquency measures are self report. Marijuana use : Respond marijuana

PAGE 35

35 use is coded for the number of days marijuana was used ranging from 0 to 30 days 2 On average, juvenil es reported using marijuana more than 13 days out of the 30 days prior to their entering the facility (mean=13.5753) with a standard deviation of a little more than 14 days (stdev= 14.1575). The most common answer was no use in 30 days prior to entering th e facility (mode=0). 3 Diversity of delinquency theory purports to be able to account for all kinds of involvement with crime and delinquency, specific crimes, as well as cri mes generally. Recall that for this study the specific behavior of marijuana use is measured in terms of frequency use over the 30 days prior to placement in the residential facility. The summed frequency of involvement across a list of delinquent acts pro duced an extremely skewed distribution for this sample, so a different approach is taken. To measure juvenile delinquency, participants were asked to report their involvement in a 14 deviant or delinquent activities: (1) beat somebody up o used check or credit cards (10 (11 committed or 2 Because of a skip pattern, some respondents who had not used marijuana in the past 30 days had to be reclaimed from the question asking if they had ever used marijuana and coded as a zero in the new variable. 3 Marijuana use in this sampled yielded a bimod al distribution with many respondents reporting no use in the 30 days before placement and many others reporting what amounted to daily use.

PAGE 36

36 attempted homicide (13) committed or attempted sex by force or (14) been in a gang or posse fights Additionally, they reported the number of times they had committed the act and the number of times they had been arrest for each crime. For the purpose of the current research each delinquent act was dichotomously recoded 0 for zero times and 1 for one or more other times. Then an index was created from a count of the dichotomous variables. This index was coded from 0 to 14 for the number of criminal/ deviant acts that the juv enile participated in. This creates a diversity of juvenile delinquency index instead of a frequency measure or scale of delinquency 4 Juveniles reported committing at least three types of delinquency/deviance on average (mean=3.34; median=3) with a standa rd deviation of over two offense types (stdev=2.54). Independent Variables The independent variables in the current study are social structure social learning measurements, with a specific interest in how peer and family (parents and siblings) influences i nvolving drug use or delinquency and delinquency. Social structure v ariables Social structural measures will come from two structural dimensions from SS SL theory: differential location in the social structure and d ifferential social location. Although these structural dimensions have names that are similar, they differ in what they measure. 4 The diversity index was used instead of an additive scale of frequency of use across the various kinds of delinqu ency/deviance because the scale was highly skewed. The diversity index was correlated the additive frequency scale (r=.350), and the diversity measure provided a less biased measure.

PAGE 37

37 A. D ifferential location in the social structure : Differential location in the social structure measures characteristics of th e individual that have a group effect such as ( Lanza Kaduce and Capece 2007 ) Differential location in the social structure (characteristics of the individual that have a group effect) will consist of age, sex, and race. Age is a continuous meas ure that is coded from 13 to 18 to represent the range in years of age of the participants. The average age in this sample is about 16 with a standard deviation of 1.3 years. Sex is a dichotomous measure that is coded 0 for females and 1 for males. The maj ority of participants (82%) in the sample are male (n=120). Race is a categorical measure that originally included 19 race and ethnic options. From the original measure, three dummy variables (white, black, and other) were created. White is coded 1 for whi te and 0 for everyone else. Black is coded 1 for Black and 0 for everyone else. Other is coded 1 for all other non white/ non black races or ethnic groups White serves as the reference group for analysis. The racial breakdown of the participants in the sam ple is 35% white, 44% black, and 21% other (not white or black). B. Differential social location : D ifferential social location levels of social organization. The focus is the recognized groups with which individuals deal, as well as r eference groups with w h ( Lanza Kaduce and Capece 2007 ) measures characteristics of the group such as its size or the type of members that belong. For this study, d ifferential social location (characteristics of the group) consists of family structure measures: differential physical lodging, broken home indicator, and

PAGE 38

38 residential stability. Differential physical lod ging is measured by asking the juvenile to created and coded 0 for residing in the ho use/home of at least one parent and 1 for not residing with at least one parent. Fewer juveniles reported living apart from their parent(s) (29%) than reported living with their parent(s) (72%). A broken home indicator was created from a question asking ju was recoded in the reverse: 0 for married to each other and 1 for not married to each other. The majority of respondents (78%) indicated they lived in a broken home where there parents are not married. Residential stability is measured by asking juveniles to report the number of times this The number of times is truncated four or more moves to eliminate outliers and decrease bias in the distribution of the variable This variable was coded from 0 for no moves to 4 for 4 or more moves in the year prior to entering the facility. On average, the respondent had moved at least once in the year prior to entering the facility (mean=1.27). The four dimensions of social learning theory (definitions, differential reinforcement, differential association, and imitation/modeling) are included in the study. Social learning process v ariables differential reinforcement, differential association, and imitation/modeling. Each component was operation al ized for this research

PAGE 39

39 A. Definitions : For definitio ns the researcher will look at statements that inquire about whether juveniles think certain actions are right or wrong depending on the circumstance. For the analysis of marijuana of use, the questionnaire item to agree or disagree with the following statement: This item is coded from unfavorable to favorable definition s of marijuana use The response format is coded 1 for strongly disagree 2 for disagree, 3 for agree and 4 for strongly agree On average, juveniles reported that they more agree than disagree that it is For the delinquency analysis the statements were focused on definitions favorable or unfavorable for four dev iant activities two of which were reverse coded One definitional statement rea d : beat up, threaten or use a weapon to hurt someone definitional more definitions were worded i break the law if you are not really hurting anyone. justified or necessary to break The re s ponses were coded 1 for strongly disagree 2 for disagree, 3 for agree and 4 for stron gly agree To form a scale, the responses for the first and second items were reversed so that the coding for all four definitions was in the same direction. The coding, unlike that for the marijuana definition, runs from favorable to unfavorable definitio ns of delinquency. The four items were summed and the sum was divided by four to create an additive scale. Analysis of this scale found a reliability coefficient that is 685). After recoding the two items, the scales score indicate that o n average, juveniles reported that they tend to disagree more than agree that it is

PAGE 40

40 wrong to partak e in the listed delinquent activities (mean=2.1459). In other words, they averages sug gest that the respondents hold definitions favorable to delinquency B. Differential reinforcement : For differential reinforcement the researcher will look at questions w here juveniles were asked about their perceptions of the rewards and consequences of certain behaviors. For the analysis of marijuana use, respondents were asked The response alternative For the delinquency analysis the questions were focused on three deviant activities: happens to a teenager who : tries to or beats somebody up or threatens someone with a weapon steals money or property and breaks into a house, building or car The response alternative s were coded 1 for 3 for Differential reinforcement of delinquency variables were added together and then divided by the number of items to create a 3 item scale Analysis of this scale found a reliabi l ity coefficient that is acceptable for use On average, juveniles reported that the effect on a teenager who uses drugs teenager who commits delinquent acts C. Differential association : Two approaches were taken to tap differential association. For differential peer association the researcher will look at questions where the juvenile reporte d their perceptions of their friends behavior s and experiences For

PAGE 41

41 the marijuana analysis, the respondents were asked committed to this facility, how many of your best friends, or the friends you spend the most time with had use marijuana ? The response alternative were coded 1 for none, 2 for some, 3 for half, 4 for most, and 5 for all of their friends. Only the question about differential association of peers who use marijuana was included in the marijuana use model s. On average, juveniles reported that more than half of their friends had used marijuana recreationally (mean=3.77). For the delinquency analysis the differential association questions were focused on 1 2 indicators of deviant activities measured through a series of questions : for been suspended from school been picked up by the police (3) ap peared in juvenile court for a crime been put in juvenile detention property stolen t h ings worth a lot of money (7) stolen a car, truck or motorcycle (8) tried to or beat somebody up or threat en ed someone with a weapon stole small things not worth much money been put into a juvenile facility or program (11) been on probation or (12) been transferred to adult court? The response alternatives were coded 1 for none, 2 for some, 3 for half, 4 for most, and 5 for all of their friends. Differential peer association variables measuring deviant behavior indicators were added together and then divided by the number of items to create a 1 2 item scale Analysis of this scale found a reli ability coefficient that is acceptable for use On average, juveniles reported that between some and half of their friends had committed the listed delinquent activities (mean=2.5616)

PAGE 42

42 In addition to the respective marijuana use and diversity of delinquency measures of differential peer association, the researcher included a measure of pro social religious behavior. Respondents were asked committed to this facility, how many of your best friends or the friends you spend the from 1 for none 2 for some, 3 for half, 4 for most, and 5 for all friends participated in church religious activities. On average, juveniles reported that at least some of their D. Imitation/modeling : To capture modeling or imitation influences, questions asked about parental and sibling use of drugs and official arrests for crim e. For the marijuana analysis, the researcher will look into parent and sibling drug use as measure s of modeling/imitation 5 The questionnaire did not include items about specific drugs like marijuana for parents or siblings, but did ask about drugs gener ally In one question juveniles were asked to report the drug use o f either or both of their parents This measure is cod ed 1 for yes if either or both parents use drugs or 0 for if neither parent use s drugs. To measure sibling drug use the researcher use s a question where juveniles were asked to report the drug use ed to the facility, This measure is coded 1 for yes r 0 for none For the marijuana use models, 5 Parental (and/ or sibling) involvement with drugs is usually conceptua lized as imitation or modeling in social learning analysis Akers, R., M. Krohn, et al. (1979). "Social Learning and Deviant Behavior: A Specific Test of a General Theory." American Sociological Review : 636 655. That a parent or sibling uses does not capture the quality of interactions for associations and is, therefore, not considered to be a form of differential association

