1 PLANNING AND EVALUATION OF A PRIMARY GRADE VIOLENCE PREVENTION PROGRAM By HEATHER CHRISTIAN A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2007
2 2007 Heather Christian
3 TABLE OF CONTENTS page LIST OF TABLES................................................................................................................. ..........6 ABSTRACT....................................................................................................................... ..............8 CHAPTER 1 LITERATURE REVIEW.........................................................................................................9 Statement of the Research Problem..........................................................................................9 Review of Related Literature..................................................................................................12 Current Conceptualiz ation of Violence...........................................................................12 Outcomes of Youth Aggression and Violence................................................................15 Role of Prevention Science in Combating Youth and School Violence................................16 Assumptions of Youth Violence Prevention...................................................................16 Goal of Youth Violence Prevention Efforts....................................................................17 Introduction to Risk Factors............................................................................................18 Risk Factors and Interventions........................................................................................19 Individual risk factors...............................................................................................19 Community/Environment.........................................................................................21 Peers.........................................................................................................................22 School.......................................................................................................................22 Family.......................................................................................................................23 Introduction to Protective Factors...................................................................................24 Empirically Supported Protective Factors.......................................................................25 Individual protective factors.....................................................................................25 Peer-related protective factors..................................................................................25 School-related protective factors..............................................................................26 Family protective factors..........................................................................................26 Theory Related to Violent Behaviors.....................................................................................27 Behavioral Theory...........................................................................................................27 Cognitive Behavioral Skills Training Theory.................................................................28 Attribution Theory...........................................................................................................28 Social Learning Theory...................................................................................................29 Social Control Theory.....................................................................................................29 Social Information Processing Theory............................................................................29 Review of Violence Preventi on Interventions and Programs.................................................31 Program Chosen for the Present Study...................................................................................32 Violence Prevention Program Planning and Evaluation........................................................43 Pre-Planning Methodology to Promote In tervention and Evaluation Fidelity................44 Planning Classroom-Based Preventive Interventions: Evaluative Focus........................44 Determine the focus of the intervention...................................................................45 Strategy and plan for implementa tion and interven tion fidelity..............................45 Rationale for the Study........................................................................................................ ...47
4 Type of Intervention and Needs of Students, Teachers, and Schools.............................47 Department of Education Criteria...................................................................................49 Theory......................................................................................................................... .....49 Opportunities for Collaboration......................................................................................50 Use of an Experiential Add-On.......................................................................................50 Use of Group Observational Method..............................................................................51 Significance of the Study...................................................................................................... ..52 Exploring the Effectiveness of a Cl assroom-Based, Universal Prevention....................52 Exploring the Effectiveness of an Experiential Add On.................................................53 Contributions to Practice.................................................................................................54 Evaluating Classroom-Based Viol ence Prevention Interventions..........................................55 Hypotheses..................................................................................................................... .........57 2 METHODS........................................................................................................................ .....66 Participants................................................................................................................... ..........66 Intervention/Treatment Groups..............................................................................................66 Procedures..................................................................................................................... ..........68 Curriculum..................................................................................................................... .........70 Intervention................................................................................................................... ..........71 Apparatus...................................................................................................................... ..........72 The PATHS Knowledge Survey.....................................................................................73 Best Acting Report of Social Behaviors..........................................................................76 Peer Nomination Survey..................................................................................................79 Social Skills Rating System (SSRS): Teacher Report and Student Report....................81 Best Acting Behavioral Observation Checklist...............................................................84 Statistical Analysis........................................................................................................... .......87 Hypotheses 1...................................................................................................................87 Hypothesis 2................................................................................................................... .88 Assumptions.................................................................................................................... .......88 3 RESULTS........................................................................................................................ .....106 Hypothesis 1................................................................................................................... ......106 Results of the Final Mulivari ate Ananlysis of Covariance...................................................106 Hypothesis 2................................................................................................................... ......107 4 DISCUSSION..................................................................................................................... ..115 Benefits to Students Participat ing in the Program/Hypothesis I..........................................115 Absence of Program Benefits?.............................................................................................116 Preexisting Risk Factors................................................................................................117 Individual risk factors.............................................................................................117 Family and environmental risk factors...................................................................118 Measurement Error and Difficulties M easuring Social Behavior Change....................118 Intervention Fidelity......................................................................................................121 Benefits to Students Participat ing in the Program/Hypothesis 2..........................................123
5 Potential Contributions a nd Areas of Development.............................................................124 Limitations.................................................................................................................... ........125 APPENDIX A PATHS KNOWLEDGE SURVEY......................................................................................128 B BEST ACTING REPORT OF SOCIAL BEHAVIORS.......................................................131 C PEER NOMINATION QUESTIONAIRE...........................................................................133 D BEST ACTING BEHAVIORAL OBSERVATION CHECKLIST.....................................134 LIST OF REFERENCES.............................................................................................................135 BIOGRAPHICAL SKETCH.......................................................................................................148
6 LIST OF TABLES Table page 1-1 Florida violence statistics for 2003 and 2004....................................................................58 1-2 Columbia County violen ce statistics for 2003 and 2004...................................................59 1-3 Intervention program summary..........................................................................................60 2-1 Students not include d in final sample................................................................................90 2-2 Students by treatment, gender, and race............................................................................91 2-3 Treatment levels........................................................................................................... ......92 2-4 State standardized testing results and free and reduced lunch percentage.........................93 2-5 Incidents of crime and violen ce in Florida elementary schools.........................................94 2-6 PATHS lessons.............................................................................................................. ....95 2-7 Data collection timeline................................................................................................... ..96 2-8 PATHS knowledge survey descriptives.............................................................................97 2-9 Percent correct on PATHS knowledge survey...................................................................98 2-10 PATHS knowledge survey descri ptives following item analysis......................................99 2-11 Best Action Social Behavi or Survey question content....................................................100 2-12 Best Acting Social Beha viors Survey descriptives..........................................................101 2-13 Peer nomination correlations: Spearmans rho...............................................................102 2-14 Descriptive statistics for SSRS reports............................................................................103 2-15 Pearson correlation coefficients for the SSRS data.........................................................104 2-16 Behaviors and definitions................................................................................................105 3-1 Data sources for the proposed MANCOVA....................................................................109 3-2 Descriptives for propos ed dependent variables...............................................................110 3-3 Descriptives for the dependent variables.........................................................................111 3-4 Frequency of aggression pre-and post-intervention.........................................................112
7 3-5 Frequency of prosocial beha vior pre-and post-intervention............................................113 3-6 Logistic regression predicting behavior al response from teacher assignment, treatment level, and school attended................................................................................114
8 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy PLANNING AND EVALUTION OF A PRIMARY GRADE VIOLENCE PREVENTION PROGRAM By Heather Christian August 2007 Chair: Thomas Oakland Major: School Psychology We implemented an 18-lesson problem-solv ing with 129 students in seven 4th grade classrooms, in two rural elementary schools. At one school, three classrooms received the curriculum plus video intervention and another classroom served as a no treatment control. At a second school, four classrooms received the cu rriculum without video intervention. Students completed measures assessing curriculum knowle dge, frequency of violence in school, peer nomination, and the Social Skills Rating System-s tudent form. Classroom teachers reported on individual childrens behavior using the Social Skills Rating System-teacher form. Observers blind to each of the three treatment conditions completed behavioral obs ervations. Aggressive and prosocial behaviors were tallied using an opportunity-d riven method for the entire classroom. Children who received curriculum and the curriculum with video did not demonstrate gains in knowledge and social ski lls and decreases in observed, teacher-rated, and student-reported violence. Results were examin ed in terms of impact on individual and group functioning.
9 CHAPTER 1 LITERATURE REVIEW Statement of the Research Problem In 1996, yearly costs of violence-related e xpenses, including medical treatment, lost productivity, rehabilitation, and dire ct costs to the United States ju stice system were estimated to exceed $60 billion (Clayton, Ballif-Spanvill, & H unsacker, 2001). In 1998, the Office of the Surgeon General estimated yearly costs associated with direct and indirect violence exceeded $425 billion each year (United Stat es Public Health Service, 2003). In 2000, suicides and homicides were the third and fourth leading cause s of death in the United States. From 1993 to 2004, youth violence and victimization rates dec lined (Centers for Di sease Control, 2005); however, North American youth die and are involved in serious violence at much higher rates than youth in all other industria lized countries (Furlong, Paige, & Osher, 2003). In Florida where this investigation occurre d, rates of violent crime have d eclined overall since the early 1990s. However, incidents of murder and aggr avated assault have risen (Table 1-1 for 20032004 statistics). In Columbia County, where the two intervention schools are located, rates of domestic violence have decreased overall (Table 1-2). However, the rate per 100,000 is much greater than for the Florida popul ation. The rates per 100,000 for the population of Florida were 707.0 in 2003 and 683.8 in 2004 (i.e., years of intervention and data collection). United States educational instituti ons have not been protected from the spread of violence. During the mid 1990s, educational institutions becam e the settings for multiple school shootings. Although homicides in schools are relatively rare (i .e., less than 1% of vi olent youth deaths occur in schools), over the past two decades school sh ootings have shifted considerable resources toward violence prevention efforts in an attempt to lower the rate of yout h and school violence. School shooting incidents during th e 1990s highlighted the need to prevent high inte nsity violent
10 acts as well as low-level aggressi on and violent climates in school s (Furlong et al., 2003; Weist, 2003). Following well-publicized school shootings, la wmakers and educators began to address youth violence through school safety initiatives. Lawmakers, administrators, and teachers expressed concern about the threat of school violence. Students also expressed fear concerning past and future violent acts in schools. Cons equently, researchers be gan investigating links between violence and/or the threat of violen ce in schools and school behaviors, including achievement. Exposure to violence has been found to have pervasiv e and profound effects on student development (Salzinger, Feldman, Stockhammer, & Hood, 2002). School violence generally negatively impacts students abilities to learn (Clayton et al ., 2001) and students ability and desire to learn is lessened when they are worried about their sa fety (Cirillo, Pruitt, & Colwell, 1998; Hintz & Shapiro, 1999). Additio nally, perpetration of violence contributes to underachievement, school drop out, poor career pe rformance, and committing future acts of physical and sexual abuse (C layton et al., 2001). School safety is an educational right, and these safety needs have been demonstrated to be paramount for academic achievement (Morris on, Furlong, & Smith, 1994). The identified link between school violence and achievement is partic ularly salient when contemplating the fact that children in the United States are required to attend school. Thus the promotion of school safety becomes a legal obligation (Hermann & Finn, 2002; Yell & Rozalski, 2000). The implementation of youth violence preven tion programs represents one method used by educators and mental health prof essionals to promote safe sch ools and the subsequent reduction in violence. For individuals, violent behaviors tend to remain stable over time. However, psychosocial and environmental buffers can ha ve a powerful effect on the development of
11 chronic violent behaviors (Dodge, 2001; Eddy, Reid, & Fetrow, 2000; Frey, Hirschstein, & Guzzo, 2000; Tolan & Gorman-Smith, 2002). Preven tion programs can be designed to inject buffers into the individuals environment to co mbat existing risk fact ors. Although many risk factors associated with violence are external to schools, the inte rnal school environment either can buffer or exacerbate environmental or intrai ndividual risk factors (Adelman & Taylor, 2003; Furlong, Paige, & Osher, 2003; Pettit, 2004). Add itionally, the importance of addressing the risk factors of violence in an atte mpt to reduce violence itself, as well as other co-occurring problematic adolescent behavi ors (e.g., drug use, school drop out, delinquency, and teen pregnancy) has been demonstrated (Barrios, 2001; Cunningham, 2000). The number of violence prevention programs for youth is estimated to exceed more than one thousand. However, the vast majority have not been evaluated (Ellickson & Mcguigan, 2000; Kerns & Prinz, 2002; DeVoe et al., 2003). Violence prevention ef forts have taken many forms and have targeted differe nt populations. For example, programs can target an entire population (universal or primary pr evention), only those most at ri sk for violence (selected or secondary prevention), or thos e individuals who already de monstrate violent behaviors (indicated or tertiary prevention) Intervention methods also diffe r in their underlying theory or learning method (e.g., classroom instruction, peer mediation, parenting program, group therapy, community education, or a combination of met hods). To date, some programs have shown positive results and are empirically supported (Dodge, 2001). The Early Warning Guide, published by the Unit ed States Department of Education in 1998, encouraged educators and administrators to implement evidenced based prevention programs as part of a safe school plan (Furlong et al., 2003). This report cited teaching positive interaction skills, specif ically social problem solving, as its first recommended strategy for
12 reducing the potential for violen ce among high-risk children (Hunt er, Elias, & Norris, 2001). A number of web resources as well as empirical reviews outline program effec tiveness. However, these resources often are based on a very small number of outcome studi es. Unfortunately, few published studies examine the effectiveness of universal violence prevention programs implemented with elementary-aged students (A ber, Brown, & Jones, 2003; Smith, Larson, & Nuckles, 2006; United States Public Health Service, 2003). The major focus of the proposed study is to determine if the implementation of an elementary school, universal, classroom-based, vi olence prevention curriculum as well as an experiential add on component result in observe d and reported improvements in students social skills and reductions in observe d and reported aggression. Review of Related Literature The following literature review begins with a di scussion of the current conceptualization of violence, youth violence and sc hool violence, followed by a disc ussion of the role of schools regarding violence. Outcomes of youth aggres sion and violence are presented next followed by a discussion of the role of prevention in comb ating youth violence. The review continues by presenting current theories regarding the occu rrence of violence as well as theories of intervention to reduce violence. The next componen t in this section review s current literature on violence prevention programs and their evalua tion. Violence prevention program planning and evaluation are then discussed, followed by the sign ificance and rationale of the current study. Current Conceptualization of Violence Prior to embarking upon an investigation of elementary school violence prevention programs and the effects of implementing such programs, conceptualization of violence is important. Government agencies such as the Centers for Disease Control (CDC) traditionally have provided guidance in defining and conceptualiz ing public health issues such as violence.
13 The CDC currently defines violen ce as threatened or physical fo rce initiated by any individual that results in or has a high like lihood of resulting in physical or psychological injury or death. Violence also has been defined as one or more acts that are interpersonal, situational, or predatory in nature or a combin ation of the three (Acosta et al., 2001; Tolan & Guerra, 1994). Youth violence The mean age of violent perpetrato rs has decreased during the past three decades (Fields & McNamara, 2003; Pettit & Do dge, 2003). The homicide rates between 1985 and 1991 reached record levels and further i llustrated the seriousne ss of youth violence. Additionally, American youth die and are involved in serious vi olence at much higher rates than among children and youth in all ot her industrialized countries (D ahlberg, 1998; Furlong, Paige, & Osher, 2003). The aforementioned statistical tre nds have led to further distinctions concerning violence and the age of the perpetrator. Th e downward extension of violent perpetration among younger children is termed youth violence and usually refers to ho micide or violent crime (Dahlberg, 1998). School violence. The term school violence refers to concern about youth violence and the outcomes of violence reflected in the education pr ocess. To scholars, the term refers to many different phenomena including delinquent behavior crime on school campuses, victimization at school, school disciplinary pr actices, weapons possession at sc hool, gang violence, and zero tolerance policies. The term school violence has expanded to refer to crim inal acts of violence as well as aggression in schools which affect development, learning, and school climate detrimentally (Acosta et al ., 2001; Furlong & Morrison, 2000). During the years 1995-1998, an average of five multiple victim school shootings occurred yearly, as compared to one multiple victim even t per year in the 3 years before 1995 (Fields & McNamara, 2003). Information from the United States Department of Justice (2005) indicate
14 recent fatal youth victimization rates show youth ages 5 were victims of 22 school-associated violent deaths from July 1, 2001, through June 30, 2002 (17 homicides and 5 suicides). Alternately, in 2003, students ages 12 were vi ctims of about 1.9 million nonfatal crimes at school, including about 1.2 million thefts and 740,000 violent crimes (simple assault and serious violent crime). 150,000 of which we re serious violent crimes (rape, sexual assault, robbery, and aggravated assault) (United States Department of Justice, 2005). Notably, recent statistics also indicate stude nts are more likely to be victims of serious violence or a homicide away from school sugge sting schools are generally more safe than a students living environment. For example, in 2003, students ages 12 reported being victims of serious violence at a rate of 12 crimes per 1,000 students away from school and 6 crimes per 1,000 students at school. Additionally, in each school year from July 1, 1992, through June 30, 2002, youth ages 5 were over 70 times more likely to be murdered away from school than at school (United States Department of Justice, 2005). Generally, current data trends indicate student victimizati on has decreased over the last decade. The nonfatal victimization rate for stude nts ages 128 at school by and large, declined between 1992 and 2003 (this was true for the total crime rate and for thefts, violent crimes, and serious violent crimes.) However, for y ears 2002 and 2003, no differences were detected between in the rates of total victimization, viol ent victimization, or thef t at school. For fatal victimization, between July 1, 1992, and June 30, 2002, the number of homicides of school-age youth at school declined as well. Fo r example, between the 1998 and 1999 school years, the number of homicides of school-age youth at school decl ined from 33 to 14 homicides. Since that time, there have been between 12 and 17 homicides in each school year through 2001 2002 (United States Department of Justice, 2005).
15 Schools also are settings for lo w-level aggression that can put children at risk for greater levels of aggression (Batsche, 1999). Low-level aggression also is refe rred to as school violence. Furlong et al. (2001) showed the prevalence of on e type of low-level a ggression: bullying. A nationally representative sample of 15,686 sixth through tenth grade students were surveyed, and 30% reported moderate or frequent involvement in bullying as eith er the bully (13 %), the victim (11%), or both (6%) (Furlong et al., 2003). Outcomes of Youth Aggression and Violence Empirical support for the detrimental effects of violence is overwhelming. Scholars have demonstrated children exposed to violence experience pervasive and profound negative developmental outcomes (Fields & McNamara, 2003; Lynch, 2006). Children who are victims or witness violence are more likely than nonvictims and non-witnesses to experience chronic fear, insecurity, avoidance, f eelings of alienation, and hopelessness. Further, they may demonstrate an inability to form secure attachme nts and trust, leading to subsequent difficulties in relationships throughout life (Clayton et al., 20 01). Psychosocial consequences of acute and chronic violence exposure include increases in depression, posttraumatic stress symptoms, aggressive behavior, memory impairment, wit hdrawal, and difficulties concentrating (Embry, 2002; Guerra, 2006; Salzinger, Fe ldman, &, Stock hammer, 2002). Victims of school violence have demonstrated a decreased ability and desire to learn (Clayton et al., 2001; Hintze & Sh apiro, 1999). Additionally, childre n who are victims of bullies more often have problems with self-esteem, anxiety, and depression (Taub, 2001; Vernberg & Gamm, 2003). Additionally, perpetrators of school violence who demonstrate persistent patterns of aggression towards peers as well as their victims demonstrate poorer psychological adjustment than individuals not involved in chronic vi ctimization (Embry, 2002; Werner & Crick, 2004). Children who are aggressive and im pulsive are more likely to be neglected by their peers (Taub,
16 2001). Moreover, childhood physical aggression is the single most predictive factor for delinquency, crime, and substance abuse during adolescence and adulthood (Acosta et al., 2001). Children who are bullies in school are more likely to commit antisoc ial, delinquent, and criminal acts in adolescence and adulthood. They also are more likely to underachieve, drop out of school, experience poor career performance, and commit acts of physical and sexual abuse (Clayton et al.; 2001). Role of Prevention Science in Co mbating Youth and School Violence Prevention science refers to an approach that combines developmental theory with Bronfenbrenners contextual theory (Bronfenbr enner, 1979), public health and developmental psychopathology (Greenberg et al., 2001). Th e following section will elaborate on the assumptions underlying youth viol ence prevention, the goa ls of violence prevention programs, and risk and protective factors. Assumptions of Youth Violence Prevention According to the United States Surgeon Genera l (United States Public Health Service, 2003), youth violence is a complex and ever-c hanging public health problem. However, researchers and educators have demonstrated prevention programs can prevent and reduce incidents of violence. Scholars have identified risk and protective factors that, if present, can either increase or decrease the likelihood children a nd adolescents will engage in violent behavior. A risk factor is anything that in creases the likelihood a child or adolescent will demonstrate violence or aggression. A protective factor is anything that decreases the likelihood the risk factor will cause aggres sion or violence (United States P ublic Health Service, 2003). Researchers have identified many risk factors for violence and potenti al protective factors which may buffer the effect of risk. However, these risk factors are no t predictive in and of themselves. That is, we are unable to predict wi th reasonable certainty if a particular child will
17 engage in violence based on the presence of particul ar risk factors. The predictive value of risk (or protective factors) change s, depending on when the factors occur in a young person's development, in what social c ontext, and under what circumstances Risk and protective factors may be intraindividual, environmental, or result from an interaction between the individual and the environment. Additionally, the effects of so me risk or protective factors may be dependent upon the individuals stage or level of development (U nited States Public Hea lth Service, 2003). Goal of Youth Violence Prevention Efforts The use of prevention programs to reduce the in tensity and frequency of violent behavior in youth is not a 21st century invention. Prevention progr ams have been implemented regularly since the 1970s (Dahlberg & Po tter, 2001). In fact, preventiv e interventions have been implemented to combat many social and educatio nal problems including obesity, illiteracy, drug use, dating violence, pregnancy, and school drop out. Prevention scientists seek to promote wellness and buffer risk in these areas as well as others (Morrison et al, 1994) by reducing the impact and occurrence of risk f actors and providing protective f actors to buffer risk (Farrell, Meyer, Sullivan, & King, 2003). Prevention theorists generally support a deve lopmental-ecological theory of violence. They believe violence results from adults failing to provide an appropriate environment in which children can grow and develop. As discussed prev iously, violence has been shown to be stable over time. However, psychosocial and environmen tal influences can have a powerful effect on disrupting the development of viol ence (Strein & Hoagwood, 2003). These theorists operate under several assumptions. They are as follows. (1) The family serves as the primary support and socializing force for children. (2) The challenges families face and ways they solve problems are dependent upon the age of the child. (3) Families and children are influenced by the larger social, cultural, and political environments. (4) Strengthening
18 families and family-focused interventions are among the more powerful ways to effect positive child outcomes and prevent violence or antisoci al behavior. (5) Individual development is influenced by the qualities of the individual s social systems ((Br onfenbrenner, 1979; 1988; Smith et al., 2004). This theoretic al approach often is used in conjunction with other cognitive oriented approaches in designing preventive in terventions. However, the costs of these developmentally and ecologically integrated progra ms are high in terms of resources. Because of the high cost of these programs many sc hools and agencies implement more focused interventions such as school, cla ssroom, or after school programs with or without parent and community support (United States P ublic Health Service, 2003). Introduction to Risk Factors Scholars (e.g., Boyd, Cooley, Lambert, & Ialonga, 2003; Dahlberg & Potter, 2001; Herrenkohl, Maguin, & Hill, 2000) have identified risk factors for youth violence through the use of longitudinal investigati ons that track individual charact eristics and social conditions. Scholars then determine if the individual charact eristics and social cond itions predict violent behavior. Research on risk factors has revealed most are environmental in nature and do not appear to have a strong biological basis. The presence of a singl e risk factor may promote an undesirable outcome. However, a combination of risk factors generally is thought to cause violent and aggressive behaviors. Researchers believe a combinat ion of risk factors has additive effects. That is, the more risk factors to whic h a child or adolescent is exposed, the greater the likelihood that individual will become violent. Fo r example, a 10-year-old exposed to 6 or more risk factors is 10 times as likely to be violent by age 18 as a 10-year-old exposed to only 1 factor. Research also has highlighted the importance of the timing of emergence of risk factors. Different risk factors may have more predictiv e power dependent upon when they emerge during
19 the course of development (Boyd, Cooley, La mbert, & Ialonga, 2003; Herrenkohl, Maguin, & Hill, 2000). Risk factors may be powerful tools for iden tifying and locating populations and individuals with a high potential for becoming violent. They also may provide valuable targets for program interventions designed to prevent or reduce violence. However, knowledge about and use of risk factors is limited. Researchers ha ve not identified one si ngle risk factor or set of risk factors sufficient enough to predict with cer tainty that children or adoles cents will become violent (Kam, Greenberg, & Walls; 2003). Risk Factors and Interventions An understanding of what puts ch ildren at risk for violence is paramount to understanding where to focus prevention efforts. Risk factor s can originate from the individual, family, school, peers, and community. Risk f actors for violence also are comm on to other personal and social problems including drug use, school drop out, delinquency, and teen pregnancy (Cunningham, 2000). Individual risk factors Social skills deficits. Early aggression beginning at a bout ages 8 to 12 is one of the strongest predictors of later violence (Dahlb erg & Potter, 2001; Molin a, Dulmus, & Sowers, 2005; Ikeda, Simon, & Swahn, 2001; Vitale, 2001). Aggressive behavior in childhood predicts later delinquency, substance abuse, depressi on, school drop out, and premature parenthood. Researchers also have reported st artling stability in aggressive and violent behavior over time and across generations. They have linked the st ability of aggressive behaviors and hostile attributional biases, normative beliefs supportiv e of aggression with social problem solving deficits (Dahlberg & Potter, 2001; Vitale, 2001). The link between social skills deficits and aggression is supported by the fact that children with social skills problems are more likely to
20 respond impulsively when faced with an interpersonal problem, thus placing them at greater risk for violence. For example, poking and pushing other children in elementary school as well as the frequent use of self-centered verbal responses to others (e .g., interrupting and blurting out thoughts) have been identified as early warning si gns of later aggressive and impulsive behavior (Taub, 2001). Abuse victims. Physically and emotionally abused children have been found to be more highly aggressive than other children as well as more prone to oppositional behavior, delinquency and criminality. Additionally, abused children may also show self-injurious and suicidal behaviors, substance abuse, and emotional problems. They often experience difficulties in relationships such as fighting with peers a nd being defiant toward teachers. Many of these problems are long lasting (Lis ak & Miller, 2003; Maxwell & Maxwell; 2003; Sneddon, 2003). Some researchers hypothesize childre n who are abused are more like ly to interpret behaviors as hostile regardless of intent. They may be quicker to react in a retaliatory manner, making them greater risk for aggression (Dahlberg, 1998; Dodge, 2001). Educational risk. Some of the most powerful, empiri cally supported child risk factors of violent behaviors are related to education, including low co mmitment to school and school failure (Dahlberg, 1998; Dahlberg & Potter; 2001). However, the causal relationship between academic success and violence remains tenuous. In fact the relationship is quite difficult to untangle. Researchers such as Scott, Nels on, & Liaupsin, (2001) point out that academic achievement is strongly related to social status and other social-environmental factors (Scott, Nelson, & Liaupsin, 2001) Use of alcohol and drugs. Additionally, the use of alcohol and/or drugs puts children and adolescents at risk for violence (Dahlberg, 1998). For example, a recent study of randomly
21 sampled 14to 18-year-old adolescents (3319 boys, 3890 girls) illustrated a powerful link between, smoking, drunkenness, peer drug use, and vi olence and violence re lated injury (Mattila, Parkkari, & Rimpel, 2006). Biology. Some researchers suspect antisocial beha viors, including violence against others, are connected to biological vulne rabilities that affect behavior and increase the likelihood of social deficits. These social deficits may in crease the likelihood of aggression exhibited by a child (Dodge & Rabiner, 2004). Additionally, earl y neglect also may put a child at biological risk as the limbic system can be affected detrim entally by lack of stimul ation (Dahlberg & Potter, 2001). Race and gender Some scholars also have hypothe sized a link between race/gender and violence. For example, African American male s may be at particular risk because many are raised in single, female-headed households, expo sed to elevated parental stress, have fewer community resources, are under a burden of pove rty and joblessness, often lack positive role models, and are more likely to have a father who has been incarcer ated (Okwumabua, Wong, Duryea, & Howell, 1999). Further, the vast majo rity of violent perpet rators are boys and young minority men. However, during the past two decade s girls have been arrested for violent crimes at higher rates than before (Chesney-Lind & Belknap, 2004). Relationships between race, gender, and violence are more complex than some other risk factors. Scholars know relatively little about the developmental c ourse of violence and associat ed risk factors like gender, socioeconomic status, race, and ethnicity (Aber, Brown, & Jones, 2003). Community/Environment Poverty puts children at great ri sk for later violent behavior. However, the relationship is complex (Dahlberg, 1998). The components of pove rty that affect child rens socio-emotional functioning detrimentally include experiencing discrete and ch ronic stressors, punitive and non-
22 supportive parenting, and feelings of inferiority (Dahlberg & Potter, 2001). Poverty also increases the likelihood a child wi ll behave aggressively because poverty often is coupled with peer group instability, frequent moves, lack of cognitive stimulation, harsh discipline, and drug and alcohol abuse (Knoff & Batsche, 1995; Dwyer 1999). The availability of firearms and exposure to violence in the community also puts a child at risk (Dahlb erg, 1998; Vitale, 2001). Peers Peer rejection. Children rejected by their peers are at greater risk for violent behaviors because they may seek out a negative peer group with whom to associate. Peer rejection and subsequent negative peer choice have been shown to strongly predict violence (Dahlberg, 1998; Dahlberg & Potter; 2001; Kerns & Prinz, 2002). Peer group choice. Peer groups generally are benefici al. However, children are more likely to engage in inappropriate behaviors wh en their peer group supports these behaviors. Aggressive behaviors elevate the adolescents social status in some peer groups (Dahlberg, 1998; Dahlberg & Potter, 2001). The risk for violence is substantially increased when children have negative peer influences as we ll as a weak family environmen t (Dahlberg & Potter, 2001). School The National Research Councils (Riess & Rot h, 1993) report identifies three empirically supported school characteristics th at may contribute to violence: relatively high numbers of students in a limited amount of space, a lesse ned capacity of schools to avoid student confrontations, and inadequate building design features (Morri son et al., 1994). Schools also may foster aggression by permitting undisciplined classrooms and lax enforcement of school rules (Dahlberg, 1998). The current investigation takes pl ace in a rural school district in Florida. Interestingly, the voters on Florida agreed to address large class sizes by amending the state constitution in
23 November, 2002 (two years before data collection). At the time of data collection, the average number of students (4th through 8th grades) in Columbia County was 23.71 students in 2003 and 20.05 students in 2004. The average for the st ate was 24.16 in 2003 and 22.43 in 2004 (Florida Department of Education, 2006). Family Family characteristics also can contribute to our u nderstanding of why particular children are at risk for violence and why patterns of viol ence may be stable over time. Family influences that increase risk are poor family manageme nt and parenting, defi cits in communication, spousal/partner violence, and a climate of c onflict (Dahlberg, 1998; Ma xwell & Maxwell, 2003; Greenhoot, McCloskey, & Glisky; 2005). Exposure to multiple forms of violence in the home has a particularly strong effect (Dahlberg, 1998; Vitale, 2001). Family violence also can interfere with a childs social development, ma king the engagement in prosocial behaviors with peers difficult (Vitale, 2001). Parental characteristics and practices The following intraindividual parental characteristics are positively associated with la ter aggression: parental attitude accepting of violent behaviors (Dahlb erg, 1998; Vitale, 2001), an antisocial personality disorder, history of criminal behavior, and drug or alcohol abuse. In addition, the following parenting techniques or practices are positively associated with later aggression: inadequate monitoring and supervision, inadequate communication, defici encies in problem solving, and harsh discipline (Dahlberg, 1998). In general, children who are rejected, ne glected, or maltreated by their parents are at greater risk for aggressive beha vior. Further, children who ar e neglected may be even more likely to be violent compared to children who are physically abused (Gorman, 2002; Tolan & Gorman-Smith, 2002).
24 Introduction to Protective Factors Protective factors have been conceptualized as absence of risk, and are qualities at the opposite end of the continuum from a risk factor For example, an appropriate parent-child relationship might be considered a protective factor because it is the opposite of an inadequate parent-child relationship, a known risk factor. However, a simple linear relationship of this sort (where the risk of violence decreases as parent -child relations improve) blurs the distinction between risk and protection. The view that protec tion is conceptually distinct from risk defines protective factors as characteristic s or conditions that interact w ith risk factors to reduce their influence on violent behavior. For example, a fam ilys low socioeconomic status is a risk factor for violence and a warm, supportive relationship w ith a parent may be a protective factor. Although the warm relationship does not improve th e childs economic status, it does buffer the child from some of the adverse effects of povert y (United States Public Health Service, 2003; Tolan & Gorman-Smith, 2002). Some theorists be lieve protective factors promote resilience. Youth who are considered resilient in the vi olence prevention field are those children and adolescents who are insulated from a life of vi olence due to protective factors and who do not engage in violent behaviors (Fields & McNamara, 2003). Interest in protective and risk factors emerged from research in developmental psychopathology. Researchers in this field bega n to ask why particular children exposed to multiple risk factors often escap ed their impact (Kuperminc & Br ookmeyer, 2006). This led to a search for the characteristics or conditions that might confer re siliencethat is, factors that moderate or buffer the effects of risk. Protec tive factors offer an explanation for why children and adolescents who face the same degree of risk may be affected differently (Pettit, 2004; United States Public Health Service, 2003). The identification and measurement of the effects of protective factors constitutes a new area of vi olence prevention researc h. Thus, information
25 about these factors is limited. Because they buffer the effect of risk fact ors, protective factors comprise an important tool in violence preventio n. Protective factors, like risk factors, may differ at various stages of development, they may interact, and they may exert cumulative effects. Just as risk factors do not necessa rily cause an individua l child or young person to become violent, protective factors do not guarant ee that an individual ch ild or young person will not become violent. Protective factors simply reduce the probability that groups of young people facing one or more risk factor will become invol ved in violence (United States Public Health Service, 2003) Empirically Supported Protective Factors Research delineating specific protective factors is less well developed than research in the area of risk. The following discussion includes cu rrent, suspected, protective factors in the areas of the individual, peers, school, and family. Individual protective factors Several within-child characteristics have been identified as protective factors. They are being female, having a high IQ, having a resilien t temperament, positive social skills and orientation, the ability to bond to systems, clear standards for behavior (Cunningham, 2000); having an intolerant attitude toward devian ce, and perceiving th ere are sanctions for inappropriate behavior. Other researchers have abbreviated and named these within-child characteristics social competence, autonomy, pr oblem solving skills, and sense of purpose and future (United States Public Health Service, 2003). Peer-related protective factors Children who have friends who engage in soci ally desirable behaviors possess a protective factor for violence (Dahlberg & Potter, 2001). Therefore, th e converse of delinquent peer association is an empirically suppo rted protective factor or buffer.
26 School-related protective factors When an entire school is impacted negatively by a particular subcu lture of violence, a school may influence students in a positive way by promoting and establishing conditions that may buffer the school culture from a subculture of violence. Such condi tions include increasing the academic capability of student s, supporting a culture that valu es academic achievement, and using firm but fair discipline while being sure th e focus is not on excessive security procedures (Furlong, Morrison, Chung, Bates, & Morrison, 1 999). Schools also may promote protective factors to buffer violence by promoting appropria te social skills, provi ding supportive personal connections with individuals, opportunities for skill application and service, opportunities to master subject matter, and early and intensive in terventions (Furlong et al., 2003; Morrison et al., 1994). Bonding to positive social systems provides buffe rs against the development of violence. Researchers have identified three conditions that facilitate bonding to a school, thus reducing risk. A child must have an o ccasion to belong to a group. The ch ild must be able to experience success in their chosen role. The child must be recognized for his or her contributions to the school (Furlong, Morrison, & Pavelski, 2000). Family protective factors Warm and supportive relationships with pare nts or other caregiver s effectively buffer children from risk. Further, children whose pare nts engage in regular pa rental monitoring also are protected from the risk of violence. Parents positive evaluation of their childs friends also serves as a protective factor (Dah lberg, 1998; Ellickson & McGuigan, 2000). In conclusion prevention scientists seek to re duce risk and amplify protective factors in the individuals they serve. However, this comm on overall goal of prevention does not translate easily into a common effort and method. Generall y, the developmental-ecological theory drives
27 the reason for the prevention program; howeve r, there are differing methods and underlying theories which provide the tool s for the program. To better u nderstand the nature of differing methods which address violence reduction, let us look at theori es regarding the occurrence of violence. Theory Related to Violent Behaviors A culture of violence is said to exist in the United States because of the following conditions: individual skills deficits, dome stic abuse, poverty, racism, unemployment, inadequate classrooms, easy access to weapons easy access to alcohol and other drugs, alienation from a cultural heritage lack of supervision and cons tructive outlets for young people, reduced influence of socializing institutions (e.g., churches and the family), and a media that models and glorifies aggressive solutions (Dwyer, 1999; Goldstein & Conoley, 2004;). Numerous theories discuss the mechanism conn ecting environmental influences and violent responding. However, no one theoretical stance app ears to be prominent in explaining why some intervention programs are more efficacious th an others (Fields & McNamara, 2003). The following sections include current theories comm only used to explain the occurrence of and promote the prevention of violence. Behavioral Theory The belief that the use of violence to solve in terpersonal problems is a learned behavior is widespread (DuRant, Barkin, & Krowchuk, 2001). Relatedly, traditional behavioral approaches have been used to reduce aggressive res ponding. However, these approaches have not demonstrated long-term success or cross-setting generalization. Scholars believe behavioral approaches do not promote long-term success becau se the focus of the intervention is on overt behavioral change and does not alter social-cognitive patterns that presumably guide the aggressive behavior. The impact of overt behavior management and contingent reinforcement is
28 likely to fade when aggressive children return to their natural environmen ts. Further, unaltered maladaptive cognitive patterns are believed to co ntribute to the maintenance of aggressive behavior and work against gene ralization (Nangle, Er dley, & Carpenter, 2002). Boot camps, an example of popular interventions following a behavioral model, have been found to be ineffective in reducing violent behavior and main taining long-term effects (United States Public Health Service, 2003). Cognitive Behavioral Skills Training Theory This type of intervention focuses on modi fying thought processes underlying aggression (Lochman, 1992; Spivak & Shure, 1974). Althou gh cognitive behavioral interventions promote improvements in cognitive skills and behavior, th ey have not consistent ly demonstrated that improvements in cognition result in behavior change. The recognition of anger related physiological cues often is a key skill taught in programs using this theoretical foundation. Students involved in this type of interventi on also are instructed on how to employ coping strategies to control anger. Th e coping strategy component of c ognitive behavioral programs is considered particularly importa nt because the emotional aspect of anger and its associated arousal are viewed as critical precipitants of aggressive resp onding. The Anger Coping program is a good example of this type of program (Nangle et al., 2002). Attribution Theory Some scholars believ e youth make inaccurate assumptions of violent intent in those around them (Daley & Onwuegbuzie, 2004; Fields & McNamara, 2003; Heider, 1959). Attribution theo rists attempt to explain aggre ssive responding in the following way. As people try to identify the causes or underpinnings of violen ce in their communities, some of their beliefs are accurate and others are distorted. For example, violent youth often believe others possess malevolent intent wh ile their assumptions generally are unfounded. Attribution theorists believe prev ention is possible if youth are trai ned to rethink these situations
29 and realize that adversity does not arise solely due to the bad inte ntions of others around them. The BrainPower program (Hudley et al., 1998) is a good example of one that uses attribution theory as its guiding princi ple (Fields & McNamara, 2003). Social Learning Theory Social learning theorists belie ve youth closely observe and model violent behaviors of others, including people on televisi on. These theorists hold that, in order to prevent violence, youth need to more critically eval uate the inadequate behavioral m odels to which they have been exposed and to learn new alternative models to help them behave more appropriately (Bandura, 1963). Children are believed to be more violent today due to a culture of violence portrayed in American media (Fields & McNamara, 2003). Res earch supporting this notion is mixed (Mercy, Krug, & Dahlberg; 2003). Second Step (may be accessed at http://www.cfchildren.org/cfc) is a popular intervention program designed upon the founda tion of social learni ng theory (Nangle et al., 2002). Social Control Theory The primary goal of interventions built on social control theory (Reiss, 1951) is to promote pros ocial relationships between youth and prosocial role models. Programs such as Big Brothers, Big Sisters are good examples of this type of intervention. Youth are required to make a commitment to soci ally appropriate goals and become involved in activities that directly compete with their ability to participate in deli nquent behaviors (Kerns & Prinz, 2002). Social Information Processing Theory Social information processing theory blends cognitive perspectives with an information processing paradigm. Theorists suggest cert ain deficits in soci al-cognitive information processing reduce aggressive child rens abilities to cope effi ciently with everyday social problems (Crick & Dodge, 1994; Dodge & Rabine r, 1994). These theorists suggest people
30 process social information through several step s by applying specific soci al-cognitive activities (Nangle et. al., 2002). A person is thought to typically deal eff ectively with information using minimal cognitive effort and is able to produce situation-relevant behavior. However, deficiencies in social-cognitive activities may re sult in processing errors or shortcomings that increase the likelihood of empl oying inappropriate social prob lem-solving strategies and behavior focusing on aggression. These research ers suggest processing so cial information is guided by latent social-cognitive memory structures. Individual differences in problem solving abilities result from differences in these social-c ognitive memory structures. According to this theory, the development of memory structur es depends upon environmental factors, intraindividual differences, and de velopmental level (Dodge, 2001; Dodge & Rabiner, 2004). Theorists ascribing to this vi ew suggest several steps are n eeded to process information (i.e., orientation towards the social problem, in terpretation of the situ ation, goal formulation, strategy generation, strategy evaluation, and behavi oral enactment). Many violence prevention programs have created problem solving steps comme nsurate with this model as aggression could result during any of these information proces sing steps (Dodge, 2001; Dodge & Rabiner, 2004). This theory is supported by the following resear ch findings. Aggressive children tend to be less attentive to relevant social cues, less accu rate in interpreting peer intention cues, more likely to attribute hostile intent to the actions of peers, more likely to endorse social goals that damage rather than enhance rela tionships, and have a social stra tegy repertoire comprised mostly of aggressive alternatives (Dodge, 2001). Aggressive children also believe they are good at being aggressive, that aggression leads to positive out comes, and that aggression is a legitimate response. Although these findings are correlational in nature some scholars believe these findings provide evidence for the existence of such cognitive processes (Dodge, 2001).