PAGE 43

43 14% of juveniles reported that either or both of their pare nts had used illegal drugs (mean=.14) and 29% reported that any of their brothers and sisters had used drugs (mean=.29). For the delinquency model, juveniles were asked to report the level of involvement that their parents and siblings had with the legal system in the year befor e the juvenile was incarcerated The original measures specifically ask about mom/female guardian, dad/ male guardian, and siblings. For the purposes of this study, a new variable was computed to represent an arrest for either or b oth parents and was coded 1 for yes either or both parent w ere arrested in the past year or 0 for if neither parent was arrested. The measure of sibling arrests was coded 1 for yes any sibling was arrested in the past year or 0 for if no sibling was arrest ed. For the delinquency models 13% of juveniles reported that either or both of their parents had been arrested in the year prior to them entering the facility and 29% reported that any of their brothers and/or sisters (including natural or step) had been arrested in the year prior to them entering the facility (mean=.29). Procedure and Analysis To start the analysis several descriptive analyses are run to get information regarding the distributions of the major variables such as juvenile marijuana use, juv enile delinquency, social structure measures and social learning process measurements Then the basic bivariate statistics are obtained in the form of a correlation matrix. From there, several sets of Ordinary Least Square (OLS) regression models are run separately for both of the dependent variables ( marijuana use and diversity of delinquency ). Each of the two dependent variables will include three OLS regression models that address the respective research hypotheses In the first set of

PAGE 44

44 OLS models socia l structure measures are regressed on the dependent variables to isolate the effects of the social structure variables In the s econd set of models only social learning process measures are regressed the dependent variables to see what their impacts are In the final set of OLS model s, both the social structure and social learning process variables are regressed on the dependent variables so that the mediation hypothesis can be tested. The researcher look ed to see what measures are significant in each mod el 6 as well as whether mediation occurs for the social structure through the social learning process variables. Variables that are not significant will not be removed from the analysis because of their theoretical importance to SS SL. The first set of ana lyses, in line with the first research question and related hypotheses (1 and 1a 1d) predict s the dependent variables ( marijuana use and diversity of delinquency ) separately using the four social structure variables (differential physical lodging, broken home indicator, residential mobility, and demographic information) to see how they relate to marijuana use and diversity of delinquency respectively The second set of analyses, in line with the second research question and related hypotheses (2 and 2a 2 f ) predict s the dependent variables ( marijuana use and diversity of delinquency ) separately using variables for each of the four social learning process es (definitions, differential reinforcement, differential association, and imitation/ modeling) to see how they relate to marijuana use and diversity of delinquency respectively. 6 The researcher reported .01, .05, and .10 levels of significance G iven the relatively small sample size the large number of independent variable s and the search for mediation even the .10 level of significance might help interpret the results

PAGE 45

45 The third set of analyses, in line with the third research question and hypothesis 3 predict s the dependent variables ( marijuana use and diversity of delinquency ) separately on both the social structure variables and the social learning process variables to see how all together they relate to marijuana use and diversity of delinquency respectively. If the mediation hypothesis is correct, the effects of the social structure varia bles should be diminished substantially from what they were in the models using only social structure variables for drug use and delinquency, respectively. Missing Values Because so many variables are included to create scales and indices, an occasional mi ssing value could result in a number of lost cases if a list wise deletion of cases with missing values is employed In the current study, t hree cases were excluded because they were missing data on the marijuana us e dependent variable. Because the diversi ty of delinquency index was a count across 14 types of delinquent act s, the count was made of indications of delinquency with missing values counted as zeroes. The analyses therefore, are performed on a 146 cases for both dependent variables To deal with missing values on the independent variables several steps were taken. F irst, f requencies were run to determine the best way to handle missing values. Second it was decided th at all missing values would be replaced with the median value of the respective variable because the questionnaire items used to measure or construct independent variables had only occasional missing cases. Therefore, it was decided that all missing values would be replaced with the median value of the respective item. Before any va l ues were replaced, a filter was put in place to include only the 146 cases that had measures for both dependent variables in the analysis. After this

PAGE 46

46 change, six variables still had missing values: residential stability (n=142) definitions for marijuana use model (n=143), DA peer religious behavior (n=145) definitions for delinquency (n=141), differential reinforcement of delinquency (n=14 5 ), and DA peer delinquency (n=1 36 ) Both imitation measures (imitation/modeling parent criminal behavior and imitati on/modeling sibling criminal behavior) were constructed as indicators and coded 1 if there was any indication of or parental or sibling involvement in crime or 0 if there not. Therefore, neither of these variables had missing values. For t he six variables with missing data me dian substitution was used. Across all these variables, i f list wise deletion of cases was used the result would have been a substantial decrease in the results of the number of cases included in each regression. Since no single variab le was responsible for losing a lot of cases m edi a n substitution was used. It was selected over mean substitution becaus e the median was not as affected by extremes as the mean was

PAGE 47

47 Table 4 1. Dependent Variable: Marijuana Use Purpose Question Makeup Determine the number of days in the 30 days prior to them entering the incarceration center that they used marijuana "How many days in the 30 days before you were committed, did you use marijuana?" responses can range from 0 to 30 days Table 4 2. Dependent Variable: Diversity of Delinquency Index Purpose Question Makeup Determine the types of delinquency that juveniles have committed in the past year prior to them entering the incarceration center "Tell me whether you were involved in each activi ty during the year before you came to this facility?" 1. Damaged or set fire to school property 2. Damaged or set fire to other property 3. Broke into a house, building, or car in order to take something (when no one was home) 4. Robbed or held up a place of busines s (store, gas station, bank taxi, etc.) 5. Robber a person (street, robbery, mugging, purse snatching, hold up in a house or car; exclude business robberies) 6. Tried to beat up somebody or threaten someone with a weapon (tried to shoot, stab, cut, beat, strangl e, or strong arm someone, even if no one was hurt 7. Stole or boosted something (stole from a till, shoplifted, picked pockets, or took something w/o owner's knowledge; exclude 8. Stole a car, truck, or motorcycle 9. Used check or credit cards illegally (forgery, used a stolen or bad credit card, passed a bad check) 10. Ran cons or scams (defrauded person, business or government) 11. Provided sex for money (prostitution) 12. Committed or attempted homicide 13. Committed or attempted sex by force (RAPE) 14. Been in gang or posse fights

PAGE 48

48 Table 4 3. Model 1: Can social structure s explain juvenile marijuana use and diversity of delinquency? Purpose Independent Variables Dependent Variables Analysis Plan To determine i f S SS L will significantly predict marijuana use and delinquency Differ ential Social Location Diff. Physical Lodging Broken Home Residential Stability Diff. Location in Social Structure A ge Black and Other R ace, Sex "How many days in the 30 days before you were committed, did you use marijuana?" OLS Multiple Regression Equation for Marijuana (coded for # of days ranging from 0 30) And Delinquency (coded from # of activities they admitted to) "Tell me whether you were involved in each activity during the year before you came to this facility? Table 4 4. Model 2 : Does social learning significantly predict marijuana use and diversity of delinquency? Purpose Independent Variables Dependent Variables Analysis Plan To determine is SLT will significantly predict juvenile drug use and delinquency Social Learning Theor y Differential Peer Association Definitions Differential Reinforcement Imitation/ modeling "How many days in the 30 days before you were committed, did you use marijuana?" OLS Multiple Regression for Marijuana (coded for # of days ranging from 0 30) A nd Delinquency (coded from # of activities they admitted to) "Tell me whether you were involved in each activity during the year before you came to this facility?"

PAGE 49

49 Table 4 5. Model 3: Will the impact of the family structure variables be mediated by the social learning variables? Purpose Independent Variables Dependent Variables Analysis Plan To determine if social structure measures will be substantially mediated by social process variables. Social Structure Measures Diff Physical Lodging Broke n Home index Residential Stability A g e Black or Other R ac e S ex Social Learning Theory Diff Peer Association Definitions Differential Reinforcement Parent Imitation/ model s Sibling Imitation/models "How many days in the 30 days befo re you were committed, did you use marijuana?" OLS Multiple Regression for Marijuana (coded for # of days ranging from 0 30) And Delinquency (coded from # of activities they admitted to) "Tell me whether you were involved in each activity during the year before you came to this facility?"