31 Review of Violence Prevention Interventions and Programs Determining the focus of a violence prevention intervention requires ca reful selection of a program which meets the needs of the targeted populat ion. The following section includes Table 1-3 which summarizes recent empirical invest igations on the effec tiveness of violence prevention interventions. This table is followed by a more in depth look at the program selected for the current intervention. The investigations described in the table ar e (1) studies of schoolbased violence prevention/social skills building programs, (2) studi es with experimental, quasiexperimental or pretestpostte st designs, (3) studies of progr ams with the specific aim of reducing or preventing violent or antisocial behavior, (4) and programs for children in elementary school. Table 1-3 contains the following components: study (i.e., authors of the original study conducted), program name (i.e., name of the pr ogram used in the study cited), level (i.e., universal, selected or selective, or indicated), ethnic/racial ma jority (i.e., reports the largest percentage of children from a particular ethnic group participating in th e study), total N (i.e., number of children participating in the study); males (i.e., number of boy children participating in the study); lessons (i.e., lists duration and fr equency of the intervention), approximate number of sessions (i.e., lists number of average sessions for comparison), primary SES; grade or range; the source (i.e. information in the chart in cludes the original study and may also include information from a review article or meta-analysi s); and rating (indicates program effectiveness). Meta-Analyses or Review Ratings I ndicating Program Effectiveness. Derzon, Wilson, & Cunningham (1999) conducted a meta-analysis in which program outcomes were converted into effect sizes. The authors coded each study, us ing1 through 4 stars, to indicate the degree to which treatment groups derived benefit. Hi gher ratings indicate the treatment groups demonstrated higher program benefits.
32 The Greenberg, Domitrovich, & Bumbarger (20 01) review included studies they deemed effective or promising by examining the origin al study and determining if the program was successful in reducing aggressive or antisocial behavior. The Ha milton Fish Institute completed a review of programs and rated them effective and noteworthy ba sed on the Derzon et al. (1999) meta-analysis. Youth Violence: A Report of the Surgeon Genera l (United States Public Health Service, 2003) delineates programs as effec tive or ineffective. It categori zed effective programs as either model (meet very high standards of demonstr ated effectiveness) or promising (meet a minimum standard). Researchers considered experimental rigor, re plication, degree of statistical significance reported, an d outcome measures when determining program status. They also distinguish effective programs as those th at have demonstrated effects on violence and serious offending (level 1) and t hose that have demonstrated eff ects on risk factors for violence (level 2). Program Chosen for the Present Study According to program developers (Greenberg et al., 1998), the Promoting Alternative THinking Strategies (PATHS) curriculum provide s a comprehensive program that promotes emotional and social competencies and reduces aggression and behavior problems in elementary schoolaged children. It also is reported to simultaneously enha nce the educational process in the classroom. The PATHS curriculum contains le ssons that seek to provide children with the knowledge and skills within thr ee major areas: readiness and self-control, feelings and relationships, and problem solvi ng. The lessons include instruc tion in identifying and labeling feelings, expressing feelings, assessing the intensity of feelings, managing feelings, understanding the difference betwee n feelings and behaviors, dela ying gratification, controlling impulses, reducing stress, self-talk, reading a nd interpreting social cu es, understanding the
33 perspectives of others, using steps for probl em-solving and decision-making, having a positive attitude toward life, self-awareness, nonverbal communication skills, an d verbal communication skills. The curriculum is designed for use by e ducators and counselors in a multiyear, universal prevention model that concentrates primarily on school and classr oom settings but also includes information and activities for use with parents. Ideally, the developers su ggest that the program should be initiated at the start of schooling and continued through sixth gr ade (Greenberg et al., 1998). PATHS has been field-tested and researched in general education classrooms, with a variety of special-needs student s (deaf, hearing impairment, learning disability, emotional disturbance, mild mental retardedation, and gifted), and among African-American, Hispanic/Latino, Asian-American, Pacific Isla nder, Native American, and Caucasian children (Greenberg & Kusch, 2006). The following paragraphs outline PATHS development and theoretical underpinnings. The first version of PATHS was developed in 1980 to fill the need for a comprehensive, developmentally based curriculum to promote soci al and emotional competence. It was also developed to prevent or redu ce behavior and emotional prob lems. According to program developers (Greenberg & Kusch, 2006), rese arch prior to 1980 s uggested that children experiencing behavior problems as well as thos e who were not, would benefit from universal, school-based curricula for the purposes of both promoting emotional comp etence and decreasing risk factors related to later aggression, violence and other forms of mala djustment (Greenberg & Kusch, 2006; Greenberg, Kusch, & Mihalic, 1998). The PATHS curriculum is based on five concep tual models (Greenberg & Kusch, 2006; Greenberg, Kusch, & Mihalic, 199 8). The first is the ABCD m odel. According to program
34 developers, this model is based upon psychodyna mic developmental theory, developmental social cognition, cognitive developmental theor y, cognitive social le arning theory, and attachment theory. This mode l stresses the importance of ch ildren developing integration of affective and emotional language and behavior as well as cognitive understanding to promote social and emotional competence. This model a ssumes that a childs copi ng, as reflected in his or her behavior and internal regu lation, is a function of emotiona l awareness, affective-cognitive control and sociological understa nding. PATHS activit ies reportedly help children learn about and identify affect, behavior, and cognition (Gre enberg & Kusch, 2006; Greenberg et al., 1998). The second conceptual model with which PA THS was developed is the Eco-Behavioral Systems Model. PATHS is said to follow an eco logically oriented program because aside from teaching the socioemotional skills, implementati on should also provide meaningful real-life opportunities to use the skills and provide reinfor cement for effective skill application. PATHS was designed to be implemented in the school environment as it serves as an ecology where social change can occur (Greenberg & Kusch, 2006; Greenberg et al., 1998). The third conceptual model deals with a child s neurological devel opment. Researchers (Luria, 1976; White, 1965) suggest that learning experi ences in the contex t of meaningful relationships during childhood influe nce the development of neural networks between different areas of the brain, which effect self-control and emotional awaren ess. PATHS strategies were designed to optimize the nature and quality of te acher-child and peer-peer interactions that are likely to impact brain development and learni ng. Specifically, researchers (Dawson, 1994) have indicated that optimal neurolog ical development of the frontal lobes allows these lobes to regulate emotional expression. In sufficient socialization in child ren is often associated with deficiencies in the connections between the frontal lobes and the am ygdala (in charge of
35 powerful emotions) and the motor cortex (whi ch controls movement). When adequate socialization occurs, adequate st ructuralization between these ke y areas occurs. Frontal lobe processes then promote the ability to automatica lly utilize inner speech which allows verbal thought to serve as a mediator fo r behavioral self control. Th ese frontal lobe processes also enable children to have great er independence, to plan ah ead, and to take on greater responsibility. PATHS proposes to support adequate structurali zation by engaging children in simple motor control in the early years and incr eased use of language a nd cognition activities in later years. Additionally, PA THS seeks to support neurological communication between each hemisphere of the brain. While the left hemisp here is responsible for processing receptive and expressive language as well as the expression of positive affect, the right hemisphere is specialized for processing both comfortabl e and uncomfortable receptive affect and uncomfortable expressive affect. Linguistic in formation is often processed without conscious awareness, unless we verbally think about it. PATHS helps children verbally label emotional experiences, and thus become consciously awar e of them. This proc ess is hypothesized to promote increased communication between left and right hemispheres and should assist in managing feelings, controlling be havior, and improving hemispheri c integration (Greenberg & Kusch, 2006). The fourth model involves the applica tion of psychodynamic theory to enhance developmental growth, promote mental health, an d prevent emotional distress. The developers believe that when teachers express an interest in childrens feelings and emotional experiences as well as show respect for their opinions, children are impacted in a profound way. This model focuses on the teacher-student relationship and how a positive one may enrich and enhance learning. This model also emphasizes interna lization and the subsequent development of a
36 childs conscience. PATHS is meant to promot e the internalization of prosocial values by helping children understand why these values are important, rather than memorizing rules (Greenberg & Kusch, 2006; Greenberg et al., 1998). The fifth model addresses a problem many child ren encounter when facing a problem such as being teased. Children appear to be much less likely to engage in socially competent behavior if they are unable to both accurately proce ss the emotional content of the situation and effectively regulate his or her emotional arous al so that he and she can think through the problem. This fifth model is referring to emotional competence (also termed emotional intelligence). PATHS activities are meant to help children understand their own and others emotions which is a key component in problem solving. Additionally, the five models which underlie the PATHS curriculum are meant to build protective factors that reduce maladjustment (Greenberg & Kusch, 2006). Greenberg and associates (CPPRG, 1999; Domitrovich & Greenberg, 2000; Kam, Greenberg, & Kusch, 2004; Kam, Greenber g, & Walls, 2003; Kusch, 2003; Greenberg & Kusch, 1997, 1998; Greenberg, Kusch, Cook, & Quamma, 1995; Riggs, Greenberg, Kusch, & Pentz, 2006) have conducted research in the area of PATHS effectiv eness, implementation efficacy, and general utility. Additionally, resear chers not associated with PATHS development also have investigated the us e of PATHS with a variety of populations (Johannes, 2004; Kelly, Longbottom, Potts, & Williamson, 2004). The following paragraphs summarize noteworthy research which has examined the PATHS curriculum. In a 1995 investigation, Greenberg, Kusch Cook, and Quamma (1995) examined the effectiveness of the PATHS curriculum on the emotional development of school-aged children. The study included 30 classrooms in a randomized design and involved the assessment of 286
37 children from Grades 2 and 3. The participants were in self-contained special needs classrooms (33%) and in regular education (67%). Teach ers trained in the intervention model provided PATHS lessons during most of 1 school year. Th e investigators indicated that the intervention was effective for both lowand high-risk child ren in improving the range of vocabulary and fluency in discussing emotional experiences, efficacy beliefs regarding the management of emotions, and developmental understanding of some aspects of emotions. In some cases, greater improvement was shown in children with highe r teacher ratings of psychopathology (Greenberg et al., 1995). In a randomized controlled trial with 200 second and third grade regular education students, researchers (Greenberg & Kusch 1998) found that PATHS produced significant improvements in social problem solving and unders tanding of emotions at post-test. Compared to control, general education intervention child ren demonstrated 1-year follow-up improvements on social problem-solving, emotional understandi ng, self-report of conduct related to social planning and impulsivity. They also reported th ese improvements were maintained at 1-year follow-up. Additionally, significant reductions in teacher and student reports of conduct problems appeared at 2-year follow-up. For chil dren with special needs, results indicated posttest improvement on teacher-rated social comp etence, child report of depressive symptoms, and emotional understanding and social-cognitive skill s. They also found at 1-year and 2-year follow-up, both teachers and children separately reported significant improvements in both internalizing (e.g., depression and somatic complaints) and externa lizing behavior problems, as well as improved social planni ng and decreased cognitive impul sivity (Greenberg & Kusch, 1998).
38 In a 1998 investigation, Greenberg & Kusch (1998) examined the e ffectiveness of the PATHS curriculum on the social, cognitive, and behavioral status of elementary school-age deaf children. The intervention field trial included a quasi-experimental, wait-list control design involving 57 first through sixth gr aders in 11 self-contained classr ooms. Teachers were trained in the intervention model and pr ovided PATHS lessons during most of one school year. The investigators found that the intervention led to significant improvement in students' social problem-solving skills, emotional recognition sk ills, and teacherand parent-rated social competence. They did not find an effect on t eacher or parent rated psychopathology. One and two year posttest results indicated maintenance of effects. In a 2003 study, Kam, Greenberg, & Walls (2003) investigated the quality of implementation and effectiveness of the PATH S curriculum in a high risk urban area. Participants were 350 1st grader s (47% males) in 6 inner-city public schools. Three schools implemented the intervention and the other 3 were comparison schools from the same school district. The investigators stat ed that intervention effects were not found for all the intervention schools. Reportedly, the intervention was effective in improving children's emotional competence and reducing their aggression in schools which eff ectively supported the intervention. This study also found an effect regarding curriculum implementation. They found that there were two factors whic h contributed to the success of the intervention: adequate support from school principals and a high degree of classroo m implementation by teachers. They propose that the findings highlight the impo rtance identifying multiple factors influencing the implementation of school-based interv entions (Kam, Greenbe rg, & Walls, 2003). The Conduct Problem Prevention Research Group which is associated with Greenberg and colleagues (CPPRG, 1999) has also examined the effects of PATHS in the context of a larger
39 conduct problem preventive interven tion called the Fast Track Progr am. In this study, 198 first grade classrooms from high-crime neighborhoods were randomly assigned to use the PATHS curriculum, while another 180 first grade cl assrooms from the same neighborhoods were randomly selected to serve as a control group. T eachers in the control group pursued their usual lesson plans, teachers in the treatment classr ooms delivered a 57-lesson version of the PATHS curriculum emphasizing self-contro l, emotional awareness of peer relations, and solving problems. The evaluators then assessed the imp act of the PATHS curriculum on students social competence using three distinct outcome meas ures: teacher reports (compiled through two structured interviews with each participating teach er), individual sociometric interviews with all children providing parental consent, and obs erver ratings (compiled by having impartial observers rate the classroom atmosphere as a whol e on a scale of 1 to 5). Multivariate statistical analyses were then used to control for variou s differences among individual students and to compare the social competencies of students in the treatment and control classrooms. The researchers reported that their results indi cated students in the intervention classrooms were less aggressive according to peers. Additionally they re ported the intervention classrooms were rated by observers as more having a more positive atmosphere than control classrooms (CPPRG, 1999). Program dosage was measured by asking teachers to report the number of lessons that they taught each week. Further, pr ogram fidelity was assessed by the project staff with direct observations of the teachers in the context of their ongoing consultation. Four dimensions were coded: quality of teaching PATHS concepts, modeling of PATHS concepts throughout the day, quality of cl assroom management during PATHS lessons, and openness to consultation from staff members. They reported no effects for dosage. They also found ratings of teacher skill in program implementation and classroom management predicted classroom
40 differences in positive program outcomes. Specifically, the teachers rated skill in teaching PATHS concepts, managing the classroom, and modeling and generaliz ing PATHS concepts throughout the classroom day were all significantly related to teacher ratings of AuthorityAcceptance (p < .001, p < .001, p< .001, respectively). These three measures were also related to observer ratings of classroom atmosphere (p < .01, p < .01, p < .01, respectively) (CPPR, 1999). In a more recent study, Kam, Greenberg, & Kusch (2004) examined the long-term effectiveness of the PATHS curriculum on the ad justment of school-age children with special needs. Eighteen special edu cation classrooms were randomly a ssigned to treatment and control conditions in this controlled trial. Teachers received both training and ongoing consultation and provided PATHS to students in Grades 1 through 3. The researchers reported data were collected before the intervention and for 3 succes sive years. They reported the intervention reduced the rate of growth of teacher-reported internalizing and externalizing behaviors 2 years after the intervention and produced a sustained reduction in depres sive symptoms reported by the children (Kam, Greenberg, & Kusch, 2004). The evaluators did not find an observable treatm ent effect in overall social competence and problem solving. They suggest this may be due to several mitigating factors: the PATHS curriculum being used was a single-year model, which may be less effective than longer PATHS programs; the curriculum being used was also an early version of th e PATHS program, which lacked several key lessons on dealing with difficu lt peer-related situations; and the instruments used to rate teachers perceptions of social co mpetence may not have been sufficiently sensitive (Kam, Greenberg, Kusch, 2004).
41 PATHS has also been the subjec t of dissertation research. In 2004 a disserta tion abstract Johannes (2004) reported on the us e of PATHS in an after sc hool program. At the time, researchers had not examined the effectiveness of PATHS delivered in the after school setting. The purposes of this study were to determine th e influence of PATHS af ter school on participant social behaviors and the social environment, a nd to determine the relation among dimensions of social behavior and environment. A quasi-exp erimental design was employed in three after school sites who agreed to implem ent the intervention. They were compared to two after school sites from similar small, rural Midwest comm unities who did not implement the intervention. Sites were compared at pre and th en post intervention (20 weeks). The social behaviors of children (N = 74, age range = 9-11, mean age 10 years) participating in each site were measured by be havioral observation. Staff assessment of four dimensions of social behavior was measur ed by the Teacher Observation of Classroom Adaptation-Revised and Social Health Profile (TOCA-R & SHP). Additionally, an independent research assistant completed the Classroom Ratin g Scale (CRS) yielding three dimensions of the social environment. The study found that social behavior did not significantly improve over time when comparing the intervention and nonintervention sites. Further, no significant differences were found in social environment between sites (i ntervention CRS posttest mean = 2.90, comparison CRS = 2.95; p = .05). However, when interventi on sites were categori zed by the quality of program, quality implementation si tes significantly improved in di mensions of social behavior (social competence t = -4.669, p < .05, cognitive c oncentration t = -3.516, p < .05, social contact t = -3.044, p < .05) (Johannes, 2004).
42 In 2006, Riggs, Greenberg, Kusch, & Pentz (2006) investigated the link between neurocognition and PATHS. They feel that c onsideration of neuropsyc hological factors holds promise in the construction of comprehensive, developmental models for the promotion of social competence and prevention of problem behavior They propose neuropsychological models of behavior suggest that childrens neurologica l functioning affects th e regulation of strong emotions, as well as performance in social, cognitive, and behavioral spheres. Their investigation examined the underlying neurocogn itive conceptual theory of PATHS. They hypothesized that inhibitory cont rol and verbal fluency would me diate the relationship between program condition and teacher-repo rted externalizing and intern alizing behavior problems. Participants were 318 regular ed ucation students enrolled in the second or third grade. They reported a series of regression analyses pr ovided empirical support for the effectiveness of the PATHS curriculum in promoting inhibitory control and verbal fluency and a partial mediating role for inhibitory control in the re lation between prevention condition and behavioral outcomes. They report, based on the results of this study, that program s designed to promote social and emotional development should consid er comprehensive models that attend to neurocognitive functioning and development. Additio nally, they state that lack of consideration of neurocognitive pathways to the promotion of social competence may ignore important mechanisms which impact intervention effectiven ess. They suggest developers of socialemotional prevention programs should design curricula to explicitly promote the developmental integration of executive functioning, verbal proc essing, and emotional awareness. They propose doing so may enhance prevention outcomes partic ularly if those preventions are implemented during a time of peak neurocognitive developm ent (Riggs, Greenberg, Cook, & Prinz, 2006).
43 Kusch, one of the developers of PATHS, pr oposed the program could be used to develop ego development, object relationships, and cortic al integration in chil dren. Kusch reported (2002) that employing psychoanalysis in the se rvice of violence prev ention could be made available to a large segment of the population by using the univ ersal prevention of PATHS. According to Kusch, PATHS teaches emotional literacy. She suggests emotional literacy is critical for an individuals basic knowledge repe rtoire, and that children who are not taught this information will likely be at a distinct so cial disadvantage as adults (Kusch, 2002). In 2004, Kelly et al. (2004) described their collaborative action re search project in a primary school. It was an exploratory qualitativ e study of the PATHS curri culum. Their project was aimed at promoting emotional competence in children. They chose PATHS because of its clear conceptualization of em otion, its emphasis upon cognitive and developmental aspects and its research history. One class of 9 and 10 year olds took part in the project. Target children were selected from within this group for clos er monitoring. The outcomes suggest that PATHS was rated very positively by class teachers, pupils and other staff involved in the project. The researchers attributed positive emotional, soci al and behavioral changes at the class and individual level to the effects of PATHS. They believed PATHS developed a more positive classroom environment (Kelly et al., 2004). Violence Prevention Program Planning and Evaluation Knowledge of the theory suppor ting the intervention as well as the research supporting the intervention allows those implementing and eval uating the intervention to share a common goal and conceptualization of the problem they are atte mpting to prevent. These pieces of research are paramount to implementing the best prog ram for the children and adults who are participating. However, there are also methodol ogical best practices which are supported in the
44 literature. The following paragraphs include methodological supports to promote successful program implementation. Pre-Planning Methodology to Promote In tervention and Evaluation Fidelity Several best practices in inte rvention program planning have em erged from research in this area (Greenberg et al., 2001). Th ey are as follows. (1) The pr eventive intervention should span the course of at least half a school year. Re searchers suggest that short term preventive interventions produce limited effects for at risk ch ildren. (2) Preventive interventions are best directed toward risk and protectiv e factors rather than at proble m behaviors. (3) Interventions should be aimed at multiple domains (e.g., school, home, and community) and result in changing institutions, environments, and individuals. (4 ) Programs should educate the child and instill positive changes across both the school and home environments. (5) If the intervention is designed for school-aged children, the school setti ng should be a central focus of intervention. (6) School personnel in conjunction with community providers should provide settings with fully integrated models (community care systems) (Greenberg et al., 2001). The current intervention (1) spanned more than ha lf of the school year. (2) It was directed towards the protective factor of impr oved social skills. (3, 4, & 5) It was aimed at children in the school setting with parent support and participat ion. Additionally, (6) ch ildren participating in the intervention were attending school in a district participating in a district wide risk reduction program. Planning Classroom-Based Preventive Interventions: Evaluative Focus A successful school-based violence preventi on intervention requires careful planning. Interventionists must determine the focus of the in tervention (i.e., type of violence to target and target population) and identify the strategy and plan for their integration in to school or classroom settings (Flannery, 2003; Kerns & Prinz, 2002).
45 Determine the focus of the intervention Interventionists must determine the population they intend to target for intervention. This decision should be guided by investigation into the types of violence occurri ng at school as well as the risk and protective factors present in the population (Thorton, Craft, Dahlberg, Lynch, & Baer, 2002). The focus of the intervention in th e current study was determined by reviewing the results of a needs assessment, a pilot study condu cted a year prior to the current study, and by reviewing violence prevention programs and their evaluation. The school di strict in which the pilot study was conducted was alrea dy providing a wide array of pr evention services. Results of the needs assessment indicated that a universal preventive intervention targeted at daily, low intensity aggression was appropriate. Strategy and plan for implementation and intervention fidelity Intervention fidelity is key for evaluation e fforts and requires an integration strategy. Important qualities that can promote high interv ention fidelity in a school setting can be organized into three areas: congruence between district/sch ool goals and the intervention, support and involvement of school staff, and quality of intervention materi als (Kerns & Prinz, 2002). District and school goals. The intervention program s hould link school interests with program goals. This link promotes staff and faculty program acceptance and may contribute to the promotion of an enthusiastic team approach The design should be incorporated into the school vision or mission statement as well as th e academic curriculum. To increase program acceptance, core faculty should be recruited to advocate for the program and to aid in determining which staff will implement the inte rvention with integrity (Fagan & Mihalic, 2003; Kerns & Prinz, 2002). In the current investig ation, the intervention wa s supported by the school district as part of an overall prevention effort Additionally, school administrators and teachers
46 chose to participate in the intervention based on their desire to increase social skills in the children attending their school. Staff involvement and support. Strong relationships with school staff also promotes intervention fidelity. If possible, conceptualize the school staff as collaborative consultants. This approach requires the in tervention team to be committed to common issues (e.g., youth violence) they are targeting. St aff also should be enthusiastic have a shared vision, and set common goals. They then should develop a strate gic plan for reaching the goals, take action to implement the plan, and evaluate progress towa rd attaining the goal (V ernberg & Gamm, 2003). In the current study, school staff were recruite d based on their desire to participate. The voluntary nature of the project was a strength in terms of co mmitment to helping children overcome social skills deficits a nd improve classroom functioning. Staff training Careful selection and training of pe rsonnel also is necessary for effective implementation and evaluation. Staff should receiv e supervision regularly and use standardized manuals and protocols as intended (Dodge, 2001). In the current study school staff were trained by the primary investigator. See Methods for further information. Curriculum choice for a classroom-based prevention program. Research indicates curriculum content should address daily aggressio n, victimization, as well as more high intensity events (e.g., physical altercations and verbal th reats of serious physical harm) occurring in elementary schools. This emphasis is important because some children demonstrate early in their development relatively mild forms of aggression that later transform into more violent behaviors in middle and high sc hool (Leff, Power, Manz, Cost igan, & Nabors, 2001). In the current investigation, fourth grade students were chosen as participan ts in an attempt to inoculate them against violent acting out in later grades.
47 Curriculum choice should implicitly and explic itly explain staffs beliefs concerning why violence occurs (Hunter et al., 2001). Curriculum should be evalua ted and chosen in terms of the degree to which intervention fidelity may be main tained, given lesson plans and other resources. Curriculum content should reflect sensitivity to ethnically and culturally diverse students and families (Alkon et al., 2001; Farrell, Meyer, Sullivan, & Kung, 2003). In the current investigation, school guidance c ounselors were consulted rega rding these issues. See the following section for more detail. Rationale for the Study The rationale for implementing and evaluati ng a universal, classroom-based violence prevention intervention for fourth graders and selecting the Promoti ng Alternative Thinking Strategies (PATHS) (Kusch & Greenberg, 1994) problem solving module for implementation in the current study is based on four factors: the needs of participating students, teachers, and schools; United States Department of Educa tion criteria; theoretical underpinnings of the program; and opportunities for collaboration with program developers. Additional purposes of the study include investigation into the use of an experiential add on component (i.e., video production) and the utility of group observationa l method. The rationale for these unique components also will be discussed below. Type of Intervention and Needs of Students, Teachers, and Schools This intervention was implemented as part of a district wide preven tion partnership called Columbia Acting Together for Children or Pr oject CATCh. Project CATCh was a community wide partnership focusing on violence and substance abuse prevention among school aged children. The project also provi ded and coordinated individualiz ed intervention services for identified children and families. Project CATC h was federally funded through the Safe Schools/ Healthy Students Initiative. The goals of Project CATCh were to reduce overall juvenile crime
48 rate, reduce juvenile alcohol and drug use, improve student/family/school communication and collaboration, increase school safety by reduci ng school-based aggressi on and violence, and increase academic achievement and student succe ss. The University of Floridas National Rural Behavioral Health Center (NRBHC) was a princi ple partner and provided multiple services to Project CATCh including the implementation of th e intervention discussed in this paper. The NRBHC also provided guidance in the developm ent, implementation, and evaluation of other school-based prevention programs such as Too Good for Drugs. Information regarding Too Good for Drugs may be accessed at the following website: http://www.mendezfoundation.org/educationcente r/tgfd/index.htm. The NRBHC also was responsible for assessing stude nts referred to the Prevention Management Team and providing therapy services to selected st udents and their families (tertiary prevention component). Further information may be accessed at http://www.projectcatch.org/. The participating Columbia County School Distri ct staff and Project CATCh research staff determined a process curriculum would best meet the needs of the schools due to teacher time constraints and existing high demands on fourth grade curriculum content. This decision was made based on the results of an informal survey with participating teacher s and results from The Risk Incidents for Schools Inventory (RIScI) (Na tional Rural Behavioral Health Center, 2003). A process curriculum requires inst ructors to devote a specific time to teaching the foundation abilities, principles, and one or more of the problem solving processes as a separate course with a distinct curriculum and daily lesson plans. A review of the districts needs assessment, informal interviews w ith the school counselor and participating teachers, as well as school discipline records indicated students in the participating fourth grade classrooms may benefi t from a universal intervention designed to
49 promote age appropriate problem solving skills and decreases in aggression. PATHS (Kusch & Greenberg, 1994) was chosen as it is a classr oom-based, universal pr evention designed for classroom use with elementary school childre n. The problem solving unit of PATHS provides instructors with systematic, developmentally ba sed lessons and material s designed to promote comprehension and interpretation of social cu es, perspective taking, and problem-solving. Department of Education Criteria The PATHS problem solving unit was chosen fr om a list of interventions identified by the Florida Department of Educati on to meet requirements for th e Safe and Drug Free Schools and Communities Act. The program list was compiled by the United States Department of Education following investigation of reviews from the follo wing agencies: the Collaborative for Academic, Social, and Emotional Learning (CASEL), Center for the Study and Prevention of Violence (CSPV)Blueprints for Violence Prevention, Cent er for Substance Abuse Prevention (CSAP), National Resource Center for Safe Schools, US De partment of EducationOffice of Educational Research and Improvement, the National Diffu sion Network, and the US Department of Education's Expert Panel on Safe and Drug Free Schools. PATHS has been designated as a Model Program by CASEL and the Substan ce Abuse and Mental Health Services Administration, and as a Promising Program by US Department of Education's Expert Panel on Safe and Drug Free Schools and by the Surge on Generals report on youth violence (United States Public Health Service, 2003). Theory The problem-solving intervention is expected to promote social skil ls acquisition in the children who participate. Prev ention models evolved from developmental psychopathology and developmental ecological perspectiv es provide the theoretical rationa le for the appropriateness of the intervention. In the presen t study, the creation of new soci al skills, namely appropriate
50 problem solving, is the mechanism that is theo rized to reduce aggressive responding. Social skills acquisition is meant to provide a buffer agains t risk factors present in the child and in the childs environment. Buffers or protective factors are within ch ild variables or environmental variables that are likely to inhi bit the development of maladaptiv e behaviors or psychopathology. Researchers have suggested that protective fact ors are effective in reducing violence because they may directly decrease dysfuncti on, interact with risk factors to buffer the effects of the risk interaction, disrupt the chain that leads from risk to dysfunction, or prevent the appearance of risk factors all together. Researchers have de monstrated suitable social-cognitive and social skills together with appropriate peer relations hips are powerful protective factors serving as effective buffers of aggression and violence (D ahlberg & Potter, 2001; Greenberg, Domitrovick, & Bumbarger, 2001). Opportunities for Collaboration The study was implemented with the support of The National Rural Behavioral Health Center (NRBHC) at The University of Florida as well as the Columbia County School District (Florida). The NRBHC worked with Mark Gr eenberg, Ph.D., program developer and colleagues at the Prevention Research Center at Pennsyl vania State University. Dr. Greenberg and colleagues provided the curriculum for use in the study and offered the opportunity for consultation with NRBHC, CATCh and Co lumbia County School staff as needed. Use of an Experiential Add-On Experiential learning refers to learning that is rooted in do ing. It is a learning which illuminates an experience and provides direction for decision making. Experiential learning also refers to the structuring and seque ncing of learning that leads to the increased effectiveness of the experience (Moon, 2004). The sc ript development, practice, and filming in the curriculum plus video group are conceptuali zed as experiential learning. Ch ildren in the curriculum only
51 group experienced this learning proc ess as well by engaging in role plays. However, not to such a large degree. Experiential learning is thought to be an eff ective for learning new skills due to several factors. When students are given the freedom to design their own expe riences, they may take more ownership in the process increasing intrinsic motivation for the task. Motivation for the task may also arouse excitement and self-directed learning. When interest, excitement, and selfdirected learning are present, students require gui dance as well. Adults can foster the learning process by asking students directive questions, providing appropriate problems that will maintain interest, and resources to answer their questions. The video pr ocess used in the current study was designed to increase intrinsic motivation, ar ouse excitement and produce self-directed learning (Wurdinger, 2005). The process will be further discussed in the Methods section. Use of Group Observational Method Direct observation methods are commonly used by professionals in the school setting to assess child characteristics. The utility of dire ct behavioral observati ons have more recently become a topic of interest for researchers due to the emergence of the problem solving approach in school psychology circles. Researchers ha ve suggested that student assessment and intervention evaluation techniques need to reflect an experiment al process in order to more accurately identify target behaviors, functions of behaviors, reinfor cers, and intervention success. Some student behaviors ar e legitimate problems and targets for interventions. This assumption allows for the direct assessmen t of the behaviors and for the direct evaluation of treatment effects. The direct method of observation fo r this study was develope d in order to assess students social behaviors. Researchers have called for th e development of behavioral methodologies to better understand and apply direct observation procedures. The development
52 of a group observational method which was bu ild upon functional behavioral assessment techniques is in keeping with current rese arch needs (Skinner, Dittmer, & Howell, 2000). Behavioral theory does not assume that a students behavior will be stable over repeated measurements; however, interpreting data from dire ct observations does requi re inferences to be drawn regarding the presence of the behaviors ou tside the data collection period. Use of the behavioral observation method in this study also requires this inference. Rather than narrative method, this study employed an empirical method of data recording to allow for simple comparisons across the intervention groups. The direct behavioral observation method will be further discussed in the Methods section (Skinner, Dittmer, & Howell, 2000). Significance of the Study Scholars have discussed several areas of de velopment for future research on violence prevention. The proposed study seeks to inform research by providing data concerning the effectiveness of a universal, classroom-based curriculum (i.e., the problem solving unit of Promoting Alternative Thinking Strategies) as well as the effectivene ss of an experiential component linked to the curriculum (i.e., video pr oduction). It also provides an example of a unique program evaluation tool (i.e ., group behavioral observations). Exploring the Effectiveness of a Cla ssroom-Based, Universal Prevention The potential value of violence prevention is undisputed, given violen ce statistics in our country. Additionally, childhood physical aggre ssion is the single most robust factor in predicting delinquency, crime, and substance ab use during adolescence and adulthood (Werner & Crick, 2004). Although many professionals believe implementing violence prevention services may be a good idea, data that inform d ecision-makers about which specific strategies or combinations of strategies work best with which populations are li mited (Dahlberg, 1998; Farrell, Meyer, Sullivan, & Kung, 2003; Tolan & Guerra, 1994). Because violence prevention
53 programs may be universal (i.e., implemented w ith an entire populati on), selected (i.e., implemented with students considered at risk fo r violence), indicated (i.e., implemented with children exhibiting violence), or a combination of the three, e ducators often have difficulty making informed choices. For example, Derzon, Wilson, & Cunningham (1999), reported school-based violence prevention programs were su ccessful in reducing anti-social behavior such as aggression, regardless of type of program. However, there are few published evaluations of universal, classroom-based violence prevention programs. The literature contains almost 2 times more review articles than independent studies (Derzon, W ilson, & Cunningham, 2001). A 2001 review of youth violence prevention interventions identified 1,168 empi rical articles focused on the description and treatment areas of youth violence. The review indicates most interventions focus on describing and treating areas of youth violence, assessing different t ypes of violence, asse ssing the effects of exposure to violence, and unders tanding the sequalae of violen ce rather than validating the effectiveness of treatment programs or analyzin g prevention efforts. A 2007 Psych Info search using keywords violence prevention a nd evaluation (narrowed by school based interventions) yielded six results dated 2004 to 2006 (two of which were original research articles). Given the paucity of research on progr am effectiveness, educators experience difficulty making reliable inferences about which programs are effective. Exploring the Effectiveness of an Experiential Add On Engagement of young people in learning skills and processes intended to enable them to manage and resolve conflict constructively constit utes one of the fundamental challenges of violence prevention programs (Gorman, 2002). Researchers have suggested interventions may be more successful in increasing childrens pr osocial behaviors and decreasing aggressive behaviors when the interventions require children to engage in more experien tial components
54 such as social skills training thr ough role plays, social skills pract ice with a peer role model, and psychodrama (Bosworth, Espelange & DuBay, 1998; Cirillo, Pruitt & Colwell, 1998; Clayton, Ballif-Spanvill, & Hunsucker, 2001; Tobin, & Sprague, 2000). The proposed study seeks to engage students in becoming independent pr oblem solvers throughout the learning process by using extensive role-plays, script development, and the promise of a permanent product in which they star (Taub, 2001). Contributions to Practice This study proposes to inform educators concerning the value of a classroom-based violence prevention intervention. Multiple shootings at US schools have focused the attention of lawmakers, politicians, parent s, educators and students on school violence. Funds have been allocated and mandates have been established th at require schools to prepare for and prevent violence. In turn, researchers and the private sector have pub lished programs designed to solve problems associated with school violence. School ad ministrators have been left with the task of selecting and implementing a program from th e many available. Administrators often concurrently oversee other school programs designe d to reduce tobacco, alcohol, drug use, and teen pregnancy. Thus, it is not surprising that some schools choose the shortest, least complicated, least expensive, and least dema nding violence prevention program possible to fulfill their districts requirement (Daiute, St ern, & Lelutiu, 2003; Guerra, 2003; Merrell, 2002). Individuals investigating viol ence prevention research may experience confusion when faced with beliefs (e.g., Dahlberg, 1998) which posit youth violence problems occur as a result of multiple risk factors and require comprehensive prevention programs. Dahlberg believes the most successful violence prevention programs are family-based, occur early, and address multiple components including individual risks, family dynamics, family coping, and parenting skills. These broadly focused programs often are unrea listic or impossible to implement in many
55 schools due to multiple barriers. This study s uggests a classroom-based, violence prevention program along with an experiential component may contribute to increased problem solving skills and decreased aggression at school (Chapin & Coleman, 2003; Dahlberg, 1998). Evaluating Classroom-Based Violence Prevention Interventions Undertaking an evaluation of a violence prevention intervention requires careful consideration of possible confoundi ng conditions that may impact intervention effectiveness. These conditions may occur in any of the interven tion fidelity issues stated above and also may be due to empirically supported moderating conditions. Moderating conditions have the potential to dilute intervention e fficacy. That is, some conditions in the childs environment may have a greater impact than the intervention. Several moderating factors have been studied regarding youth violence. For example, undesirabl e parental characterist ics and skills, negative school, elevated level of existing behavior prob lems, and perception of few program benefits may have an important moderating influence (E mbry, 1997; Kerns & Prinz, 2002). The pre post intervention design in the current evaluation may account for preexisting be havioral problems in the students. However, adding further evaluati on to the current evalua tion was not possible due to lack of resources. Future evaluations may co nsider the use of a school climate measure and or teacher attitude survey. Effective evaluation of a violence prevention intervention also requires researchers to use appropriate outcome measures and sound met hodology. At the minimum, research should include a measure of violent behavior, random assi gnment, and reliable and valid instruments. These qualities are particularly important when determining program effectiveness and aid in understanding whether program outcomes may be due to methodology, fidelity, or the nature of the intervention (Kerns & Prinz, 2002). Outcome measures are considered appropriate when they measure the desired behavior and when they are psychometrically sound. The outcome
56 measures chosen should reflect the goal of the intervention. For example, intervention programs may be evaluated in terms of any or all of th e following outcomes: changes in student attitudes concerning violence, knowledge of the curriculum, and/or change s in behaviors (Durant et al., 2001). A researcher also should interview and/or su rvey the participants. Student, teacher, and peer reports commonly are used in violence preven tion research and have been shown to be valid indicators of student behavior. In fact, self-r eport of conduct problems is one of the strongest predictors of later maladjustment, including delinquency conduct disord er in adolescence. Scholars suggest use of multiple sources and data comparison from these sources (Aber et al., 2003). The current study used a standardized measure of problem behavior and social skills as well as several other surveys which were administ ered to teachers and students. Additionally, direct behavioral observation met hods were used to provide inform ation from an outside source. The instruments are discussed in depth in th e methods and results section. Unfortunately, random assignment was not possible due to school district desires. However, pre and post intervention methods were used in the resulting quasi experimental design to strengthen outcome measurement. The major focus of the proposed study is to determine if the implementation of an elementary school, universal, clas sroom-based, violence prevention curriculum with and without an experiential add on results in observed and reported improvemen ts in students social skills and reductions in observed and reported aggression. The study will examine the following hypotheses.