PAGE 50

50 CHAPTER 5 RESULTS Univariate Analysis The values for the univariate analysis have already been shared in the previous chapter as each variable was described. For further reference, Table 5 1 provides a summary of the univariate analysis Bivariate Analysis The bivariate analysis consists of zero o rder correlations between the independent and dependent variables. For marijuana use, the following variables were significantly correlated at a .1 level or higher : d ifferential physical lodging ( r= .235), male ( r= .197), definitions ( r= .232) [coded from unfavorable to favorable definition s of marijuana use] differential reinforcement ( r= .239), differential association peer drug use ( r= .473), differential association peer religious behavior ( r= .174), and imitation/modeling parent drug use ( r= .186). For delinquency, the following variables were significantly correlated at a .1 level or higher: Black ( r= 191 ), age ( r= .180 ), definitions r= ( 3 18 ) [coded from favorable t o unfavorable definitions of delinquency ( opposite the direction of that used for marijuana definitions ) ] differential reinforcement ( r= 235 ), and differential association peer delinquency ( r= 482 ). For further reference, Table s 5 2 and 5 3 provide summ aries of the bivariate analys e s. Multivariate Regression Analysis The results that follow are from Ordinary Least Squares Regression predicting marijuana use and diversity of delinquency. One set of multivariate regression models is focused on predicting juvenile marijuana use and a second set is focused on predicting diversity of delinquency. The analysis discusses each hypothesis and provides an

PAGE 51

51 explanation of the regression results presented in the table s (See Table 5 4 for Marijuana Use and Table 5 5 for Diversity of Delinquency) Predicting Marijuana Use Th e first analys e s examine the ability for structural measures to predict marijuana use, the ability for social process measures to predict marijuana use, and the ability of the social p rocess varia bles to mediate the effects of structural variables in predict ing marijuana use. The hypotheses that were posed earlier in the paper will be discussed in regards to marijuana use, as well as an interpretation of the corresponding model s A. Hypothesis 1 u sing social structures only : The first hypothesis states that tructure variables reflecting differential social location in family structures (e.g., differential physical lodging, broken home indicator, and residential mobility) and social locatio n variables (e.g., sex, age, and race) will significantly be related to marijuana Alth ough only sex (male) was sig nificantly related to marijuana use (r=.197), several familial structural variables were significantly related in the multivariate anal ysis. The first column of figures in Table 5 4 presents multivariate results that address this first hypothesis. Two structural dimensions (differential social location and differential location in the social structures) are included in the regression. O v erall M odel 1 explains 8% of the variance (adjusted R 2 =.081) and is significant at the .01 level (F=5.833). The constant is 8.325 days, but is arbitrary because it is negative. The multivariate analysis provides, at best, weak support for the first gener al hypothesis. The results under Model 1 in Table 5 4 also address sub hypotheses concerning structural variables. Juveniles who do not live at the dwelling of at least one parent (differential physical lodging) will be significantly more likely th an those who do to have also participated in juvenile marijuana use.

PAGE 52

52 dwelling decreased the number of days that they had participated in marijuana use. On average, juvenile s who do not reside with their parents decrease the number of days in the past 30 days that they have used marijuana by almost 7 days (b= 6.964), in comparison with juveniles who do reside with their parents and with all other variables held constant. Note that since SS SL is silent as to the direction of the relationship between this structural vari able and marijuana use, the existence of a relationship allows the mediation hypothesis to be examined (See Hypothesis 3 below). Juveniles whose parents are no t married (broken home indicator ) will be significantly more likely than those whose parents are married to have also participated in juvenile marijuana use. This was supported and it was found that juveniles used marijuana more days in the 30 days prior t o entering the facility if their parents were not married On average juveniles who come from broken homes used marijuana at least four days more in the 30 days before placement (b=4.63), in comparison with juveniles who came from intact homes with all ot her variables held constant. Juveniles who have changed residences more often (residential stability) will be significantly more likely than juveniles who have not changed residence to have also participated in juvenile marijuana use. This sub hypothesis w as not supported since the residential mobility variable was not significant in the model. Juveniles who are white (race), older (age), or male (sex) will be significantly more likely than juveniles who are of other races, younger, or female to have als o participated in marijuana use. This sub hypothesis was partially supported ; males used drugs more days in the 30 days prior to entering the facility than did females. On average, juveniles who are male increase the number of days in the past 30 days that they have used marijuana by

PAGE 53

53 7 days (b=7.089), in comparison to juveniles who are female and with all other variables held constant. Age and race (as measured by dummy variables for Blacks and Others) were not significantly related to marijuana use. Of th e variables that were significant in this model, their relative importance (.i.e. the standardized effect of each variable) is ordered as follows: differential physical lodging (beta= .233), male (beta=.199), and broken home indicator (beta=.140). B. Hypot hesis 2 using social learning processes only : The second hypothesis ocial l earning process variables (definitions, differential reinforcement, differential association, and imitation) will be significantly Six var iables we re used to predict marijuana use: definitions, differential reinforcement, differential association (peer drug use and religious behavior), and imitation modeling (parents and siblings). At the bivariate level, definitions (r= .232) differential reinforcement (r = 2.39), differential association with marijuana using peers (r=.473), differential association with religious behavior (r= .174), and parental modeling imitation (r=.186) were statistically related to marijuana use. The relevant multivar iate results to address the second hypothesis (and it s associated sub hypotheses) are found in the second column of figures in Table 5 4 under M odel 2 O verall M odel 2 explains 25% of the variance of marijuana use (adjusted R 2 =.253) and is significant at the .01 level (F=9.173). The second general hypothesis receives modest support but not as many social learning variables are related to marijuana use in this sample as expected. The constant in the multivariate analysis is 7.550, and signifies that on ave rage juveniles use marijuana more than 7

PAGE 54

54 days out of the 30 days prior to them entering the facility, controlling for the independent variables. Th e second hypothesis has several sub hypotheses reflecting what is expected for each of the social learning components Juveniles with definitions unfavorable to drug use will be significantly less likely than those with definitions favorable to marijuana use to have also participated in juvenile marijuana use This was not supported since the definitions varia ble was not significant in the model. The finding is not expected by the theory and the strength of its relationship to marijuana use (as measured by the zero order correlation of .232) was diminished and was no longer significant in the multivariate anal ysis. Juveniles who are differentially reinforced for marijuana use will be significantly more likely than those who are not to have also participated in juvenile marijuana use. This was not supported since the dif ferential reinforcement variable was not significant in the model. This finding is not expected by the theory and the strength of its relationship to marijuana use (as measured by the zero order correlation of .318) was diminished and was no longer signif icant in the multivariate analysis. Juveniles who associate with more peers who participate in marijuana use will be significantly more likely to have also participated in marijuana use. This was supported and it was found the more peers that used marijua na, the more days juveniles used marijuana in the 30 days prior to entering the facility. On average, for every one unit increase amount of friends who use marijuana in the past year, juveniles increase the number of days in the past 30 days that they hav e used marijuana by more than 4 days (b=4.317), with all other variables held constant.

PAGE 55

55 Juveniles who associate with more peers who more often participate in religious activities will be significantly less likely to participate in marijuana use. This was not supported since the imitation/modeling siblings variable was not significant in the model. The finding is not expected by the theory and the strength of its relationship to marijuana use (as measured by the zero order correlation of .174) was diminish ed and was no longer significant in the multivariate analysis. Juveniles whose parents have used illegal drugs will be significantly more likely to have also participated in marijuana use than juveniles whose parents do not use illegal drugs. This was supp orted and it was found that juveniles whose parents used illegal drugs used marijuana more days in the 30 days prior to entering the facility than juveniles whose parents did not use illegal drugs. On average, juveniles whose parents had used illegal drug s in the year prior to them entering the facility increase the number of days in the past 30 days that they have used marijuana by more than 5 days (b=5.594), in comparison to juveniles whose parents had not used drugs and with all other variables held con stant. Juveniles whose siblings have used drugs are significantly more likely to have also participated in marijuana use than juveniles whose siblings have not used drugs. This was not supported since the imitation/modeling siblings variable was not signif icant in the model. In the multivariate analysis, t wo of the independent variables were significant in predicting marijuana use: differential association peer marijuana use and imitation/modeling parents. Those who reported more association with marijua na using peers and who had at least one parent who used drugs tended to use marijuana more regularly before their placement. Of the variables that were significant in this model,

PAGE 56

56 their relative importance (.i.e. the standardized effect of each variable) is ordered as follows: differential association peer marijuana use (beta=.402) and imitation/modeling parents (beta=.144). C. Hypothesis 3 mediating social structures with social learning processes : The third hypothesis expects the effects of s ocial struct ure measures on marijuana use to be substantially mediated by social learning process variables The sub hypothesis for marijuana use states: The effects of social structure measures on juvenile marijuana use will be substantially reduced in models that in clude social process measures The relevant results are derived from comparing the coefficients and probabilit ies presented for M odel 3 i n the third column of figures in Table 5 4 with those from M odel 1 (presented in the first column) Both the social stru cture and the social learning proces s measures are included in M odel 3 but only the social s tructure ones are presented in M odel 1 Th e hypothesis was partially supported since not all the structure variables experienced a substantial reduction in effect. The broken home indicator variable was substantially mediated in that the significant relationship to marijuana use in Model 1 was rendered insignificant in Model 3 (and the standardized regression coefficient or beta shrunk from .14 to .085). T he sex (mal e) variable was partially mediated ; it was significantly related to mar ijuana use at the .05 level in M odel 1 but only at the .10 level in M odel 3 (and the beta or standardized coefficient dropped from .199 to .151). D ifferential physical lodging was not mediated at all ; its relationship to marijuana use was signifi cant at the .001 level in both m odels 1 and 3 (and its standardized regression coefficients were stable at .233 and .214 in the two models The other structural variables were not significant in the first model and therefore could

PAGE 57

57 not be mediated since the ir effect was never significant. Additionally, the process variables differential association peer marijuana use and imitation/modeling parents increased slightly in their effect after the s tructure variables were added to the model and differential reinforcement became significant. The overall analysis of M odel 3 explains about 30% of the variance in marijuana use (adjusted R 2 =.296) and is significant at the .01 level (F=5.688). The constant is 2.466 days, but is arbitrary because it is negative. Four of the independent variables were significant in predicting marijuana use: differential physical lodging, sex (male), differential association peer marijuana use, and imitation/modeling paren ts. Each of these can be interpreted. On average, juveniles who do not reside with their parents decrease the number of days in the past 30 days that they have used marijuana by more than 6 days (b= 6.415), in comparison to juvenile who do reside with the ir parents and with all other variables held constant. On average, juveniles who are male increase the number of days in the past 30 days that they have used marijuana by 5 days (b=5.398), in comparison to juveniles who are female and with all other variab les held constant. On average, for every one unit increase in negative reinforcement to delinquent activities, juveniles decrease the number of days in the past 30 days that they have used marijuana by more than 2 days (b= 2.519), with all other variables held constant. On average, for every one unit increase amount of friends who use marijuana in the past year, juveniles increase the number of days in the past 30 days that they have used marijuana by more than 4 days (b=4.060), with all other variables hel d constant. On average, juveniles whose parents had used illegal drugs in the year prior to them entering the facility

PAGE 58

58 increase the number of days in the past 30 days that they have used marijuana by more than 6 days (b=6.153), in comparison to juveniles w hose parents had not used drugs and with all other variables held constant. Of the variables that were significant in this model, their relative importance (.i.e. the beta or standardized effect of each variable) is ordered as follows: differential associa tion peer marijuana use (beta=.378), differential physical lodging (beta= .214), imitation/modeling parents (beta=.158), male (beta= 151) and differential reinforcement (beta= 133). Predicting Diversity of Delinquency The second set of multivariate regr ession models is focused on predicting the diversity of juvenile delinquency. T h is section discusses the hypothese s and the corresponding models to examine the ability for structural measures to predict diversity of delinquency the ability for social lear ning process measures to predict diversity of delinquency and the extent to which the social learning process variables mediate the effects of the structural variables A. Hypothesis 1 using social structures only : The first hypothesis states that l s tructure variables reflecting differential social location in family structures (e.g., differential physical lodging, broken home indicator, and residential mobility) and social location variables (e.g., sex, age, and race) will significantly be related The biva riate relationships between delinquency and r ace (Bl ack) (r=1.191) and age (r= .180) were confirmed in the mult ivariate analysis presented in M odel 1 (the first column o f figures) presented in Table 5 5. The general hypothesis, however, is only wea kly supported. T he structural model (M odel 1) explains only 4% of the variance ( adjusted R square=.040); its F value (1.864) is significant at the .10 level. The constant

PAGE 59

59 is 8.155 types of delinquency. It can be interpreted from the co nstant that on average, juveniles participate in at least 8 types of delinquency, controlling for the other variables. This structural hypothesis has several sub hypotheses. Juveniles who do not live at the dwelling of at least one parent (differen tial physical lodging) will be significantly more likely than those who do to have also participated in a greater div ersity of delinquent activities This was not supported since differential physical lodging was not found to significantly predict diversit y of delinquency. Juveniles whose parents are not married (broken home indicator) will be significantly more likely than those whose parents are married to have also participated in a greater diversity of delinquent activities. This was not supported since broken home indicator was not found to significantly predict diversity of delinquency. Juveniles who have changed residences more often (residential stability) will be significantly more likely than juveniles who have not changed residence to have also pa rticipated in greater diversity of delinquency. This was not supported since residential stability was not found to significantly predict diversity of delinquency. Juveniles who are white (race), older (age), or male (sex) will be significantly more lik ely than juveniles who are of other races, younger, or female to have also participated a greater diversity of delinquent activities. This sub hypothesis was partially supported S ex was not significant in the model, but race and age did significantly pre dict a greater diversity of delinquency. In line with the hypothesis, it was found that juveniles who are Black had less diversity of delinquency than other juveniles (White or Other) On average, juveniles who are Black decrease their diversity in delinqu ency by almost one type of delinquent act, in comparison to all other juveniles and with all other variables held constant.

PAGE 60

60 Contrary to the hypothesis, it was younger rather than older juveniles who tended to participate in a greater diversity of delinque ncy. On average, for every one unit increase in age (1 year), juveniles in this sample decrease their diversity in delinq uency by less than one type of delinquent act, with all other variables held constant. Two of the structural variables were significant in predicting diversity of delinquency in the multivariate analysis : race (Black) and a ge. Of the variables that were significant in this model, the relative importance (i.e. the beta or standardized effect of each variable) is ordered as follows: race ( B lack ) (beta= .169) and a ge (beta= .163). B. Hypothesis 2 using social learning processes only : The second general ocial l earning process variables (definitions, differential reinforcement, differential association, and imitation) will be significantly related to In the bivariate analysis, definitions (r=.318), differential reinforcement (r= .235), and differential association with delinquent peers (r=.482) were significantly related to delinquency Only definitions and differential association with delinquent peers remained significant in the mult ivariate analysis presented in M odel 2 (the second column of figures in Table 5 5 ) The overall model explains nearly 30% of the variance in diversity of delin quency (adjusted R 2 =.297) and is significant at the .001 level (F=11.197). The constant is .1119, but is arbitrary because it is negative. This hypothesis has several sub hypotheses. Juveniles with definitions unfavorable to delinquency will be significa ntly less likely to have also participated in diversity of delinquency T he relationship was statistically significant (at the .001 level) ; this sub hypothe s is was supported because juveniles with definitions favorable to delinquency reported

PAGE 61

61 significantly more diversity of delinquency. This direction is consistent with theoretical expectations. On average, for every one unit increase in the strength that juveniles agree that it is alright to commit delinquent activities, the juveniles increase their dive rsity of delinquenc y in the past year by at least one additional type of delinquency (b=1.302), with all other variables held constant. Juveniles who are differentially reinforced for delinquency will be significantly less likely than those who are not to have also participated in diversity of delinquency. This was not supported since the differential reinforcement variable was not significant in the model. Juveniles who associate with more peers who participate in deli nquent acts will be significantly more likely than those who associate with fewer peers to have also participated in diversity of delinquency. This was supported and it was found the more peers who participated in delinquency, the greater the diversity in delinquency. The relationship was significant at the .001 level. On average, for every one unit increase amount of friends who committed delinquent acts in the past year, juveniles increase their diversity of delinquency in the past year by at least 1 ad dition type of delinquency (b=1.312), with all other variables held constant. Juveniles who associate with peers who more often participate in religious activities will be significantly less likely to have participated in a diversity of delinquency. This w as not supported since associating with religious friends was not significant in the model. Juveniles whose parents have been arrested in the year prior to the participated in a greate r diversity of delinquency than juveniles whose parents had not been arrested.

PAGE 62

62 This was not supported since the imitation/modeling of parent deviant experiences was not significant in the model. Juveniles whose siblings have been arrested in the year prior to the incarceration will be significantly more likely to have also participated in greater diversity of delinquency than juveniles whose siblings hav e not been arrested This was not supported since the imitation/modeling of deviant experiences was not significant in the model. Of the variables that were significant f or the delinquency analysis in M odel 2 their relative importance (.i.e. the beta or standardized effect of each variable) is ordered as follows: differential associatio n peer delinquency (beta=.442) and definitions (beta= .264). C. Hypothesis 3 mediating social structures with social learning processes : The third hypothesis expects the effects of s ocial structure measures on the diversity of delinquency to be substa ntially mediated by social learning process var iables. The sub hypothesis for this count of involvement with different kinds of delinquency states: The effects of social structure measures on juvenile delinquency will be substantially reduced in models th at include social process measures This hypothesis was partially supported. The relevant results are derived from comparing the coefficients a nd probabilities presented for M odel 3 in the third column of figure s in Table 5 5 with those from M odel 1 (prese nted in the first column). Both the social structure and the social learning process measures are included in M odel 3 but only the social structure ones are presented in M odel 1. The hypothesis was partially supported since some but not all the structure v ariables yielded a substantial reduction in effect. The a ge variable was significant at the .1 level in Model 1 (beta=.161) but not significant in Model 3 (beta= .063). I ts effect had been absorbed by the social learning

PAGE 63

63 variables so the mediation hypo thesis is supported The race (Black) variable was significant at the .1 level in Model 1 (beta= .169) and became actually more significant (at the .05 level ) in Model 3 but its beta weight remained almost exactly the same (beta==.170) The effects of bei ng Black on delinquency were not mediated by the social learning variables. O verall M odel 3 which combines both the structural and social learning variables explains 32% of the variance in diversity of delinquency (adjusted R 2 =.323) and is significant a t the .01 level (F=6.317). The constant is .424 types of delinquency. Three of the independent variables were significant in predicting delinquency: race (Black), definitions, and differential association peer delinquency. Each of these can be interpreted. On average, juveniles who are Black decreased their diversity of delinquency by almost one additional type (b= .839 ), in comparison to other juveniles (Whites or Others) when all other variables held constant. On average, for every one unit increase in th e strength that juveniles agree that it is wrong to commit delinquent activities, the juveniles increase their diversity of delinquency in the past year by at least one additional type of delinquency (b= 1. 541 ), with all other variables held constant a fin ding contrary to social learning theory On average, for every one unit increase amount of friends who committed delinquent acts in the past year, juveniles increase their diversity of delinquency in the past year by at least one addition type of delinque ncy (b= 1.273 ), with all other variables held constant. Of the variables that were significant in this model, their relative importance (.i.e. the standardized effect of each variable) is ordered as follows: differential association peer delinquency (beta= .4 29 ), definitions (beta=. 312 ) and Black ( beta= .17 0 ).