57 Hypotheses Hypothesis 1. Individual students are expected to demonstrate different program benefits according to their intervention program involve ment (i.e., no treatment, curriculum only, and curriculum plus video). (A) Compared to studen ts in the curriculum only group, students in the curriculum plus video group will demonstrate gr eater program benefits. (B) Compared to students in the no treatment group, students in the curriculum only group will demonstrate greater program benefits. Program benefits will be measured by performance on outcome measures: peer nomination classification, Social Skills Rating System-student report standard scores, Social Skills Rating System-teacher repor t standard scores, number of correct content knowledge survey questions, and frequency of student reported aggressive and prosocial behaviors. Hypothesis 2 Intervention program involvement level (i.e., no treatment, curriculum only, and curriculum plus video) will predict students' response during an actual problem solving situation as measured by class wi de behavioral observations. (A ) Compared to students in the curriculum only group, students in the curriculum plus video group will demonstrate greater program benefits. (B) Compared to student s in the no treatment group, students in the curriculum only group will demonstrat e greater program benefits.
58Table 1-1. Florida violen ce statistics for 2003 and 2004 Year Population Murder Forcible sex offense Robbery Aggravated assault Total % Change Rate per 100K 3 17,071,508 924 12,756 31,512 79,044 124,236 -2.9 727.7 4 17,516,732 946 12,427 29,984 80,340 123,697 -0.4 706.2
59Table 1-2. Columbia County viol ence statistics for 2003 and 2004 Year Population Criminal homicide manslaughter Forcible rape, sodomy, fondling Aggravated assault Aggravated. stalking Threat/ Intimidation Simple assault Stalking Total Rate per 100K 3 58,890 1 14 92 0 0 400 10 517 877.9 4 60,453 0 5 102 0 8 385 1 501 828.7
60Table 1-3. Intervention program summary Study Program Name Level Ethnic/ Racial Majority Total N Males Lessons # of Sessions Primary SES Grade Rating Source Lochman (1985), Lochman, & Wells (1996) Anger Coping I Black U/K U/K 1 x per week for 10 weeks 10 V 35 Effective Clayton, Ballif-Spanvill, & Hunsaker (2001) Lochman (1985) Anger Coping Program with goal setting I Black 38 38 1-2 x per week for 12 weeks 18 V 47 Effective www.hamfish.org Lochman (1985) Anger Coping Program with goal setting I Black 38 38 1-2 x per week for 12 weeks 18 V 47 *** Derzon, Wilson, & Cunningham (1999) Tanner & Holliman (1998) Assertiveness Training I Caucasian 24 15 1-2 x per week for 3 weeks 4.5 U/K 13 ** Derzon, Wilson, & Cunningham (1999) Hudley (1991) BrainPower Program I Black 48 48 1-2 x per week for 6 weeks 9 V 57 *** Derzon, Wilson, & Cunningham (1999) Hudley & Graham (1993) BrainPower Program I Black 78 78 U/K U/K L 35 Effective Clayton, Ballif-Spanvill, & Hunsaker (2001) through **** indicate the degree to which treatment groups derived benefit. More indicate the treatment groups demonstrated higher program benefits.
61Table 1-3. Continued Study Program Name Level Ethnic/ Racial Majority Total N Males Lessons # of Sessions Primary SES Grade Rating Source Hudley, Britsch, Wakfield, Smith, Demorat, & Cho (1998) BrainPower Program I Black 166 166 1-2 x per week for 6 weeks 9 L 3-6 Positive results Fields & McNamara (2003) Hudley, Britsch, Wakfield, Smith, Demorat, & Cho (1998) BrainPower Program I Black 166 166 1-2 x per week for 6 weeks 9 L 3-6 Effective www.hamfis h.org Hudley, Britsch, Wakfield, Smith, Demorat, & Cho (1998) BrainPower Program I Black 166 166 1-2 x per week for 6 weeks 9 L 3-6 *** Derzon, Wilson, & Cunningham (1999) Coats (1979) Cognitive Self Instruction Training I U/K 16 16 1 x per day for 2 weeks 14 U/K 3 ** Derzon, Wilson, & Cunningham (1999) Roseberry (1997) Conflict Resolution Training U Black 173 81 1-2 x per week for 21.5 weeks 32 L 4-6 ** Derzon, Wilson, & Cunningham (1999) Feshbach (1983, 1984) Empathy Training Program U Caucasian U/K U/K 10 weeks U/K V 1-5 Promising Clayton, BallifSpanvill, & Hunsaker, (2001)
62Table 1-3. Continued Study Program Name Level Ethnic/ Racial Majority Total N Males Lessons # of Sessions Primary SES Grade Rating Source Dolan, Kellam, Brown, WerthamerLarson, Rebok, Mayer, Laudolff, Turkkan, Ford & Wheeler (1993) Good Behavior Game U Black U/K U/K U/K U/K U/ K U/K Effective Clayton, BallifSpanvill, & Hunsaker (2001) Based on multiple studies Good Behavior Game U U/K U/K U/K U/K U/ K U/K U/K Promis-ing Level 2 Youth Violence, A report of the Surgeon General (2003) WerthamerLarsson, Kellam, & Wheeler (1991) Good Behavior Game U Black 363 178 3-4 x per week for 24 weeks 84 V 1 Effective www.hamfis h.org WerthamerLarsson, Kellam, & Wheeler (1991) Good Behavior Game U Black 363 178 3-4 x per week for 24 weeks 84 V 1 ** Derzon, Wilson, & Cunningham (1999) Based on multiple studies Good Behavior Game U U/K U/K U/K U/K U/K U/ K U/K Promis-ing Mihalic, Irwin, Elliot, Fagan, & Hansen (2001)
63Table 1-3. Continued Study Program Name Level Ethnic/ Racial Majority Total N Males Lessons # of Sessions Primary SES Grade Rating Source Based on multiple studies I Can Problem Solve U U/K U/K U/K U/K U/ K U/K U/K Promising Level 2 Youth Violence, A report of the Surgeon General (2003) Based on multiple studies I Can Problem Solve U U/K U/K U/K U/K U/K U/K U/K Effective www.hamfis h.org Based on multiple studies I Can Problem Solve U U/K U/K U/K U/K U/K U/ K U/K Uncoded Derzon, Wilson, & Cunningham (1999) Based on multiple studies I Can Problem Solve U U/K U/K U/K U/K U/K U/ K U/K Promising Mihalic, Irwin, Elliot, Fagan, & Hansen (2001) Shure (1997), Shure & Spivak (1988) I Can Problem Solve U U/K U/K U/K U/K U/K U/K U/K Effective Greenberg, Domitrovich, Bumbarger (2001) Prinz, Blechman, & Dumas (1994) Prosocial Coping Skills Training U Unknown 25 classes Unknow n Weekly U/K U/K 1-3 Promising Clayton, BallifSpanvill, & Hunsaker (2001) Marvel, Mareda, & Cook (1993) Rules for Fighting Fair U Black 193 U/K 3-4 x per week for 7 weeks 24.5 L 4-6 Derzon, Wilson, & Cunningham (1999)
64Table 1-3. Continued Study Program Name Level Ethnic/ Racial Majority Total N Males Lessons # of Sessions Primary SES Grade Rating Source Grossman, Neckerman, Koepsell, Liu, Asher, Beland, Frey, & Rivara (1997) Second Step U Caucasian 790 424 1-2 x per week for 16-20 weeks 27 V 2-3 Mixed results, overall effective Fields & McNamara (2003) Grossman, Neckerman, Koepsell, Liu, Asher, Beland, Frey, & Rivara (1997) Second Step U Caucasian 790 424 1-2 x per week for 16-20 weeks 27 V 2-3 Effective Cooper, Lutenbacher, & Faccia (2000) Grossman, Neckerman, Koepsell, Liu, Asher, Beland, Frey, & Rivara (1997) Second Step U Caucasian 790 424 1-2 x per week for 16-20 weeks 27 V 2-3 Noteworthy www.hamfis h.org Grossman, Neckerman, Koepsell, Liu, Asher, Beland, Frey, & Rivara (1997) Second Step U Caucasian 790 424 1-2 x per week for 16-20 weeks 27 V 2-3 Derzon, Wilson, & Cunningham (1999) Taub (2001) Second Step U Caucasian 620 U/K 2 x per week for 8 weeks U/K V 3-6 Effective by authors Taub (2001)
65Table 1-3. Continued Study Program Name Level Ethnic/ Racial Majority Total N Males Lessons # of Sessions Primary SES Grade Rating Source Grossman et.al (1997) Second Step Same as above U Caucasian 620 U/K 2 x per week for 8 weeks U/K U/K U/K Effective Greenberg, Domitrovich, & Bumbarger (2001) Gesten, Rains, Rapkin, Weissberg, Flores, Cowen, & Bowen (1982) Social Problem Solving Curriculum U Caucasian 133 72 1-2 x per week for 9 weeks 13.5 V 2-3 Derzon, Wilson, & Cunningham (1999) Weissberg, Gesten, Rapkin, Cowen, Davidson, de Apodaca, & McKim (1981) Social Problem Solving Training U Caucasian 243 U/K 3-4 x per week for 14 weeks 49 V 3 *** Derzon, Wilson, & Cunningham (1999) Weissberg, Gesten, Carnulce, Toro, Davidson, & Cowen (1981) Social Problem Solving Training U Caucasian 563 U/K 3-4 x per week for 14 weeks 49 U/K 2-4 ** Derzon, Wilson, & Cunningham (1999) Nelson & Carson (1988) Social Problem Solving Training U Caucasian 101 U/K 1-2 x per week for 18 weeks 27 V 3-4 Derzon, Wilson, & Cunningham (1999) Coie & Dodge (1988) Social Relations Training S Black 24 12 1-2 x per week for 17 weeks 25.5 L 4 *** Derzon, Wilson, & Cunningham (1999)
66 CHAPTER 2 METHODS Participants Participants in the Best Acti ng Video Project included 129 f ourth grade children in eight classrooms in two schools in the Columbia C ounty School District. The intervention groups were selected based on principals and classroom teachers willingness to participate. Some students for whom data were collected were not in cluded in the final sample as mentioned above. Table 2-1 includes a summary of st udents who were and were not incl uded in the final data set. These students were those who withdrew from the assigned school, enrolled following the start of the intervention, or who could not participate in the interventi on. Demographic characteristics of the sample are presented in Table 2-2. Ov erall, 66 boys and 63 girls constitute the final sample. In the curriculum only group, 42 (57%) boys and 31 girls participated. 74% (54) were white, 19% (14) were black, and 7% (5) were Hi spanic. In the curriculum plus video group, 17 (41%) boys and 25 (59%) girls part icipated. In this group 81% we re white, 19% were black, and none were Hispanic. In the no interventi on (control) group, 7 (50%) boys and 7 girls participated. Eighty percent (11) were white, 13 % (2) were black, and 7% (1) was Hispanic. Intervention/Treatment Groups Table 2-3 summarizes information on levels of treatment for all groups. Students participating in the no treatment control group rece ived no intervention associated with the study. Further, this group received no violence preven tion interventions administered by the school district or other agency. Becau se the district was participa ting in the Safe Schools Healthy Students Initiative, many other schools in the district were engaged in other preventative interventions such as Too Good for Drugs and T oo Good for Violence. Further, because the district was participating in a uni versal prevention effort, district personnel promoted the use of
67 interventions in as many classes as possible. Therefore, classrooms not engaged in these or similar programs were not chosen for the no tr eatment intervention. Additionally, the trained observers were not available to travel to the one othe r available site due to funding limitations. Students participating in the curriculum groups received instruction with the PATHS curriculum for 18 weeks. These lessons spanned a 40 minute interval and occurred once a week. Students participating in the curr iculum plus video groups received instruction fo r 18 weeks with the PATHS curriculum and participated in scri pt development and vi deo production. Script development included five one hour sessions of script writing a nd curriculum review. During these sessions students gathered in small gr oups, were given a written example of an interpersonal problem. The guidance counselors a nd primary investigator first encouraged the children write down a positive and negative outcome of the problem situation. They were then asked to read the PATHS problem solving steps wh ich were posted in the classroom. They were then asked to take one step at a time and write wh at a child in the proble m situation would say to themselves or others during a given step. Th e guidance counselors and primary investigator circulated throughout the room to provide guida nce and feedback to the children on how to incorporate the steps in to their script. Video production included five sessions of practice and one session of filming. Practice sessions were limited to 20 minutes and filming se ssions varied from 20 minutes to one hour as some students remembered their parts better than others. Researchers have suggested interventions may be more succe ssful in increasing ch ildrens prosocial behaviors and decreasing aggressive behaviors when the interventions require children to engage in more experiential components such as social skills training through role plays, social skills pr actice with a peer role
68 model, and psychodrama (Bosworth et al., 1998 ; Cirillo, Pruitt, & Colwell, 1998; Clayton, Ballif-Spanvill, & Hunsucker, 2001; Tobin, & Sprague, 2000). School Characteristics. School characteristics based on data for the 2004-2005 intervention year are summarized in Tables 2-4 (i.e., FCAT and fr ee lunch data) and 2-5 (incidents of reported crime and violence in Fl orida elementary schools). Level 3 or higher FCAT math scores were achieved by 36% of st udents at Eastside Elementary and 57% of students at Fort White Elementary. Level 3 or higher FCAT reading scores were achieved by 81% of students at Eastside elementary and 68% of students at Fort White. Passing scores on FCAT writing were achieved by 90% of students at Eastside Elementary and 87% at Fort White. Free and reduced lunch statistics were available for the schools at which the interventions occurred. Children from families with incomes at or below 130 percent of the poverty level are eligible for free meals. Those with incomes be tween 130% (i.e., 24,506 for a family of four) and 185% (i.e., 34,873 for a family of four) of the pove rty level are eligible for reduced-price meals, for which students can be charged no more than 40 cents. The percentage of children who qualified for free and reduced lunch at East side Elementary was 59.8% and Fort White Elementary was 62.2% (Florida Department of Education, 2007). Table 2-5 summarizes the reported violence statistics for the two schools. For the intervention year, 28 incidents of fighting/harassment were reported in district elementary schools with 2 occurring Eastside Elementary and 1 at Fort White Elementary. Thus the incidents of serious offenses recorded at the intervention sites were low. Procedures Choice of the curriculum, measurement instru ments, implementation schedule, and video process were guided by the proce ss and results of a 2003-2004 pilot st udy that occurred at one of
69 the intervention sites (School 2 or Eastside Elementary). Furthe r detail regarding these issues will be discussed in following sections. The number and training needs of the two obser vers, two guidance counselors, as well as time constraints of the primary investigator requ ired staggered implementation of the program. Additionally, consecutively occurring hurricanes in the fall of 2004 delayed its start. Therefore, video plus curriculum groups (which required more intervention weeks) were needed at the first implementation site: Eastside Elementary. East side Elementary also was the obvious choice for the video plus curriculum groups as the sc hool possessed video and editing equipment. Additionally, Eastside Elementary was the choice for the no treatment group. School staff chose to include one classroom in the no interventi on group due to their overwhelming desire to participate in the intervention. At Eastside Elementary, 11 ne w students enrolled following pretesting and an additional five students in the curriculum plus video group withdrew from the school. Four students in the no intervention group enrolled following pretesting and one withdrew from school. These students were not included in the final data set. The primary investigator collected data from students who were out of school at the time of po sttesting during the foll owing two school days. One student in the curriculum plus video gr oup was removed from the pool at the parents request. The guidance counselor co nsulted with the parents concerni ng their desire to have their child excluded from the intervention. The parents reportedly were experiencing discipline problems at home and withdrew the child as a punishment. The guida nce counselor supported the parents right to do so a nd spoke to the parents regardi ng the possible benefits of the program. The parents elected to remove the child.
70 Students at Fort White Elementary were pa rt of the curriculum only group. Classroom enrollment was fairly stable, with one child entering school during th e intervention and one student leaving each classroom during the spring se mester. These students were excluded from the study. Four students were abse nt on the day of the posttest for school 1. These students were excluded from the study due to the end of the school year and resulting inab ility to collect these data. Additionally, two students we re excluded from the study at school 1 due to an inability to complete preand post-intervention materials. As mentioned previ ously, preand postintervention materials were administered by the primary investigator and one of two guidance counselors. All efforts were taken to ensu re the students followed along in answering the questions. At each school, if a student appear ed to be falling behind when completing the administered materials, the prim ary investigator or guidance couns elor worked individually with the student to complete them. However, two stud ents at Fort White Elementary were unable to complete these materials even w ith individual assistan ce. These students a ppeared to adopt an impulsive response style and were un able to explain to the primary investigator the content of the questions being asked. Curriculum Curriculum choice was guided by theoretical underpinnings, empirical evidence, and practical considerations. The I Can Problem Solve (ICPS) curriculum was used during the 2003-2004 pilot study. This curriculum was divided into early and late elementary curricula. The older elementary participants demonstrated difficulty understanding the many problem solving steps presented in the curric ulum. Furthermore, the early elementary curriculum activities and language were not judge d to be age appropriate and lacked efficacy with fourth grade students Additionally, the guida nce counselor and primary investigator experienced difficulty presenting the materials in their entirety during the allotted classroom time
71 and within the intervention time frame. Fi nally, the PATHS program developer, Mark Greenberg, was available for c onsultation on the project. Intervention Overview. The Best Acting Video Proj ect lessons c onsist of group didactic instruction, small group role-play, and written assignme nts. The titles and subject matter of the 18 units are found in Table 2-6. Th e intervention team consisting of the primary investigator and the two guidan ce counselors taught the classroom-based lessons. Lessons at Eastside Elementary Lessons began on November 22nd and ended on April 4th. Script development and practice began on April 11th and ended on May 16th. Filming occurred during the weeks of May 16th. At Fort White Elementary, lessons began on January 6th and ended on May 9th. Guidance counselor training and intervention integrity. The intervention team attended a training session prior to the implem entation of the intervention to discuss the curriculum and related procedures. The tr aining session began by di scussing difficulties encountered during the pilot project (i.e., difficulty of the curriculum strict timeline, time needed to film and edit) and barriers to appropriate in tervention. The interven tion team discussed how to address potential problems (e.g., school st aff/student resistance, timelines, staggered implementation, and logistics). The interventi on team chose the curriculum following a brief presentation of research and curriculum materials. The tr aining session also included a discussion of the theory behind the curriculum a nd how the theory guided the intervention. The importance of following the curriculum to promote adequate intervention integrity and the timeliness of delivery also were discussed. Th e procedures to ensure optimal levels of communication (described below) were also discussed.