PAGE 64

64 Table 5 1. Univariate Analysis for All Variables Variable N Mean Median Mode St. Dev Min Max Juvenile Marijuana Use 146 13.5753 8.5 0 14.1575 0 30 Juvenile Delinquency 146 3.3425 3 2 2.4648 0 10 Social Structure Variable for Both Models Differential Physical Location 146 .2945 0 0 .4574 0 1 Broken Home Indicator 146 .7808 1 1 .4151 0 1 Residential Stability 146 1.27 1 0 1.246 0 4 Race (Black) 146 .43 15 0 0 .497 0 0 1 Race (Other) 146 2123 0 0 .4 103 0 1 Sex (Male) 146 .8 3 00 1 1 .3 78 0 1 Age 146 15.9 726 16 16 1.2 537 13 18 Social Learning Process Variables for Marijuana Use Models Definitions 146 2.7100 3 3 .856 1 4 Diff Reinforcement 146 2.41 3 3 .722 1 3 Diff Assoc Peer MJ Use 146 3.77 4 5 1.275 1 5 Diff Assoc Peer Religious Behavior 146 2. 19 2 2 1. 2 1 5 Imitation/Modeling Parents 146 .1400 0 0 .352 0 1 Imitation/ Modeling Siblings 146 .2900 0 0 .457 0 1 Social Learning Process Variables for Delinquency Models Definitions 14 6 2.1 5 63 2.2 500 2 .25 .4 991 1 3. 75 Diff Reinforcement 14 6 2.8288 3 3 .3780 1 3 Diff Assoc Peers Delinquency 14 6 2.5616 2.4583 2.25 .8308 1 4.58 Imitation/Modeling Parents 14 6 .1 336 0 0 .2 442 0 1 Imitation/ Modeling Siblings 14 6 .29 0 0 .45 4 0 1

PAGE 65

65 Table 5 2 Variable (x1) (x2) (x3) (x4) (x5) (x6) (x7) (x8) (x9) (x10) (x11) (x12) (x13) 1. Diff Physical Location 1 2. Broken Home Indicator .088 1 3. Residential Sta bility .184** .050 1 4. Race (Black) .156* .201** .005 1 5. Race (Other) 0.068 .058 .052 .449*** 1 6. Sex (Male) .171** .074 .157* .050 .015 1 7. Age .017 .032 .002 .075 .008 .104 1 8. Definitions .082 .105 .076 .116 .096 .203** .023 1 9. Diff Reinforcement .014 .089 .043 .121 .110 .058 .001 .399*** 1 10. Diff Assoc Peer Drug Use .098 .049 .022 .006 .043 .016 .131 .175** .161** 1 11. Diff Assoc Peer Rel Beh .052 .025 .01 6 .212** .023 .186** .084 .088 .089 .287*** 1 12. Imi/Modeling Parents .121 .028 .161* .205*** .015 .013 .020 .050 .017 .104 .024 1 13. Imi/ Modeling Siblings .088 .021 .058 .117 .031 .026 .100 .059 .014 .162* .011 .078 1 14. Mariju ana Use .235*** .104 .043 .002 .069 .197** .058 .232*** .239*** .473*** .174** .186** .125 ***p < 0.01 ** p < 0.05 *p < 0.10

PAGE 66

66 Table 5 3. Bivariate Analysis of Delinquency Model Variables Variable (x1) (x2) (x3 ) (x4) (x5) (x6) (x7) (x8) (x9) (x10) (x11) (x12) (x13) 1. Diff Physical Location 1 2. Broken Home Indicator .088 1 3. Residential Stability .184** .050 1 4. Race (Black) .1 5 6 201 ** .0 05 1 5. Race (Other) 06 8 .0 58 .0 52 .449 *** 1 6. Sex (Male) .171 .0 74 .157 .0 50 .01 5 1 7. Age .01 7 .0 32 .00 9 .0 75 .00 8 104 1 8. Definitions .0 36 .162 .1 02 .12 9 .0 75 .144 .1 27 1 9. Diff Reinforcement .1 87** .0 06 .01 7 .01 4 .0 69 .164* .1 73 ** .3 8 2 *** 1 10. Diff Assc Peer Del .1 28 .07 9 .0 29 .03 7 .0 24 .0 95 .158 ** .0 59 .163** 1 11. Diff Assc Peer Rel Beh .041 .030 .008 .205** .0 32 .195* .076 .0 43 .0 79 .188 ** 1 12. Imi/Modeling Parents .139 .053 .003 .0 54 .0 70 .108 .03 9 123 .0 49 .151 112 1 13. Imi/ Modeling Siblings .05 5 .015 .0 86 .0 35 .0 68 .288 *** .0 55 .0 42 .01 2 .0 77 .0 85 .0 15 1 14. Delinquency .102 .033 .0 35 191* 109 .036 .1 80* 318** 235*** .482* ** 172 0. 101 .0 20 ***p < 0.01 ** p < 0.05 *p < 0.10

PAGE 67

67 Table 5 4. Multivariate OLS Regression Predicting Juvenile Marijuana Use Independent Variables Model 1 Model 2 Model 3 b B SE b B SE b B SE Social Structures Differential Social Location Family Structur e Differential Physical Lodging 6.964 / .233 [2.490]*** ----6.415 / .214 [2.256]*** Broken Home Indicator 4.630 / .140 [2.700]* ----2.818 / .085 [2.390] Residential Stability 1.279 / .116 [.900] ----.900/ / .082 [.801] Different ial Locati on in the Social Structure Race (Black) 1.538 / .056 [2.547] ----.374/ / .014 [2.479] Race (Other) 3.709 / .110 [3.023] ----1.280 / .038 [2.728] Sex (Male) 7.089 / .199 [2.948]** ----5.398/ / .151 [2.729]* Age .896 / .053 [.872] ----.197/ / .018 [.782] Social Processes Definitions ----1.832 / .115 [1.266] .954 / .060 [1.264] Differential Reinforcement ----2.320 / .120 [1.496] 2.519 / .133 [1.465]* Diff Assoc Peer MJ Use ----4.317 / .402 [.834]*** 4.060 / .378 [.829]*** Diff Assoc Peer Rel Beh ----.396/ / .035 [.891] .334 / .029 [.904] Imitation/ Modeling Parents ----5.594 / .144 [2.821]** 6.153 / .158 [2.916]** Imitation/Modeling Siblings ----1.205 / .040 [2.188] .960 / .032 [2.170] Constant 8.325 7.550 2.466 Model Fit F 5.833*** 9.173*** 5.688*** Adj. R square .081 .253 .296 N 146 146 146 Notes: Values shown for each model represented as follows: unstandardized coefficients / standardized coef ficients [standard error] ****p< 0.01 ** p< 0.05 p< 0.1

PAGE 68

68 Table 5 5. Multivariate OLS Regression Predicting Juvenile Delinquency Independent Variables Model 1 Model 2 Model 3 b B SE b B SE b B SE Social Structures Differential Social Loca tion Family Structure Differential Physical Lodging 523 / 097 [. 458 ] ----. 187 / .0 35 [. 405 ] Broken Home Indicator 411 / 069 [. 497 ] ----. 132 / .022 [. 425 ] Residential Stability .207 / .105 [. 166 ] ----.198 / .100 [. 141 ] Dif ferential Location in the Social Structure / / Race (Black) 837 / .169 [ .469 ] ----. 839 / 170 [. 414 ] ** Race (Other) 139 / .0 23 [. 556 ] ----. 210 / 035 [. 477 ] Sex (Male) .205 / 032 [. 542 ] ----. 447 / 070 [. 501 ] Age 317 / .1 63 [. 161 ] ----. 112 / / 063 [. 140 ] Social Processes Definitions ----1 302 / .264 [. 376 ] *** 1.541 / 312 [. 383 ] *** Dif Reinforcement ----. 456 / 070 [. 502] 283 / 043 [. 510 ] Diff Assoc Peer Delinquency ----1 312 / 442 [. 2 17 ] *** 1 273 / 429 [. 219]*** Diff Assoc Peer Rel Beh ----. 169 / .0 82 [. 148 ] 127 / / .062 [. 156 ] Imitation/ Modeling Parents ----. 112 / .011 [ .719 ] 119 / / .0 12 [. 719 ] Imitation/Modeling Siblings ----. 102 / .0 19 [. 379 ] 168 / / 031 [. 392 ] Constant 8.155 111 9 .424 Model Fit F 1. 864* 1 1 197*** 6.317*** Adj. R square .0 40 297 .3 23 N 146 146 146 Notes: Values shown for each model represented as follows: unstandardized coefficients / standardized coefficient s [standard error] ****p< 0.01 ** p< 0.05 p< 0.1

PAGE 69

69 CHAPTER 6 DISCUSSION AND CONCLUSIONS Although it is important to understand the results from a research study, it is equally important to understand those results in light of the theoretical paradigm a pplied. For the current study, the researcher Social Structure Social Learning Theory (SS SL) guides how the results can be understood by looking at what the theory suggests and what other researchers have found. Marijuana Use Models For the mari juana use models, some support was found for the hypotheses; at most the results suggest partial support for the independent variables predicting marijuana use. The first hypothesis examined i n Model 1 posited that social structure variables would relate t o marijuana use. O nly some social structure measures (broken home indicator and sex) were significant and performed in the direction suggested by theory. The other social structure variables did not hold up sugges t ing either that poor social structure me asures were selected or that family structure variables do not operate as anticipated for this sample of serious delinquents who have been placed in moderate and high risk residential treatment facilities. With the few tests of SS SL theory that have used family structures, the current research found much less significance tha n other studies ( Hoffman and Johnson 1998 ; Page 1998 ; Akers 2009 ) This could be due to several reasons, including different operationalization of family structures a smaller sample size, or the incarcerated status of the juveniles in the sample. The second hypothesis examined i n Model 2 posited that social learning process variables would predict marijuana use. O nly some of the social learning process

PAGE 70

70 measures (differential association peer marijuana use and imitation/modeling parents) were significant and performed in the direction of the theory. The other social learning process measures did not hold up as SS SL theory suggest s In most tests of social learning processes, definitions and differential association are the strongest predictors of the dependent variable. From the univaria te and the bivariate analyses, it can be seen that limited variation exists in the responses for differential reinforcement since the mean, median, and mode are all about equal to 3. This helps explain why differential reinforcement is not significant in the model and may reflect the cognitive behavioral therapeutic environment that the juveniles were in during incarceration wh en they were interviewed. The cognitive behavioral programming, especially Thinking for a Change, was designed to help youth identi fy thoughts and emotions that were paired (i.e., reinforced) with risky behaviors so the learned patterns could be disrupted and changed. That programming may have als o affected the definitions component of social learning theory. In the current research the differential association peer marijuana use variable comes in, but not the differential association religious behavior. This suggests that the specificity of friends performing actions that are being predicted about juveniles works better than gener al measures of behaviors. The imitation/modeling parent drug use variable adds support to the claim that peers are not the only important group to be aware of. The imitation/modeling sibling drug use variable was not significant in the model, but this may be d ue to the fact that their role might not be different enough from parents and peers to become significant. This is supported by the lack of zero order correlation between imitation/modeling siblings and marijuana use.

PAGE 71

71 Model 3 examined the mediation hyp othesis of SSSL. O nly one of the social structures (broken home indicator) was substantially mediated. Previous research on SS SL, has found support for the mediation hypothesis, but not all variables are mediated well ( Hwang and Akers 2006 ; Akers 2009 ) Another explanation for the partial support of the hypotheses predicting marijuana use is that the dependent variable had a high level cases that reported no marijuana use (coded 0) or daily marijuana use (coded 30), indica ting that the distribution of the dependent variable was bimodal which violates an assumption of O rdinary L east S quares R egression To adjust for this, a logistic regression was run using a dichotomous dependent variable for marijuana use (coded 1 for used marijuana in the 30 days prior to being incarcerated and 0 for no marijuana use in the 30 days prior to being incarcerated). The main difference between the results of the Ordinary Least Squares Regression and the Logistic Regression is that differential reinforcement is significant (at the .01 level) Specifically, this is seen in the Logistic Regression models predicting marijuana use using social learn ing process variables (Model 2) and using social structure and social learning process variable (Model 3). Another difference between the results of the two types of regression can be seen in Model 3. When using Logistic Regression, differential physical l odging is the only social structure that significan tly predicted marijuana use, whereas in Ordinary Least Squares differential physical lodging was significant, but other social structure measures were also significant and one structural variable (broken h ome indicator) was mediated. Diversity of Delinquency Models For the diversity of delinquency models, some support was found for the hypotheses A s with the marijuana use models the results suggest partial support for

PAGE 72

72 the independent variables predicti ng diversity of delinquency. In Model 1, testing the effects of the structural variables, only some social structure measures (Black and Age) were significant and performed in the direction of the theory. The other social structure variables did not hold u p, as was also seen in the marijuana use models. As before, this suggests either that poor social structure measures were selected or that family structure variables do not operate as anticipated for this sample of serious delinquents who have been placed in moderate and high risk residential treatment facilities. The reasons mentioned in the discussion of the marijuana use models (different operationalizations and a small sample of incarcerated juveniles) hold true here, as well. In Model 2, examining the impacts of the social learning process variables, only some of the social learning measures (definitions and differential association peer delinquency) were significant and performed in the direction of the theory. The other social process measures did not hold up as SS SL theory suggests. Just as with the marijuana use models, differential reinforcement had limited variability. As for the imitation/modeling parent and sibling variables, the lack of significance may be due to the small amount of people in the sample who had a parent (13%) or a sibling (29%) who were arrested in the year prior to the juveniles incarceration. The mediation hypothesis was tested in Model 3 O nly one of the social structures (age) was substantially mediated. As previously men tioned, the lack of mediation has been experience in other research using SS SL, but may be related to issues with the sample or the measures. In the current research, race (Black) was not mediated by social learning processes. The social learning process vari ables remain highly significant. Overview

PAGE 73

73 This research sought to answer three research questions 1. Do family structures explain marijuana use and diversity of delinquency among this group of offenders? 2. Do variables reflecting the social learning process account for marijuana use and diversity of delinquency among this group of offenders? 3. Will the impact of the family structure variables be mediated by the social learning variable among this group of offenders? The independent variables (social structure and social learning processes) are all focused around important groups (family and peers) that affect juvenile behavior. Specifically, p arents, siblings, and peers were selected because of their important role in both the structural and process factors affecting juvenile marijuana use and diversity of delinquency Social s tructures at the group level (differential physical lodging, broken home indicator, and residential stability) and location of individual s within those groups (age, sex, and rac e) provide a context within which social interactions and social process occur Social processes (definitions, differential reinforcement, differential association, and imitation/modeling) are sculpted by interactions within the same groups (parents, sibli ngs, and peers). Two dependent variables (juvenile marijuana use and diversity of delinquency ) were selected in this sample incarcerated Marijuana is a popular drug among many age groups, including juveniles regardless of whether or not they have a histo ry with the legal system Additionally, the activities that represent the diversity of delinquency range from minor to serious offenses, covering a span of possible options. Each of these dependent variables was used to test three hypotheses, respectively

PAGE 74

74 Research Limitations There are several important limitations that should be considered before attempting to generalize the findings of this study. First, the sample is made up of a s mall number of cases (n=146) in an unusual population (incarcerated juve niles) Second, the sample was purposive with non randomized selection of juvenile offenders faced incarceration and volunteered to participate in the evaluation. Third, this sample is n ot representative of the popu lation of all juvenile offenders since th e juveniles included in the study had done something severe enough to be place in a residential facility Fourth, this sample was receiving treatment at the time of data collection that may have affected responses especially to the social learning proces s items. Lastly, th ese data are collected at one point in time (i.e., are in that sense cross sectional questions asked for current and retrospective responses. For example, the definitions favorable and unfavorable to marijuana and delinquency were asked as of the time of their residential treatment but the questions about m arijuana use and delinquent behavior and associations were targeted before their commitments. Thus, these data pose potential problems of temporal ordering This leads the researcher to a ssume that parent deviant behavior precedes peer deviant behavior in affecting juvenile deviant behavior and that being incarcerated may have impacted certain variables (such as definitions and differential reinforcement) since the environme nt molds the juveniles to have unfavorable opinions about deviant behavior. Future Research If this research was to be continued with this dataset, there are several things that should be considered in future analyses. First, the decision of what to do mis sing values is difficult and other options (besides median substitution of missing values) could be

PAGE 75

75 considered to improve the analysis. Second, instead of OLS multivariate regressions, using poisson regression may provide more accurate results since it is better suited to predict probabilities of dependent variables that are counts like the ones used in the current study. If this research was to be continued with a new research project, a sample of non incarcerated juveniles would be included so that a com parison could be made between offenders and non offenders. In addition to the interview questions included in the current study, more specific questions on drug use and deviant behavior would be asked of parents, siblings, and peers. With more in depth inf ormation and a larger samples size the research would be better able extrapolate to a larger population, as well as dig deeper into how the structures and process affect different segments of the sample ( e.g. offender vs. non offender, males vs. females, and younger vs. older). Theoretical Implications Overall, the current research provide s moderate support for Social Structure Social Learning theory The social process variables had significant bivariate correlations (direct effects) and although not all the variables were significant in the multivariate analyses (net effects) it is expected for some of the variables to drop out because of the interrelation expected by the theory. In regard to the social structure variables, the theory does not suggest th at specific structures will be significant in the dimensions selected, but it does predict mediation of social structure variables in the presence of social process v ariables. This aspect of the theory did not hold up well This research shows clear suppor t for social learning theory even though it was tested on an extreme group of juvenile offenders with models explaining above 25% of the variance in the dependent variables expected beyond chance

PAGE 76

76 P olicy I mplications Despite the limitation s of the current research, there are still implications that policy makers should consider. This study highlights the importance of taking a holistic strong effect on their behavior, there are other primary groups that deserve consideration like parents and siblings. If the goal is to reduce juvenile deviant behaviors, then programs may need to incorporate how groups and processes beyond those involving peers to have optimal success.