72 The intervention team met w eekly, in person, to discuss barriers to successful implementation and possible solutions to problem s. In addition, the primary investigator supervised/participated in the classroom interven tion and was present for at least half of the lessons taught by each guidance counselor. Following each intervention lesson, the primary investigator and guidance couns elor discussed progress, pr oblems, and solutions. The implementation of the curriculum wa s staggered; therefore, the prim ary investigator was able to serve as an observer of the intervention implem entation and consultant for the first guidance counselor and later observer of the intervention implementation and consultant for the second guidance counselor. The primar y investigator took notes on sessions provided by the first guidance counselor who had implemented the pi lot project thus provi ding input and ongoing training for the second guidance counselor. Th e two guidance counselors also communicated weekly to ensure continuity of the curriculum. The guidance c ounselors also received regular email updates from the principal investigator desc ribing project progress. No formal measures were used to verify intervention integrity. The high degree of communication between the guidance counselors and the primary investigat or helped ensure intervention integrity. Apparatus The methods selected to measure intervention ou tcome represent a wide array of constructs researchers have used in order to measure childrens social behavi or. Researchers generally have sought to measure social behaviors through the use of standardized reports from teachers, parents, and students. Additionally, other met hods such as peer nomination have been used routinely. Although each method has its stre ngths and weaknesses, researchers commonly believe the use of various instruments and the co llection of data from different sources (i.e., parents, teachers, peers, and student) is optimal in order to most effectively evaluate interventions (Fox & Bolton, 2005).
73 Measures or instruments used to assess the impact of the Best Acting Video Project on childrens social competence are described below. Table 2-7 outlines the timelines for data collection. The following sections present each method chosen and the rationale supporting its inclusion in the present study. Technical info rmation, including psychometric properties, and statistical analyses performed to prepare the data for inclusio n in the current study also are presented. The data were judged to be appropriate for inclusion in the final analysis if they met the basic assumptions for the analysis. Therefor e, the following paragraphs included discussion of the datas normality and homogeneity of va riance/covariance as suggested by correlational analyses. The PATHS Knowledge Survey Students knowledge of the curriculum preand post-intervention was assessed using 21 multiple choice questions. Students were asked to recall specific information from the PATHS curriculum post-intervention. Preexisting survey di fferences were measured by the pretest. This survey was developed by the primary investigat or with assistance from staff at The National Rural Behavioral Health Center and modeled after questions embedded in the lessons. The PATHS curriculum did not contain content knowledge tests. Thus, one needed to be developed. The numbers of items were taken equally from each lesson and were based on information from the pilot project as well as a survey for a pr oject involving Too Good for Drugs. The tests were administered preand post-interventi on by the intervention team (see Appendix A). Rationale for PATHS Knowledge Survey. The goal of the Best Acting Video Project was for students to acquire and use the curriculum content in their lives, thus leading to a reduction in aggression and an increase in pr osocial behavior. The PATHS knowledge survey was used to measure the effects of the intervention on curriculum knowledge.
74 Psychometric properties and data prep aration of the PATHS Knowledge Survey The number of correct responses was computed. A Q-Q plot procedure re vealed the data were fairly normally distributed. Table 2-8 includes de scriptive statistics for the pre-and post-PATHS knowledge survey. Mean increases from preto post-tests were more substantial in the no treatment group (+2.40) and the curriculum only gr oup (+1.74) than the curriculum plus video group (+1.29). Table 2-9 shows the percent correct for each pre-intervention and postintervention question. An item analysis was conducted using a single construct (content knowledge). First, an item analysis was computed for post-intervention responses for those who complete the survey. Items Q7 (-.140), Q13 (.046), and Q3 (.046) had the lowest corrected item-total correlation. An item analysis was rerun without item Q7 to ex amine whether items Q13 and Q3 would continue have a low correlation after deleting Q7. After rerunning the analysis, Q 13 (.040) and Q3 (.058) continued to have low item-total correlations. An examination of the content of Q7 revealed wording which was confusing and the use of all of these answer choice to be inappropriate. Thus, this item was deleted. An examination of the content of item Q13 revealed students had difficulty answering an application question. Slightly more than half of the students in th e curriculum and curriculum plus video groups selected the correct answer. Le ss than half (47 %) of the c ontrol group answered the question correctly. This question did not appear to di stinguish students who knew and did not know the content. Thus, it was eliminated. An examination of the content of Q3 revealed it to be one of the more difficult questions concerning PATHS pr oblem solving steps. Students were asked to identify a step in the middle of the problem so lving process. More than half (65%) of the students participating in the curr iculum selected the correct answ er. More than half (57%) of
75 students in the no intervention group also answered the question correctly. Thus, this item was eliminated because of its low corrected itemtotal correlation and discrimination power. The Chronbach alpha was .68 for the final scale. Furthe r item analyses revealed that deletion of the 9 items with the lowest item to total correlation (below .30) woul d not increase the Chronbach alpha and instead would decrease it to .634 for the remaining 11 items. Table 2-10 includes the corrected descriptive statistics for the PATHS knowledge survey following deletion of items Q3, Q 7, and Q13. The mean number correct on the posttest for the no treatment group is numerically smallest (14. 93), followed by the curriculum plus video group (15.68), and finally, the curri culum only group (16.32). Following the deletion of the three items, th e data from the PATHS knowledge survey were assessed for normality using SPSS Q-Q plot procedure. Visual inspection of the preintervention plot revealed seve ral low values outside the line trajectory. However, the data appeared to approach normality. Inspection of the post-intervention pl ot indicated further investigation into the normality of the data was wa rranted as the data did not approach normal. SPSS was used to create a histogram for the preand post-intervention data sets. The preintervention data approached a normal distributio n. However, post-interv ention data displayed a negatively skewed distribution. According to Tabachnick & Fidell (2001 ), even with unequal sample sizes and only a few dependent variable s, a sample size of approximately 20 in the smallest cell should ensure robustness in a MANCOVA design. They propose assessing normality with good judgment in light of whether the dependent variables are expected to be fairly normally distributed. If the data are ex pected to be normally di stributed, they suggest transforming the data.
76 The distribution of the post-intervention PA THS knowledge survey data may be expected to have negatively skewed distribution. This sk ewness may occur due to the small sample in the no treatment group as well as to th e high percentage of correct ite ms even on the pre-intervention PATHS knowledge survey. According to Leech, Barrett, & Morgan (2005), skewness should be less than plus or minus 1.0. For the post-int ervention data, skewness was 1.791. These data were transformed using methods logX, X, 1/X, and 1/X. None of these transformations produced data more normally distributed than the original variable data. Additionally, the data set was recalculated and transfor med with question Q3 included. None of the transformations produced a more normal distributio n than the original variable data. Thus, the PATHS Content Knowledge survey data were judged to be inappr opriate for inclusion in the final analysis. Best Acting Report of Social Behaviors Students were asked to report the frequency of observed and experi enced prosocial and aggressive behaviors preand post-intervention. The survey re quired students to recall how often (i.e., never, sometimes, often, always) they either observed or e xperienced prosocial or aggressive behaviors during the last two weeks (see Appendix B). Rationale for the Best Acting Report of Social Behaviors. Student reports of peer behaviors validly predict the level of aggressi ve behaviors in schools (Carmona, 2005; Werner & Crick, 2004). The Best Acting Report of Social Behaviors survey wa s designed to provide information regarding the number of prosocial and aggressive be haviors the students perceived were occurring among their classmates and themse lves. The survey was designed by the primary investigator with assistance from staff at The National Rural Behavioral Health Center. The questions were based on questions embedded in The Florida Youth Risk Behavior Survey (YRBS). The YRBS, a self-administered school-based anonymous survey, has been administered to Florida public hi gh school students every two year s since 1991. It is part of a
77 national survey effort led by the Centers for Dise ase Control and Preventi on to monitor priority health risk behaviors that contribute markedly to the leading causes of death, disability and social problems among youth in the United States. In 2003, 4,080 students in 75 public high schools in Florida completed the YRBS (Florida State Department of Health, 2003). Psychometric properties and data preparatio n of the Best Acting Report of Social Behaviors. Students were asked to report the number of times they engaged in aggressive (questions 1-6) and prosocial (questions 7-13) beha viors. Additionally they were asked if one or more classmates engaged in aggression towards them (questions 14-19) or if they had seen aggressive (questions 20-25) or prosocial (questions 26-30) behavi ors directed towards classmates during the past two weeks. Stude nts selected from the following options when answering each question: many times (4), often (3 ), sometimes (2), and never (1). Therefore, higher numbers indicate more fre quent incidents of behaviors. The highest possible score was 24 for self reported aggression, 28 for self reported prosocial be havior, 24 for self reported victimization, 24 for witness of classmates aggressive behavior, and 20 for witness of classmates prosocial behavior. See table 2-11 fo r ranges of scores. Normality of these groups was examined using SPSS and the Q-Q plot procedure. Visual inspection of the Q-Q normal probability plots using SPSS indicated the scores for the pre-and post-frequency measures were fairly normally distributed and out liers were not problematic. De scriptives for these data are presented in table 2-12. In preparation for including these data in the final analysis, Pearson correlations were calculated to determine relationships between the groups of questions mentioned above. The frequency of pre-intervention stud ent self reported aggression correlated significantly with the post-intervention studen t self reported prosocial beha vior, r (127) =.192, p < .05. Pre-
78 intervention self reported prosoc ial behavior was correlated sign ificantly with post-intervention self reported aggressive behavi or, r (127) =-.214, p<.05. Reports of pre-intervention witnessing prosocial behavior correla ted significantly with post-interventi on reports of victimization, r (127) =.208, p < .05. These data appear to be appropr iate for inclusion in the final analysis. Additionally, a preliminary MANCOVA anal ysis was conducted using the groups of answer choices outlined above to assess redundanc y of the groups for inclusion in the final analysis. The use of the preliminary MANC OVA controlled for possible pretest score differences. Boxs test of equality of covari ance (p = .43) revealed differences between the covariance matrices of the dependent variables we re not significant. Additionally, Levenes test of equality of error variance indicated the assumption of homogeneity of variance was not violated for student reports of aggression, prosocial behavior victimization, and witnessing prosocial behavior. However, Le venes test indicated a violati on (p = .048) for the variable witnessing aggressive behavior. Nevertheless, the MANCOVA was conducted using the data in that Boxs test was not significant and Leve nes test just reac hed significance, The three way between-groups MANCOVA revealed that the multivariate main effect of intervention level was significant: Wilks = .841, F (124, 234) = 2.12, p = .024, multivariate = .28, indicating a small to moderate effect size. Examination of the coef ficients for the linear combinations that distinguish the three interven tions indicated that st udent self reported postintervention prosocial behavior co ntributed most to distinguishi ng the intervention groups. In particular, student self reported prosocial beha vior and student report of witnessing prosocial behavior contributed signifi cantly toward discriminating the no intervention group from the curriculum and curriculum plus video groups (p = .015 and p = .022, respectively). Thus, these variables will be used in the fina l analysis. No variables contribut ed significantly to attempts to
79 distinguish the curriculum only group from the no intervention and curriculum plus video groups. Peer Nomination Survey Students completed a peer nomination survey on which they nominated three children for each of the following seven questions: (Who) do I lik e the most, fights the most, starts fights the most, do not know how to join in a group, says mean things, shares their things, and helps others. The primary investigator administered this su rvey. (See Appendix C for the peer nomination survey). Rationale for the peer nomination survey. Peer nomination, perhaps the oldest type of peer-referenced assessment, was used to better understand how students were viewed by their peers. Social status or peer relations, a subcomponent of social competence, often is considered to result from one's expression of social skills (Merrell, 2003). Data from this type of assessment provide information regarding persons social environment within the context which the information is taken (Frederickson & Furnham, 2004; Kamphaus & Frick, 2002; Wood, Cowan & Baker, 2002), including those children with disabilities (Bellan ti & Bierman, 2000; Hall & Strickett, 2002; Mu, Siegel, & Allinder, 2000) For example, a peer nomination survey administered in a school setting would provide in formation regarding that setting. Because the social environment is crucial to a persons psyc hological adjustment and well-being, information on how peers are perceived and perceive others is viewed as valuable when attempting to understand an individuals overall social functioning (Kamphaus & Frick, 2002). Several studies have examined the efficacy of peer referenced assessment. For example, Jiang and Cillessen (2005) conducted a meta-analysi s of more than 75 studies that utilized peerreferenced assessment to examine the techniques s hort-term stability (test-retest reliability) and long-term stability, as well as to identify factors that affect stability. These researchers found
80 that, on average, childrens rankings of accepta nce, rejection, social preference, and liking exhibited moderately strong te st-retest reliability, ranging fr om .70 to .82. Although, stability decreased significantly after one ye ar, it still remained relatively stable (ranging from an average of .52 to .58). Additionally, long-term stability was higher for older than younger children (e.g., preschool). Social status becomes more stable as children ag e (Coie & Dodge, 1983). Other research has examined the predictive validity of peer nomination. For example, Nelson and Dishion (2004) found that sociometri c status, as indicated by peer nominations, acquired at ages 9 and 10 was significantly correla ted with rates of antisoc ial behavior, history of arrests, and work-school engagement in early adulthood, even when controlling for academic performance and adolescent antiso cial behavior. In addition, soci al status as determined from peer referenced assessment in middle school predic ted long-term adjustment and rates of juvenile delinquency in adolescence (Ollendick, Weist, Borden, & Greene, 1992). An earlier metaanalysis (Parker & Asher, 1987) reported si milar findings. However, the aforementioned researchers reported the use of peer-referenced assessment procedures tended to contribute to few false negative errors yet cont ributes to many false-positive erro rs. That is, peer nomination procedures accurately predic ted which children would have undesirable outcomes, but the procedure also identified other children with relations hip problems who did not demonstrate later difficulties (Ollendick et al., 1992). Psychometric properties and data preparation for the peer nomination survey The typical level of interpretation of a peer nomina tion procedure is to deri ve Z-scores from the number of times a child was nominated for a partic ular category (i.e., prosocial, popular, socially awkward, and aggressive)(Kamphaus & Frick, 2002) This procedure also was used in the current investigation. Normality of the data then was assessed using the Q-Q plot method in
81 SPSS. Examination of Q-Q plots for these items revealed further examination of normality was warranted. SPSS was used to construct histogram s for the eight variables. This procedure revealed that seven of the eight distributions we re negatively skewed. Skewness was found to be greater than 1 for all va riables with the exception of the popul arity for both pre(.609) and post(.825) intervention scores. These variables were found to be inappropriate for inclusion in the final analysis due to their non normality. In addition, correlations for vari able pairs indicate violation of assumptions for the final analysis. The Spearmans rho correlations do not require linearity a nd were calculated (see Table 2-13). The analysis revealed that each pair of preand post-nomination Z-scores correlations were significant at the p <.01 level. Preand post-correlat ions were highest for aggressive (.73), then prosocial (.57), popular (. 56) and finally socially awkward (.40). Other highly correlated pairs included popular and pr osocial (pre-intervention) (.63) and postintervention (.63). Other significant as well as substantial correlations occurred between prosocial preintervention and popular post-inte rvention (.45), popular pre-in tervention and prosocial postintervention (.44), and socially aw kward pre-intervention and aggres sive preintervention (.42). Other pairs which suggest possible negative co rrelations were confirmed by the analysis. Significant negative correlations occurred between pairs of social ly awkward and prosocial preintervention (-.28), post-interven tion (-.23). Results for the pa irs of socially awkward and popular demonstrated a similar, although not alwa ys significant trend. Correlations between the pairs of aggressive and popul ar were not significant. Social Skills Rating System (SSRS): Teacher Report and Student Report The Social Skills Rating System (SSRS) (Gre sham & Elliot, 1990) provides a standardized norm-referenced assessment of childrens prosocia l and problem behaviors. Teachers completed
82 a SSRS for each student, thus providing informati on on two behavior domains: social skills and problem behaviors. Each st udent also completed a SSRS t hus, providing information on one domain: social skills. The SSRS includes the following subscales on th e teacher and student forms: cooperation (e.g., takes turns, shares, tolerant of others, em pathy, handles conflict ap propriately), assertion (e.g., initiates conversations/interactions with peer s, introduces self, participates in games), and self-control (e.g., relating to adults in an appr opriate manner). The teacher instrument also includes externalizing, internalizi ng, and hyperactive behavior scales The student instrument also includes an empathy scale (G resham & Elliot, 1990). Rationale for use of the SSRS. The SSRS (Gresham & Elliot, 1990) was found to be an appropriate measure for use in th e current investigation based on a review of the literature. Mental Measurement Yearbook (2004) reviewers of the SSRS found validity estimates to be acceptable to screen and classify students soci al behavior. Reliability estimates (internal consistency, interrater agreement, and test-retest stability) for the teacher and student scales are acceptable (Gresham & Elliot, 1990). Internal co nsistency is .94 for the elementary teacher report and is .83 for the student report. Internal consistency at or above .80 is considered acceptable (Duffey, Salvia, Tu cker, Ysseldyke, 1981). Interrater agreement is low for the student -teacher pairing (r = .22). However, these figures are similar to those reported by othe r researchers (e.g., Achenbach, McConaughty, & Howell, 1987). When assessing social competence, the use of self report measures introduces various quandaries. One may question whether children with social skills problems are competent to reliably report their own social skills deficits and strengths. In addition, individual student characteristics may influence the agreem ent between teachers and students. Crick and
83 Dodge (1994) submit that aggressive children posse ss social processing deficits, that cause them to interpret interpersonal information differently than their peers. These deficits seemingly interfere with the perceiving, dec oding, and interpreting of informa tion within social contexts. They propose that, when comparing teacher and peer ratings, aggressive children overestimate their cognitive competence and peer relationshi ps even though they ar e judged as being less preferred than their peers. In contrast, they propose that non-a ggressive children actually tend to underestimate their cognitive and social abilitie s in comparison to teacher reports and peer nominations (Crick & Dodge, 1994). The current study used the SSRS because it was thought to provide a broad picture of vari ed social competencies. The reliability of the SSRS student self repor t measure remains an issue. Test-retest reliability for elementary reports is .68. A recent study (Diperna & Volpe, 2005) included a sample of 185 students and 24 teachers in grades 3 to 5 and reports an internal consistency estimate of .86. The test-retest correlation over six months was .58. They report this moderate correlation is acceptable due to evidence that sugges ts students social behaviors as measured by social rating scales are moderately stable ove r time per the National Center for Education Statistics (Diperna & Volpe, 2005). Data preparation of the SSRS Raw scores for the two surveys were converted into standard scores (separated by gender) using methods as described in the SSRS manual. The teacher and student SSRS forms provide an option to first combine raw scores from the surveys and then convert them into subscale scores and su bsequent behavior levels. However, the use of the behavior levels in lieu of total standard scores for social skills and problem behaviors for the teacher version and social skills for the student ve rsion is not supported in the literature (Diperna & Volpe, 2005). In light of this informati on, the teacher reported social skills and problem
84 behaviors standard scores as well as the student reported socials ski lls standard score are used in the final analysis. The SSRS data were analyzed using Q-Q box plots to determine if they were normally distributed,. The SSRS teacher and stud ent reports approached a normal distribution and contained no outliers. Table 2-14 provides de scriptive statistics for the two teacher report subscales and the student self report s cale. Table 2-15 provides correlations. Teachers 1 though 4 were in the curriculum onl y group, 5 through 7 were in the curriculum plus video group, and teacher 8 was the no interv ention group. This SSRS data set is missing values from teacher 7 in the curriculum plus video group and teacher 1 in the curriculum only group. Efforts to obtain these data were unsuccessful despite offering incentives to the teachers. These missing data are problematic in light of the final analysis for Hypothesis 1. Several methods are available to replace missing values. According to Tabachni ck and Fidell (2001), mean substitution provides a conservative best gu ess method about the true value of a variable. Using this method, the mean for the distribution do es not change. However, the variance of the data is reduced by an unknown amount dependi ng on the actual values that are missing. Therefore, the data set used for the fina l analysis did not include teachers 1 and 7. Best Acting Behavioral Observation Checklist Behavioral observations were conducted in the eight partic ipating classrooms using the Best Acting Behavioral Observ ation Checklist. The observer recorded the frequencies of prosocial and aggressive behaviors as well as student responses to prosocial and aggressive behaviors. The form and proce dure were created for use in this study and was adapted from the work of Jollivette et al. (Jollivette, Stitchne r, Sibilsky, Scott, & Ridgeley, 2002). See Appendix D. Behavioral observations rationale. Behavioral obs ervations were used to determine frequencies of aggressive and prosocial behaviors at post-inte rvention to examine if student
85 response could be predicted by intervention grou p membership. This measure was designed to better understand the overall level of appropriate in-vivo problem solving skills. The use of behavioral observations is salien t in that rates of responding are di rectly observable and sensitive to environmental changes. In addition, behaviors observed in natural settings provide ecologically valid da ta (Hintze, 2002). Current methodology in the area of classwide observations is lacking, particularly for observations geared at recording classroom data as opposed to individual child data. A PSYCH LIT search performed on Oct ober 5, 2006 using the keywords obs ervation methods and humans yielded 77 articles. Of those articles, 14 in cluded discussion and research of observation methods and none discussed observations of gr oup data. A search of the online Mental Measurement Yearbook in 2006 failed to identify measures designed to r ecord group behavior data. Due to the lack of information in this area, the primary investigator consulted with Dr. Terrance Scott in the Department of Special Education at the Universi ty of Florida. He directed the principal investigator to the idea of recording opportunities for problem solving and the resulting behaviors from the child ren involved in the interaction. This method was derived from functional behavioral assessment methods. Spec ifically, the opportunity for problem solving was conceptualized as an antecedent trigger and the childs response was conceptualized similarly to the problem behavi or or replacement behavior. The degree of change in an i ndividuals social be havior was determined using the SSRS teacher and student report, as discussed earlier in this section. The obser vational data indicated if and when prosocial or aggres sive behaviors occurred as well as the type of response that followed among groups of classmates.