PAGE 77

77 APPEND IX A VARIABLE CONSTRU CTION Variable Type Construct [Question #] Questions Coding Recode DV juvenile drug use days [137d] How many days in the 30 days before you were committed, did you use marijuana? enter # of days The zeros from 137a are reclaimed as z eros in 136d DV Juvenile Delinquency Scale [157 165,167, 169 172] # of times in the past year that you (a): Damaged or set fire to school property Damaged or set fire to other property Broke into a house, building, or car in order to take something (wh en no one was home) robbed or held up a place of business (store, gas station, bank taxi, etc. ) Robbed a person (street, robbery, mugging, purse snatching, hold up in a house or car; exclude business robberies) Tried to beat up somebody or threaten some one with a weapon (tried to shoot, stab, cut, beat, strangle, or strong arm someone, even if no one was hurt Stole or boosted something (stole from a till, shoplifted, picked pockets, or took something w/o owner's knowledge; exclude Stole a car, truck, o r motorcycle Used check or credit cards illegally (forgery, used a stolen or bad c redit card, passed a bad check) Ran cons or scams (defrauded p erson, business or government) Provided sex for money (prostitution) Committed or attempted homicide Commi tted or attempted sex by force (RAPE) Been in gang or posse fights Possessed marijuana or hashish possessed hard drugs such as cocaine crack heroin, PCP, and LSD enter # Each type of delinquency was coded 1=committed act 1 or more times and 0= never com mitted act. A diversity index was created from the sum of dichotomous recodes from the count of types of delinquency IV Differential Physical Lodging Before you entered this facility, where were you living/staying?: 1=your own house 2=your parent'(s ) 0 =live with in parent'(s) house/home 1 = not living

PAGE 78

78 Variable Type Construct [Question #] Questions Coding Recode [3] house/ home 3= another relative's house or apartment 4=a friend's house or apartment 5=an hotel 6=a rooming or boarding house, or halfway house 7=a shelter or welfare boarding home 8=a foster home 9=on the streets (abandoned building, vacant lot, park, ho meless) 10at a correction institution 11=other with parents IV Broken Home Indicator [5] Are your natural parents married to each other? 1=yes 0=no 8=DK 9=RF Reverse coded so that 1= parents not married/ broken home 2= parents married/ intact home IV Residential Stability [14] How many times did you move in the year before you were committed to this facility? # of times IV Age What is your age? enter age IV Sex Youth's sex 1=male 0 =female IV Race [2] What is your race/ethnicity? Asian American=1, American Indian=2, Black/ African American=3, Cambodian=4, Chinese=5, Cuban=6, Filipino=7, Guamanian=8, Hispanic (not Cuban)=9, Dummy variables are created for black, whit e and other. White is the reference group

PAGE 79

79 Variable Type Construct [Question #] Questions Coding Recode Japanese=10, Laotian=11, Pacific Social Learning Process Variables for MJ Use IV Definitions [202] For each of the statements I read here, please tell me how much you agree or disagree: It is wrong for teenagers to use marijuana 1= strongly disa gree 2=disagree 3= agree 4= strongly agree 8=DK 9= RF IV Differential Reinforcement [211] Whether or not you have ever used each of the following what usually happens to a teenager who uses. Marijuana 1=Mainly good, gain more than you lose 2=About as much good as bad 3=Mainly bad, lose more than you gain 8=DK 9=RF IV Differential Association Peer Drug Use [94] During the year before you were committed to this facility, to the best of you r knowledge, about how many of your best friends, or the friends you spent the most time with, used marijuana 1=none 2=some 3=half 4=most 5=all 8=DK 9=RF IV Differential Association Peer Religious Behavior [98] During the year before you were committe d to this facility, to the best of your knowledge, about how many of your best friends, or the friends you spent the most time with, had: Regularly taken part in church or religious activities 1=none 2=some 3=half 4=most 5=all 8=DK 9=RF IV Imitation Par ent [51] During the year before you were committed to the facility, did either or both of your parents/ guardians use illegal drugs? 1=yes 0=no 8=DK 9=RF IV Imitation Sibling (s) [53] During the year before you were committed to the facility, did any o f your brothers and sisters use illegal drugs? 1=yes 0=no 8=DK 9=RF Social Learning Process Variables for Delinquency IV Definitions [204 208] For each of the statements I read here, please tell me how much you agree or disagree: It is wrong to beat up threaten or use a weapon to hurt someone It is wrong to break the law It is all right to break the law if you are not really hurting anyone 1= strongly disagree 2=disagree 3= agree 4= strongly agree 8=DK 9= RF A 5 item scale was created from the average score of the items

PAGE 80

80 Variable Type Construct [Question #] Questions Coding Recode It is sometimes justified or necessary to break the law IV Differential Reinforcement [214, 216 217] Whether or not you have ever done it, what do you think usually happens to a teenager who does the following: Tries to or beats somebody up or threatens someone with a weapon Deals or delivers drugs Steals money or property Breaks into a house, building or car 1=Mainly good, gain more than you lose 2=About as much good as ba d 3=Mainly bad, lose more than you gain 8=DK 9=RF A 3 item scale was created from the average score of the items IV Differential Association Peer Drug Use [103 110, 112 115] During the year before you were committed to this facility, to the best of your knowledge, about how many of your best friends, or the friends you spent the most time with, had: Been suspended from school Been picked up (not just stopped) by the police (ex: were not free to leave) Appeared in juvenile court for a crime Been put i n juvenile detention Stolen things worth a lot of money Stolen a car, truck or motorcycle Tried to or beat somebody up or threatened someone with a weapon Dealt or delivered drugs Stole small things not worth much money Been put into a juvenile facility or program Been on probation Been transferred to adult court 1=none 2=some 3=half 4=most 5=all 8=DK 9=RF A 12 item scale was created from the average score of the items IV Differential Association Peer Rel igious Behavior [98] During the year before you were committed to this facility, to the best of your knowledge, about how many of your best friends, or the friends you spent the most time with, had: Regularly taken part in church or religious activities 1=none 2=some 3=half 4=most 5=all 8=DK 9=RF IV Imitation Parent [65,75] During the year before you were committed to this facility, was your mom/ female guardian arrested During the year before you were committed to this facility, was your dad/ male g uardian arrested 1=yes 0=no 8=DK 9=RF A new variable was created from the 2 questions so that 1=either parent had been arrested and 0=neither parent had been arrested IV Imitation Sibling (s) During the year before you were committed to this facilit y, were any of your brothers and/ or sisters (including 1=yes 0=no 8=DK 9=RF

PAGE 81

81 Variable Type Construct [Question #] Questions Coding Recode [80] natural or step): arrested

PAGE 82

82 LIST OF REFERENCES Akers, R. (1985). Deviant Behavior Belmont, California, Wadsworth Publishing Company. Akers, R. (1990). "Rati onal Choice, Deterrence, and Social Learning Theory in Criminology: The Path Not Taken." The Journal of Criminal Law and Criminology : 653 676. Akers, R. (1998). Social Learning and Social Structure: A General Theory of Crime and Deviance Boston, MA, North eastern University Press. Akers, R. (2009). Social Learning and Social Structure: A General Theory of Crime and Deviance Boston, MA, Northeastern University Press. Akers, R. and J. Cochran (1985). "Adolescent Marijuana Use: A Test of Three Theories of Dev iant Behavior." Deviant Behavior : 323 346. Akers, R. and G. Jensen (2003). Social Learning Theory and the Explanation of Crime New Brunswick, New Jersey, Transactions Publishers. Akers, R. and G. Jensen (2006). The Empirical Status of Social Learning Theo ry of Crime and Deviance: the Past, Present, and Future. Taking Stock: The Status of Criminological Theory New Brunswick, New Jersey, Transaction Publisher : 37 76. Akers, R., M. Krohn, et al. (1979). "Social Learning and Deviant Behavior: A Specific Test of a General Theory." American Sociological Review : 636 655. Akers, R., A. La Greca, et al. (1989). "Social Learning Theory and Alcohol Behavior Among the Ederly." The Sociological Quarterly : 625 638. Akers, R. and G. Lee (1999). "Age, Social Learning, and Socia Bonding in Adolescent Substance Use." Deviant Behavior 19 : 1 25. Akers, R. and C. Sellers (2004). Criminological Theories: Introduction, Evaluation, and Application Los Angeles, CA, Roxbury Publishing Company. Andrews, J., H. Hops, et al. (1993). Parental Influence on Early Adolescent Substance Use: Specific and Non Specific Effects." Journal of Early Adolescence : 285 310. Aseltine, R. (1995). "A Reconsideration of Parent and Peer Influences on Adolescent Deviance." Journal of Health and Social Beh avior : 103 121. Bahr, S., J. Hoffmann, et al. (2005). "Parental and Peer Influences on the Risk of Adolescent Drug Use." The Journal of Primary Prevention 26 (6): 529 551.

PAGE 83

83 Bahr, S., A. Marcos, et al. (1995). "Family, Educational and Peer Influences on the A lcohol Use of Female and Male Adolescents." Journal of Studies on Alcohol 457 469. Barnes, G. and J. Welte (1986). "Patterns and PRedictors of Alcohol Use among 7 12th Grade Students in New York State." Journal of Studies on Alcohol : 53 62. Benda, B. (1994 ). "Testing Competing Theoretical Concepts." Deviant Behavior : 375 396. Bishop, D. (2000). "Juvenile Offenders in the Adult Criminal Justice System." Crime and Justice : 81 167. Botvin, G., K. Griffin, et al. (2000). "Preventing Illicit Drug Use in Adolesce nts: Long Term Follow Up Data from a Randomized Control Trial of School Population." Addictive Behaviors 25 (5): 769 774. Brook, J., D. Brook, et al. (2006). Risk and Protective Factors of Adolescent Drug Use: Implications for Prevention Programs. Handbook of Drug Aguse Preventions Springer US : 265 286. Brook, J., M. Whiteman, et al. (1990). "The Role of Older Brothers in Younger Brothers' Drug Use Viewed in the Context of Parent and Peer Influences." The Journal of Genetic Psychology : 59 75. Brook, J., M. Whiteman, et al. (1986). "Some Models and Mechanisms for Explaining the Impact of Maternal and Adolescent Characteristics on Adolescent Stage of Drug Use." Developmental Psychology : 460 467. Burgess, R. and R. Akers (1969). A Differential Association Reinf orcement Theory of Criminal Behavior. Delinquency, Crime, and Social Processes New York, NY, Harper & Row Publishers : 531 556. Catalano, R., R. Kosterman, et al. (1996). "Modeling the Etiology of Adolescent Drug Use: A Test of the Social Development Mode l." Journal of Drug Issues : 429 455. Chassin, L. (2008). "Juvenile Justice and Substance Use." The Future of Children 18 (2): 165 183. Dorius, C., J. Bahr, et al. (2004). "Parenting Practices as Moderators of the Relationship between Peers and Adolescent Ma rijuana Use." Journal of Marriage and Family : 163 178. Gottfredson, M. and T. Hirschi (1990). A general theory of crime Stanford University Press. Greenwood, P. (1996). "Responding to Juvenile Crime: Lessons Learned." The Future of Children : 75 85.

PAGE 84

84 Hirsch i, T. (1969). Causes of Delinquency Berkeley, CA, University of California Press. Hoffman, J. and R. Johnson (1998). "A National Portrait of Family Structure and Adolescent Drug Use." Journal of Marriage and Family : 633 645. Hoffman, J. and S. S. Su (1998 ). "Parental Substance Use Disorder, Mediating Variables and Adolescent Drug Use: A Non recursive model." Addiction : 1351 1464. Holland Davis, L. (2006). Putting Behavior in Context: A Test of Social Structure Social Learning Model. Sociology Gainesville, Florida, University of Florida. PhD Dissertation Huizinga, D., R. Loeber, et al. (1995). Urban Delinquency and Substance Abuse. Initial Findings. Research Summary. Washington, D.C., Department of Justice: Office of Juvenile Justice and Delinquency Preven tion. Hwang, S. and R. Akers (2006). "Parental and Peer Influences on Adolescent Drug Use in Korea." Asian Criminology : 51 69. Johnston, L., P. O'Malley, et al. (2004). Monitoring the Future National Results on Adolescent Drug Use: Overview of Key Findings 2004. Bethesda, MD, National Institute on Drug Abuse. Kandel, D. (1980). "Drug and Drinking Behavior Among Youth." Annual Reviw of Sociology : 235 285. Kaplan, H., S. Martin, et al. (1984). "Pathways to Adolescent Drug Use: Self Derogation, Peer Influence s, Weakening of Social Controls, and Early Substance Use." Journal of Health and Social Behavior : 270 289. Khavari, K. (1993). "Interpersonal Influences in College Student's Initial Use of Alcohol and Drugs: The Role of Friends, Self, Parents, Doctors, and Dealsers." Substance use and misuse : 377 288. Krohn, M. (1999). "Social Learning: The Continued Development of a Perspective." Theoretical Criminology : 462 476. Krohn, M., L. Lanza Kaduce, et al. (1984). "Community Context and Theories of Deviant Behavior : An Examination of Social Learning and Social Bonding Theories." Sociology Quarterly : 353 372. Krohn, M., W. Skinner, et al. (1985). "Social Learning Theory and Adolescent Cigarette Smoking: A Longitudinal Study." Social Problems 32 (5): 455 473. Krohn, M. W. Skinner, et al. (1989). "Elaborating the Relationship Between Age and Adolescent Cigarette Smoking." Deviant Behavior : 105 129.

PAGE 85

85 Lane, J., L. Lanza Kaduce, et al. (2009). Final Report of the Florida Faith and Community Based Delinquency Treatment Intia tive (FCBDTI) Evaluation. Lanza Kaduce, L., R. L. Akers, et al. (1982). "Conceptual and Analytical Models in Testing Social Learning Theory: Reply." American Sociological Review 47 (1): 169 173. Lanza Kaduce, L. and M. Capece (2007). Social Structure Socia l Learning (SSSL) and Binge Drinking: A Specific Test of A General Theory. Social Learning Theory and the Explanation of Crime Transaction Publisher. 11: 179 196. Lanza Kaduce, L., M. Capece, et al. (2006). "Liquor is Quicker: Gender and Social Learning A mong College Students." C riminal Justice Policy Review 17 (2): 127 143. Lee, G., R. Akers, et al. (2004). "Social Learning and Structural Factors in Adolescent Substance Abuse." Western Criminology Review 5 (1): 17 34. Mallett, S., D. Rosenthal, et al. (2005 ). "Young people, drug use and family conflict: Pathways into homelessness." Journal of Adolescence : 185 199. McDermott (1984). "The Releationship of Parental Drug Use and Parents' Attitude Concerning Adolescent Drug Use to Adolescent Drug Use." Adolescenc e XIX (73): 89 97. Moon, D., M. Hecht, et al. (1999). "Ethnic and Gender Differences and Similarities in Adolescent Drug Use and Refusals of Drug Offers." Substance use & misuse : 1059 1083. Newcomb, M., E. Maddahan, et al. (1986). "Risk Factors for Drug Use among Adolescents: Concurrent and Longitudina Analysis." American Journal of Public Health 75 (5): 525 531. Page, E. R. (1998). Family Structure and Juvenile Delinquency: The Mediating Role of Social Learning Variables. Department of Criminology Gainesvil le, FL, University of Florida. PhD Dissertation Park, R., E. Burgess, et al. (1984). The city University of Chicago Press. Penning, M. and G. Barnes (1982). "Adolescent Marijuana Use: A Review." Substance use & misuse : 749 791. Petraitis, J., B. Flay, et al. (1995). "Reviewing Theories of Adolescent Substance Use: Organizing Pieces in the Puzzle." Psychological Bulletin : 67 86. Piko, B. and E. Kovcs (2010). "Do Parents and School Matter? Protective Factors for adolescent Substance Use." Addictive Behavio rs 35 : 53 56.

PAGE 86

86 Platt, A. (1977). The child savers: the invention of delinquency University of Chicago Press. Reed, M. and P. Wilcox Rountree (1997). "Peer Pressure and Adolescent Substance Use." Journal of Quantitative Criminology : 143 180. Rohde, P., C. K ahler, et al. (2004). "Psychiatric disorders, familial factors, and cigarette smoking: II. Associations With Progression to Daily Smoking." Nicotine & Tabacco Research : 119 132. Simcha Fagan, O., J. Gersten, et al. (1986). "Early Precursors and Concurrent Correlates of Patterns of Illicit Drug USe in Adolescent." The Journal of Drug Issues : 7 28. Sprott, J. (1996). "Understanding Public Views of Youth Crime and the Youth Justice System." Canadian Journal of Criminology : 271 290. Stice, E. and J. Barrera, Ma nuel (1995). "A Longitudinal Examination of the Reciprocal Relations Between Percieved Parenting and Adolescents' Substance Use and Externalizing Behaviors." Developmental Psychology : 322 334. Stormshak, E., C. Comeau, et al. (2004). "The Relative Contribu tion of Sibling Deiance and Peer Deviance in the Prediction of Substance Use Across Middle Childhood." Journal of Abnormal Child Psychology : 635 649. Sutherland, E. (1939). Criminology Philadelphia, PA, Lippincott. Svensson, R. (2000). "Risk Factors for D ifferent Dimensions of Adolescent Drug Use." Journal of Child & Adolescent Substance Abuse : 67 90. Terry, Y., C. Vanderwaal, et al. (2000). "Provision of Drug Treatment Services in the Juvenile Justice System: A System Reform." The Journal of Behavioral He alth Services & Research : 194 214. Triplett, R. and B. Payne (2004). "Problem Solving as Reinforcement in Adolescent Drug Use: Implications for Theory and Policy." Journal of Criminal Justice 32 : 617 630. Trost, M., E. Langan, et al. (1999). "Not Everyone Listens When You "just say no": Drug Resistance in Relational Context." Journal of Applied Communication Research : 120 138. Velleman, R., L. Templeton, et al. (2005). "The Role of Family in Preventing and Intervening with wubstance use and misuse: a compre hensive review of family interventions, with a focus on young people." Drug and Alcohol Review 24 : 93 109.

PAGE 87

87 Visher, C. A. (1995). Carrer Offenders and Crime Control. Criminology: a contemporary handbook Belmont, Wadsworth : 515 534. Windle, M. (2000). "Pare ntal, Sibling, and Peer Influences on Adolescent Substance Use and Alcohol Problems." Applied Developmental Science : 98 110.

PAGE 88

88 BIOGRAPHICAL SKETCH Danielle Tolson was born and raised in St. Louis, Missouri and is the second child of her parents, Fred and Ste phanie Tolson. She attended Truman State University in Kirksville, Missouri where she achieved a B.S. in justice systems, with a minor in business a dministration. After graduation from undergraduate in 2009, she started as a Research Assistant at the Unive rsity of Florida in the D epartment of Sociology and Criminology & Law As of May 2011, she will have achieved a MA in c riminology and will continue on to pursue a PhD in c riminology at the University of Florida.