86 Method of observation. Students were observed in seve ral different settings: regular education classroom during academic instruction, transition from one academic subject to another within the classroom, free or non-instruct ional time, lunch time, and transition between and within school locations. Playground activi ties were not observed because intervention groups co-mingled with other intervention and no n-intervention groups during that time. Each classroom was observed an equal amount of time in the various settings. Event recording methods were used to record opportunities for pros ocial or aggressive behaviors. Opportunities are de fined as discrete events in which a student engages another student in an interaction that can result in the need to use problem solving skills. Event recording is an appropriate technique for collecting data in this investigation because aggressive and prosocial behaviors generally are discrete and of low frequency. Researchers have promoted this method as the least disruptive technique for r ecording observations (Gre enberg et al., 2001). Table 2-16 outlines the behaviors recorded as well as their operational definitions. Training for behavioral observations. Two observers were used to observe and code classroom observations. They were informed of the definitions, provided examples of behaviors, and received training in use of the record form. The two hour training also included discussion of objective assessment, expected time commitment, and observer duties and responsibilities. The observers then took part in three hours of classroom-based training. Following training, inter-observer agreement using live trials was conducted over 1.5 hour period. The interobserver agreement during this live trial was .85 for the events recorded and .80 for agreement regarding the categorizat ion of initiation and response. The observers then collected data separately a during three-hour blocks for each cla ssroom. The observers were blind to the nature
87 of the treatment. Observers met immediately preceding and following the trial observation periods to discuss definitions, problems encountered, and possible solutions. Statistical Analysis Hypotheses 1 Students were expected to demonstrate differe nt benefits according to their intervention program involvement (i.e., curriculum only, curric ulum plus video, and no treatment). More specifically, compared to students in the curric ulum only group, students in the curriculum plus video group will demonstrate greate r program benefits. Additionally, compared to students in the no treatment group, students in the curricul um only group were expected to demonstrate more benefit from the program. Program benefits were measured by performa nce on the following five outcome measures: peer nomination, Social Skills Rating System-st udent reports, Social Skills Rating Systemteacher reports, PATHS Knowledge Survey, and the Best Acting Report of Social Behaviors. Relationships between intervention program involvement or level of intervention (curriculum only, curriculum plus video, and no treat ment,) and social behaviors, as measured by performance on outcome measures, were investigat ed using multivariate analysis of covariance (MANCOVA). This procedure is a ppropriate when pretest scores se rve as covariates for a threeway between-subjects design with treatment gr oups serving as the independent variables and posttest scores on the outcome measures serving as dependent variables. This procedure allows for comparisons among groups and for evaluating th e influence of the independent variables and their interactions on the va rious dependent variables. When using MANCOVA, clearer results often are achieved using the fewest variables. That is, the addition of variab les and the slight improvement due to their addition do not compensate for the cost in higher degrees of freed om and reduction in the pow er of the analysis.
88 The preceeding sections presented data from each of the five measures and how the data were chosen to be included in the fi nal MACOVA design for hypothesis 1. Hypothesis 2 Intervention program involvement level (i.e., curriculum only, and curriculum plus video, and no treatment) was expected to predict student s' responses during an actual problem solving situation as measured by class wi de behavioral observations. Mo re specifically, compared to students in the curriculum only group, students in the curriculum plus video group were expected to demonstrate increases in appropriate soci al behaviors and decr eases in aggression. Additionally, compared to stude nts in the no treatment group, st udents in the curriculum only group were expected to demonstrat e increases in appropriate soci al behaviors and decreases in aggression. This hypothesis was investigated using binom ial logistic regressi on. This analysis determined if the categorical independent variab les (also predictors) of no treatment, curriculum only, and curriculum with video gr oups predict students prosocia l or aggressive behavior as coded by the observers on the behavioral observation form. Binomial logistic regression was used to com pute an odds ratio for each of the independent variables in the model, resulting in specifying the probabilities of part icular outcomes. The regression equation predicted the pr obability of whether an indivi dual will fall into one or the other categories (e.g., prosocial or aggressive). This procedure requires no assumptions about the distributions of the predictor variables. Bi nomial logistic regression may be used to analyze various kinds of data as pr edictor variables (e. g., categ orical, nominal, ordinal) Assumptions Several assumptions underlie this study. They generally relate to curriculum implementation and outcome measurement. Th e first acknowledges differences in students
89 ability to learn the steps of problem solving and related facts would not exists between the intervention groups. Additionally, the curriculum instructors behaviors were assumed not to differ between classrooms; they followed the curriculum; that the primary investigators behavior did not differ between classes or sc hools; that teachers completed rating scales honestly; that students answered rating scales and que stionnaires honestly; and observers were blind to the treatment conditions and recorded be havioral information accurately and discretely.
90 Table 2-1. Students not included in final sample School Treatment group Teacher Total N N not enrolled preN not enrolled postN out postN unable to complete Total 1 20 1 1 1 1 16 2 19 0 1 1 1 16 3 23 0 1 1 0 21 4 22 0 1 1 0 20 1 Curriculum Subtotal 84 1 4 4 2 73 5 21 5 4 0 1 11 6 16 3 0 0 0 13 2 Curriculum plus video 7 22 3 1 0 0 18 Subtotal 59 11 5 0 1 42 2 No intervention 8 19 4 1 0 0 14 Total 162 16 10 4 3 129
91Table 2-2. Students by treatment, gender, and race Race Intervention level White Black Hispanic Total Curriculum Boy Girl 32 22 7 7 3 2 42 31 Total 54 14 5 73 Curriculum plus video Boy Girl 14 20 3 5 0 0 17 25 Total 34 8 0 42 No intervention Boy Girl 6 5 0 2 1 0 7 7 Total 11 2 1 14
92Table 2-3. Treatment levels Curriculum only Curriculum and video No treatment Fort White Elementary I X Fort White Elementary II X Fort White Elementary III X Fort White Elementary IV X Eastside Elementary V X Eastside Elementary VI X Eastside Elementary VII X Eastside Elementary VIII X
93Table 2-4: State standardized testing re sults and free and reduced lunch percentage FCAT Math 4th State District Eastside Fort White Students tested 19,5866 742 81 102 % scoring at Level 1 15 20 22 15 % scoring at Level 2 12 30 42 28 % scoring at Level 3 38 38 27 45 % scoring at Level 4 31 11 9 10 % scoring at Level 5 6 1 0 2 FCAT Reading 4th State District Eastside Fort White Students tested 19,5678 737 81 102 % scoring at Level 1 15 14 4 17 % scoring at Level 2 13 14 17 16 % scoring at Level 3 35 38 41 30 % scoring at Level 4 29 30 35 34 % scoring at Level 5 8 4 4 4 FCAT Writing State District Eastside Fort White Number tested 19,4661 Unavailable 80 103 Percent passing 90 90 90 87 % Free and reduced lunch 53.1 60.7 59.8 62.2
94Table 2-5: Incidents of crime and violence in Florida elementary schools State District Eastside Fort White Against persons 2,858 4 0 2 Alcohol, tobacco, drugs 445 3 0 0 Against property 2,379 2 2 0 Fighting/harassment 14,003 28 2 1 Weapons 1,028 9 2 0 Other disruptive conduct 2,239 10 0 0
95Table 2-6. PATHS lessons Lesson 1 Introduction to problem solv ing / Introduction to identify Lesson 2 Problem identification Lesson 3 Identifying feelings / Goals Lesson 4 Solutions / Choices Lesson 5 Review / Generating alternatives Lesson 6 Generating alternatives / Li stening to others alternatives Lesson 7 Consequences / Choosing the best solution Lesson 8 Making a good plan Lesson 9 Trying your plan a nd evaluating what happens Lesson 10 Trying again Lesson 11 Trying again Lesson 12 Problem solving practice Lesson 13 Problem solving practice Lesson 14 Problem solving practice Lesson 15 Problem solving practice Lesson 16 Review Lesson 17 Review Lesson 18 Review
96Table 2-7. Data collection timeline Event 1 2 3 4 5 6 7 8 Observation 01/05 05/16 01/05 05/16 01/05 05/16 01/05 05/16 10/28 05/17 10/28 05/17 10/28 05/17 10/28 05/17 Social Skills Rating System student form 01/05 05/16 01/05 05/16 01/05 05/16 01/05 05/16 11/03 05/17 11/03 05/17 11/03 05/17 11/03 05/17 Social Skills Rating System teacher form 01/05 05/16 01/05 05/16 01/05 05/16 01/05 05/16 11/03 05/16 11/03 05/16 11/03 05/16 11/03 05/16 PATHS Content Knowledge Survey 01/05 05/16 01/05 05/16 01/05 05/16 01/05 05/16 11/08 05/16 11/08 05/16 11/08 05/16 11/08 05/16 Best Acting Report of Social Behaviors Survey 01/06 05/18 01/06 05/18 01/06 05/18 01/06 05/18 10/29 05/19 10/29 05/19 10/29 05/19 10/29 05/19 Peer nomination survey 01/06 05/18 01/06 05/18 01/06 05/18 01/06 05/18 11/03 05/19 11/03 05/19 11/03 05/19 11/03 05/19
97Table 2-8. PATHS knowledge survey de scriptives Intervention group N Mean correct prepostMean difference SD prepostMinimum prepostMaximum prepostCurriculum 73 16.01 17.75 +1.74 2.16 2.01 8 9 20 21 Curriculum plus video 42 15.76 17.05 +1.29 1.96 2.60 10 9 19 21 No intervention 14 15.80 17.40 +2.40 1.66 2.69 14 11 19 21
98 Table 2-9. Percent correct on PATHS knowledge survey % Correct Question Pre-Intervention Post-Intervention % difference Q0 70.5 97.3 + 26.8 Q1 81.2 83.2 + 2.0 Q2 87.0 89.9 + 2.9 Q3 19.5 64.4 + 44.9 Q4 65.1 71.8 + 6.7 Q5 84.6 96.0 + 11.4 Q6 84.6 89.9 + 5.3 Q7 32.2 34.2 + 2.0 Q8 94.0 97.3 + 3.3 Q9 69.1 85.9 +16.8 Q10 95.3 95.3 0 Q11 92.6 97.3 + 4.7 Q12 89.9 94.6 + 4.7 Q13 53.0 59.1 + 6.1 Q14 61.7 75.8 +14.1 Q15 78.5 85.9 + 7.5 Q16 95.3 94.0 1.3 Q17 89.3 94.6 + 5.3 Q18 81.2 89.3 + 8.1 Q19 69.1 78.5 + 9.4 Q20 79.9 88.6 + 8.7
99 Table 2-10. PATHS knowledge survey de scriptives following item analysis Intervention group N 129 Mean correct prepostMean difference SD prepostMinimum prepostMaximum prepostCurriculum 73 15.10 16.33 +1.23 1.87 1.98 10 7 18 18 Curriculum plus video 42 14.88 15.83 +.95 1.80 2.01 8 9 17 18 No Intervention 14 15.07 14.93 -.14 1.49 3.22 13 7 18 18
100Table 2-11. Best Action Social Be havior Survey question content Question content Number of ques tions Range of possible scores Self report of aggre ssive behavior 6 6-24 Self report of prosocial behavior 7 7-28 Self report of victimization 6 6-24 Witness classmates aggressive behavior at school 6 6-24 Witness classmates prosocial behavior at school 5 5-20 Total of reported aggressive behavior 18 18-72 Total of reported prosocial behavior 12 12-48
101Table 2-12. Best Acting Social Behaviors Survey descriptives Self report-preintervention Witness-preintervention Self report-postintervention Witness-postintervention Intervention group Aggressive Prosocial Victim Aggressive Prosocia l Aggressive Prosocial Victim Aggress Prosocial No Intervention M SD 8.07 1.98 21.00 4.14 11.60 5.10 12.40 4.44 13.60 4.31 9.13 2.64 18.07 5.20 11.93 5.33 13.87 4.87 10.87 4.033 Curriculum M SD 8.57 2.90 21.04 4.894 9.82 3.31 12.99 4.44 13.79 4.11 8.94 3.20 19.75 4.25 10.04 4.19 13.00 4.45 13.97 3.764 Plus Video M SD 8.50 2.32 20.95 4.423 10.79 4.24 12.83 3.99 14.10 4.48 8.24 2.53 20.90 4.21 11.52 4.59 13.31 5.04 13.26 3.794 Total M SD 8.49 2.61 21.01 4.629 10.34 3.88 12.87 4.27 13.87 4.23 8.74 2.93 19.93 4.40 10.74 4.50 13.20 4.67 13.38 3.898
102Table 2-13. Peer nomination co rrelations: Spearmans rho Category Prosocial PrePopular PreAwkward PreAggressive PreProsocial PostPopular PostAwkward PostAggressive PostProsocial Pre1.0 .634** .000 -.227** .010 -.250** .004 .573** .000 .448** .000 -.330** .000 -.168 .057 Popular Pre.634** .000 1.000 -.274** .002 -.017 .847 .441** .000 .563** .000 -.270 .002 .122 .170 Awkward Pre-.227** .010 -.274** .002 1.000 .420** .000 -.232** .008 -.149 .093 .399** .000 .210* .017 Aggressive Pre-.250** .004 -.017 .847 .420** .000 1.000 -.267** .002 -.093 .292 .147 .098 .734** .000 Prosocial Post.573** .000 .441** .000 -.232** .008 -.267** .002 1.000 .633** .000 -.345** .000 -.277** .002 Popular Post.448** .000 .563** .000 -.149 .093 -.093 .292 .633** .000 1.000 .-.331 .000 -.066 .454 Awkward Post-.330** .000 -.270** .002 .399** .000 .147 .098 -.345** .000 -.331** .000 1.000 .152 .086 Aggressive Post-.168 .057 .122 .170 .210* .017 .734** .000 -.277** .002 -.066 .454 .152 .086 1.000 ** Correlation is significant at the 0.01 level (2-tailed). Correlation is significant at the 0.05 level (2-tailed).
103 Table 2-14. Descriptive stat istics for SSRS reports Teacher Statistic Preteacher report social skills Preteacher report problem behavior Postteacher report social skills Postteacher report problem beahvior Prestudent report social skills Poststudent report social skills 1 Mean 85.00 112.06 89.94 89.06 N=16 SD 14.133 19.028 14.322 12.358 Min/Max 62/117 67/137 63/116 69/110 2 Mean 96.63 94.19 98.56 103.31 112.13 108.81 N=16 SD 14.137 12.708 11.804 13.460 14.555 14.896 Min/Max 73/122 85/125 80/114 85/125 89/130 89/130 3 Mean 91.95 93.95 121.19 86.67 103.14 98.14 N=21 SD 7.486 12.936 15.536 3.246 13.683 15.477 Min/Max 77/107 72/127 61/130 85/95 69/124 74/130 4 Mean 96.65 102.65 99.10 91.40 108.45 104.75 N=20 SD 12.975 15.308 16.370 26.269 19.223 18.671 Min/Max 66/125 85/140 60/125 40/141 69/130 67/130 5 Mean 98.09 111.09 104.00 101.45 107.55 101.82 N=11 SD 12.637 9.513 17.384 9.533 14.515 9.745 Min/Max 70/117 87/122 60/119 90/118 86/128 84/118 6 Mean 81.08 112.38 93.54 104.23 111.38 104.08 N=13 SD 6.763 7.869 16.029 16.719 18.251 18.813 Min/Max 70/91 98/127 66/130 66/127 74/130 67/130 7 Mean 91.61 105.39 111.72 110.06 N=18 SD 9.525 18.066 9.067 10.586 Min/Max 71/113 59/134 95/130 80/121 8 Mean 110.14 97.86 103.50 97.14 95.93 89.29 N=14 SD 20.111 20.583 22.083 15.402 10.887 19.024 Min/Max 74/130 85/142 62/130 85/135 78/112 48/110 Total Mean 93.75 102.91 104.35 96.13 105.06 100.98 N=129 SD 14.522 16.529 18.745 17.172 16.149 16.788 Min/Max 62/130 59/142 60/130 40/141 63/130 48/130 Higher scores indicate a greater degr ee of the characteristic measured
104 Table 2-15. Pearson correlation coe fficients for the SSRS data Scale Teacher/ PreSocial Skills Teacher/ PreProblem Behavior Teacher/ PostSocial Skills Teacher/ PostProblem Behavior Student/ PreSocial Skills Student/ PostSocial Skills Teacher/PostSocial Skills .432** .000 -.440** .000 1 -.525** .000 .012 .908 .128 .216 95 95 95 95 95 Teacher/PostProblem Behavior -.394** .000 .463** .000 -.525** .000 1 -.200 .052 -.179 .082 95 95 95 95 95 95 ** Correlation is significant at the 0.01 level (2-tailed). Correlation is significant at the 0.05 level (2-tailed).
105 Table 2-16. Behaviors and definitions Target behaviors Definition Examples Verbal aggression An utterance, scream, or other vocalization with the intent to harass, frighten, threaten or intimidate another school-aged participant. Insults, threats, teasing, yelling, and humiliation. Behaviors excluded: crying out for help or crying out in anger with no observable target. Physical aggression An attack upon another schoolaged participant with sufficient physical force to attempt to cause physical harm to another schoolaged participant. Kicking at or contact with another person, throwing an object at someone or something in anger or frustration, tripping or attempting to trip a child. Behaviors excluded: self injury or throwing an object carelessly. Relational aggression Use of social interactions in an attempt to dominate or belittle another school-aged participant. Excluding a child from a playgroup by saying, You cant play. excluding a child from a playgroup by ignoring or deliberately turning away, spreading rumors, withdrawing friendship, or threatening to withdraw friendship. Verbal prosocial A verbal action that is performed in an attempt to benefit another school-aged participant without the apparent expectation of reward for oneself. The behavior must indicate the action is meant to result in a positive social outcome. Additionally, the behavior must be socially appropriate. Saying something sympathetic, or saying something to intervene during an incident of bullying (e.g., stop that, you are being a bully). Physical prosocial An apparent physical action performed to benefit another school-aged participant without the blatant expectation of reward for oneself. The behavior must indicate the action is meant to result in a positive social outcome. Additionally, the behavior must be socially appropriate. Sharing and appropriate touching/comforting.
106 CHAPTER 3 RESULTS Hypothesis 1 Hypothesis 1 states that student s were expected to demonstrat e different benefits according to their intervention program involvement (i.e., curriculum only, curriculum plus video, and no treatment). More specifically, compared to students in the curri culum only group, students in the curriculum plus video group will demonstrate great er program benefits. Additionally, compared to students in the no treatment group, students in the curriculum only group were expected to demonstrate more benefit from the program. At the outset of the study, proposed program be nefits were to be measured by data from the PATHS Knowledge Survey, Best Acting Repo rt of Social Behaviors, Peer Nomination Survey, and the Social Skills Rating System teacher and student versions postintervention while controlling for possible differences in data on the same instruments acquire d at pre-intervention. Data were analyzed using MANCOVA. Table 3-1 outlines the data sources for the proposed MANCOVA analysis. Table 3-2 outlines descriptive statistics for each of the data sources in the final sample. Results of the Final Mulivariate Ananly sis of Covariance The final MANCOVA examined the impact of three interventions (i.e., curriculum, curriculum plus video, and no treatment), the independent variables on five depe ndent variables: (1) student self reports of prosocial behavior, (2) witnessing prosocial behaviors meas ured by the Best Acting Report of Social Behaviors, (3) student report of social sk ills, (4) teacher report of social skills, and (5) problem behaviors measured by the Social Skills Rating System. Pre-intervention data for each of the dependent variables were used as covariat es. Table 3-3 includes descriptives for the five dependent variables.
107 The MANCOVA examined whether the in tervention groups differed on a linear combination of student self reports of prosocia l behaviors and witnessi ng prosocial behaviors, SSRS student report of social skills and SSRS teacher report of social skills and problem behaviors. Pre-intervention scores served as covariates. Boxs te st of equality of covariance indicated the assumption of homogeneity of co variances was met (p = .39). Additionally, Levenes test of equality of error variances indicates the assu mption of homogeneity of variance was met for all variables (p > .05). A signifi cant difference was not found for the intervention level: Wilks = .825, F (2,92) = 1.697, p = .089, multivariate = .092. Hypothesis 2 Hypothesis 2 states that inte rvention program involvement levels (i.e., curriculum only, and curriculum plus video, and no tr eatment) were expected to predict students' responses during an actual problem solving situ ation as measured by class wide behavioral observations. More specifically, compared to students in the curriculum only group, students in the curriculum plus video group were expected to demonstrate increases in appropriate social behaviors and decreases in aggr ession. Additionally, compared to students in the no treatment group, students in the curriculum only group were expected to demonstrate increases in appropriate social behaviors and decreases in aggression. Hypothesis 2 was investigated using binomia l logistic regression. Within the eight classroom groups, observers recorded 137 incide nts prior to the intervention and 198 following the intervention where students could have responde d prosocially or aggres sively. Table 3-4 and 3-5 summarize the frequencies of aggressive and prosocial behavior by intervention group. A logistic regression analysis was performed using SPSS to assess whether the variable of treatment condition or intervention level predicte d an individuals res ponse (aggressive or prosocial) during each of the 198 social interact ions observed. When the predictor variable of
108 treatment condition was considered alone, it did not predict whether students behaved prosocially or aggressively: = .487, df = 1, N = 137, p < .05, R = .004. An additional logistic regression analysis was performed, although it was not originally included in the proposed analyses, to determine if treatment condition provided prediction of response when considered with teacher assignment and school enrollment. Therefore, a logistic regression analysis was performed to assess wh ether the three variab les of school attended, assigned teacher, and treatment condition predic ted an individuals resp onse (aggressive or prosocial) during each of the 137 social interacti ons observed. When all three predictor variables were considered together, they did not pr edict whether an indi vidual student behaved aggressively or prosocially: = 7.57, df = 3, N = 137, p < .05, R = .054. Thus, the combination of the three variable s explained 5.4% of the variance. However two of the variables are significant: teacher assignment and school a ttended. Therefore, when assessing the impact of treatment condition, teacher assignment, and school enrollment, only teacher assignment and school enrollment were significant when all three predictors are c onsidered together. Table 3-6 summarizes data on the logistic re gression coefficient, Wald test, and odds ratio for each of the predictors.
109Table 3-1. Data sources for the proposed MANCOVA Data source Pre-and postPATHS Knowledge Survey Best Acting Social Behaviors Survey Peer Nomination Social Skills Rating System-student report Social Skills Rating System-teacher report Available data from each source Number correct Number correct following item analysis Self report of -aggression -prosocial -victimization Witness peer -aggression -prosocial Z transformation for classroom and sample -prosocial -aggression -popularity -social awkwardness Social skills standard score Social skills standard score Problem behaviors standard score
110Table 3-2. Descriptives for proposed dependent variables Data Source Sample Curriculum group Curriculum plus video No intervention Pre Post Pre Post Pre Post Pre Post SD SD SD SD SD SD SD SD PATHS Knowledge Survey 15.0 1.7 16.1 2.3 15.1 1.8 16.4 2.1 14.8 1.4 15.9 2.0 15.1 1.49 14.9 3.22 Best Acting Report of Social Behaviors Self Report Aggression 8.5 2.7 8.7 3.0 8.1 2.8 8.8 3.2 9.3 2.6 8.4 2.8 8.1 2.0 9.1 2.6 Self Report Prosocial 21.1 4.6 19.4 4.4 21.6 4.8 19.5 4.3 20.1 4.6 20.1 4.1 21.0 4.1 18.1 5.2 Victimization 10.8 3.8 10.1 4.3 9.6 3.3 9.5 3.9 10.2 3.8 10.4 4.4 11.2 5.1 12.0 5.3 Witness Aggression 12.3 4.3 12.9 4.4 12.2 4.3 12.5 4.0 12.4 4.3 13.1 5.2 12.4 4.4 13.9 4.9 Witness Prosocial 14.1 4.3 13.0 4.0 14.0 4.3 13.8 3.8 14.7 4.3 12.6 3.7 13.6 4.3 10.1 4.0 Peer nomination Prosocial .16 .96 .13 1.0 .07 .91 .14 .95 .29 1.0 -.02 1.2 .29 1.1 .36 1.1 Popular .07 1.0 ..20 1.0 .11 1.0 .22 1.0 .12 1.0 .08 1.1 -.16 1.1 .32 1.0 Socially Awkward 12 96 .02 1.0 .12 1.0 .06 1.0 .07 .77 -.17 .76 .18 1.0 .20 1.2 Aggressive .06 1.0 .05 1.0 .20 1.1 .-.03 .97 .20 1.1 .16 .96 .13 1.0 .20 1.3 Social Skills Rating System Student Report Social Skills 106.4 16.1 101.1 17.0 107.5 16.2 103.5 16.8 109.6 16.4 103.0 15.1 95.9 10.1 89.1 19.0 Teacher Report Social Skills 95.6 14.9 104.4 18.7 94.9 11.7 107.1 18.2 88.9 13.0 98.3 17.1 110.1 20.1 103.5 22.1 Teacher Report Problem Behaviors 100.9 15.4 95.9 17.1 97.7 14.1 92.8 18.1 111.8 8.5 103.0 14.7 97.1 20.1 97.1 15.4 For each instrument, higher scores reflect a greater quantity of the quality measured.
111Table 3-3. Descriptives for the dependent variables Group Best Acting Report of Social Beha viors Social Skills Rating System Student self report of prosocial Witness prosocial Student self report social skills Teacher report social skills Teacher report problem behavior Mean SD Mean SD Mean SD Mean SD Mean SD Curriculum 19.47 4.27 13.60 3.85 103.46 16.81 107.09 18.24 92.76 18.51 Curiculum plus video 20.08 4.13 12.58 3.76 103.04 15.08 98.33 17.14 102.96 13.69 No intervention 17.50 4.83 10.43 3.80 89.29 19.02 103.50 22.08 97.14 15.40 For each instrument, higher scores reflect a greater quantity of the quality measured.
112Table 3-4. Frequency of aggres sion pre-and post-intervention Initiated Aggression PreResponded Aggression PreTreatment Physical Verbal Relational Physical Verbal Relational Curriculum group 23 27 5 14 25 3 Curriculum group total 55 42 Curriculum plus video group 37 15 6 16 12 3 Curriculum plus video groups total 58 31 No intervention 11 3 2 7 6 0 No intervention total 16 13 Initiated Aggression PostResponded Aggression PostTreatment Physical Verbal Relational Physical Verbal Relational Curriculum group 29 14 2 14 11 1 Curriculum groups grand total 45 26 Curriculum plus video group 12 14 0 3 11 0 Curriculum plus video groups grand total 26 14 No intervention 9 3 0 3 2 0 No intervention total 12 5
113Table 3-5. Frequency of prosocial behavior pre-and post-intervention Initiated Prosocial PreResponded Prosocial PreTreatment Physical Verbal Physical Verbal Ignore Curriculum group 8 31 13 29 10 Curriculum groups total 39 52 Curriculum plus video group 9 10 21 12 13 Curriculum plus video groups total 19 46 No intervention 2 6 4 6 1 No intervention total 8 11 Initiated Prosocial PostResponded Prosocial PostPhysical Verbal Physical Verbal Ignore Curriculum group 5 20 12 26 7 Curriculum group total 25 45 Curriculum plus video group 10 15 9 28 0 Curriculum plus video group total 25 37 No intervention 4 0 4 7 0 No intervention total 4 11
114Table 3-6. Logistic regression predicting be havioral response from teacher assignmen t, treatment level, and school attended Predictor Wald p Odds Ratio Teacher -.45 6.01 .01 .64 Treatment -.34 .90 .34 .71 School 2.40 6.78 .01 11.0
115 CHAPTER 4 DISCUSSION In the present study, the creation of new social skills, namely appropriate problem solving, is the mechanism that is theorized to reduce students aggressive responding. Social skills acquisition was meant to provide a buffer against risk factors present in the child and in the childs environment. Researchers have demons trated suitable social skills together with appropriate peer relationships ar e powerful protective factors se rving as effective buffers of aggression and violence. Resear chers have suggested that prot ective factors are effective in reducing violence because they may directly decr ease dysfunction, interact with risk factors to buffer the effects of the risk interaction, disrupt the chain that leads from risk to dysfunction, or prevent the appearance of risk factors all together. (Dah lberg & Potter, 2001; Greenberg, Domitrovick, & Bumbarger, 2001). The following paragraphs pres ent a discussion of the Best Acting Video project and the relationsh ip with student social behavior. Benefits to Students Participat ing in the Program/Hypothesis I Hypothesis I states that students will demonstrat e changes in social be haviors as a result of their exposure to the intervention. More specifically, compared to students in the curriculum only group, students in the curric ulum plus video group will demonstrate greater program benefits. Additionally, compared to student s in the no treatment group, students in the curriculum only group will demonstrate more benef it from the program. Program benefits were measured by five dependent variables: (1) stude nt self reports of pros ocial behavior and (2) witnessing prosocial behaviors measured by the Best Acting Report of So cial Behaviors, (3) student report of social skills, (4) teacher report of social skills, and (5) problem behaviors as measured by the Social Skills Rating System.
116 Contrary to the hypothesis, students in the intervention groups did not differ on the five dependent variables more than would be e xpected by chance. The intervention program seemingly did not promote significant changes in the students as measured by the outcome instruments. The finding of no effect of the intervention may have occurred due to the true absence of program benefit (i.e., d ecreases in social dysfunction or increases in socially desirable behaviors) due to measurement error regarding th e outcome instruments, or intervention fidelity. The following paragraphs expand on po ssible explanations for the findings. Absence of Program Benefits? This investigation measured intervention e ffectiveness by assessing social behaviors pre and post intervention. However, it is important to mention that students may have benefited in ways which were not measured in the present st udy. Students participating in the intervention may have experienced a sense of belonging to positive group and gained clarity regarding standards for behavior which are protective f actors reducing violence (Cunningham, 2000). Scholars have demonstrated that protective factors can disrupt the chain of events that lead from risk to dysfunction. Additionally, participating in the intervention could prevent the appearance of violence all together. However, the results indicated the absence of significant differences between intervention groups and suggest that students di d not derive benefit from part icipation in the intervention. Students may not have learned the social skills a nd or did not apply the skills at school. Factors other than the curriculum may ha ve contributed to the findings as the PATHS program has been shown to be effective in settings such as t hose in the current study (Greenberg & Kusch,1998; CPPR, 1999; Kam, Greenberg, & Ku sch, 2004; Kelly et. al, 2004). Researchers have suggested preexisting risk factors may bloc k the effectiveness of an inte rvention (Boyd, Cooley, Lambert,
117 & Ialonga, 2003; Dahlberg & Po tter, 2001; Herrenkohl, Maguin, & Hill, 2000). Possible risk factors that may have affected the inte rventions success are discussed below. Preexisting Risk Factors The intervention may not have been effective due to the presence of individual, family, and community/environmental risk factors. Although t eaching positive interacti on skills, specifically social problem solving, is the primary reco mmended strategy for reducing the potential for violence in high risk students (Hunter, Elias, & Norris, 2001 ) the buffering effect of the intervention may not have been powerful e nough to provide a counteracting mechanism to existing individual, family, and community/environmental risk factors. Individual risk factors Students who display risk fact ors for violence (e.g., significant social skills deficits, particularly deficits pair ed with an aggressive problem solving style), may have derived little benefit from a universal so cial skills building/probl em solving program. These students may require additional, targeted interventions to demonstrate social skills improvements. Scholars hypothesize that many st udents who demonstrate aggressive behaviors may share neurologically based deficits which in crease the difficulty of effective social problem solving. Deficits in aggressive children such as inhibitory co ntrol and verbal regulation may require more in depth interv ention to remediate (Dodge, 2001; Riggs, Greenberg, Kusch, & Pentz, 2006). Students who engage in aggres sion provide peers with negativ e influences. Students who identify with aggressive peers are at risk for developing or escalating th eir aggressive behavior (Dahlberg & Potter, 2001). Peer rejection, wh ich may be a consequence of an aggressive environment, has been demonstrated to predic t the growth of aggression in those who are rejected (Dodge et. al, 2003). Th erefore, aggressive students in the current investigation may
118 have influenced the classroom climate negative ly, thus promoting perceived or actual hostile classrooms. Family and environmental risk factors Students who participat ed in the present study may also have been at risk due to their exposur e to frequent domestic/drug related violence and abuse (Florida Department of Law Enforcement, 2005). Many students who participated in the present study also were living at or below the poverty level (Flo rida Department of Education, 2005). Exposure to multiple forms of violence a nd poverty are considered powerful risk factors (Dahlberg, 1998; Maxwell & Maxwell, 2003; Greenhoot, McCloskey, & Glisky; 2005, Vitale, 2001). Family violence also can interfere with a childs social development, making the engagement in prosocial behaviors with peers difficult (Vitale, 2001). These family and environmental risk factors along with individual risk factors may have worked together to minimize the effects of th e intervention. Many prev ention programs include a family and or community component to strengthen the effect of the intervention and to combat existing risk. Measurement Error and Difficulties Mea suring Social Behavior Change The outcome instruments selected to assess the impact of the intervention may not have been effective in measuring program effects/benefits Accurate measurement of social behavior change is a challenge encountered by many researchers. In fact increasing the precision with which key variables are measured has been referre d to as a core issue in the field of violence prevention (Sharkey, Furlong, & Yetter, 2006). The aim of this evaluation was to examine soci al behavior change as evidenced by increases in knowledge of the curriculum, increases in ob served prosocial problem solving, improvement in peer rated social status, incr eases in teacher and student repor ted social skills, increases in student reported prosocial skills decreases in teacher reported pr oblem behaviors, decreases in
119 student reported aggression, and decreases in ag gressive problem solving. These instruments were selected to be used in combination to pr ovide a broad picture of the participants social behaviors. However, statistical evaluation of data from these in struments required several to be deleted from the final analysis. Although thes e instruments were not included in the final analysis similar measures commonly are used in other violence preventio n program evaluation to connote program benefits and to measure changes in social behaviors. For example, direct measures of social skills frequently are utilized to evaluate violence prevention programs. These scales include the Sc hool Social Behavior S cales (e.g., Grossman et al., 1997; Taub, 2001) the Social Skills Rating Sy stem (e.g., Kratochwill, Donald, Levin, BearTibbetts, Demaray, 2004), the Social Health Prof ile (e.g., Flanagan et al., 2003), and the Teacher Child Rating Scale (e.g., Kam et al., 2004). Researchers also have documented violence prevention/social skills building program effects through the use of behavior rating scales such as the Child Behavior Checklist (e.g., Kam et al., 2004; Kratochwill, Donald, Levin, BearTibbetts, & Demaray, 2004), Problem Behavior Frequency Scales (e.g., Farrall et al., 2003), Teacher Checklist (e.g., Aber et. al., 2003; Em bry, 2002; Flanagan etl al., 2003), and The Teacher Observation of Classroom Adaptation-Revised (e.g., Embry, 2002). The current study selected the Social Skills Rating System. It was chosen to reflect students perceptions of their own social skills and teachers percep tions of student social skills and problem behaviors. Students perceptions of their social skills were most positive in the curriculum plus video group, followed by the cu rriculum group, and then the no intervention group. Although students perceptions of their social skills were more positive in the intervention groups than in the no intervention group, teachers perceptions of student social skills and problem behaviors did not follow the same pattern. In fact, data from teachers in the
120 curriculum only group indicated their students displa yed higher levels of de sirable social skills and lower levels of problem behaviors compared to the other groups. Additionally, the teacher in the no intervention group believed her students possessed higher levels of desirable social skills and lower levels of problem behaviors co mpared to the curriculum plus video groups. Findings regarding teacher report of social skills and problem behaviors in the no intervention group may have occurred due to the small sample size. However, findings regarding teacher report in the curriculum and curriculum plus vi deo group are not as easily explained. Teachers in the curriculum plus video gr oups participated in the interven tion for much of the school year, whereas teachers in the curriculum only group part icipated in the intervention for half a school year. Teacher ratings may have been affected by teacher attitude or expectation. Teachers in the curriculum plus video group may have expected th eir children to demonstrate superior social skills as they had been instructed in social skills and social problem solving for the better part of a year. Scholars have suggested that primar y grade teachers place hi gh value on self control skills and cooperation (Lane, Givener, Pierson, 2004). Teachers participating the curriculum plus video group may not have perceived adequate program benefit in re lation to the loss of instructional time. Findings regarding teacher report also may have been influenced by school climate. The curriculum group and the curriculum plus video group were housed at two separate schools. These assignments were made by school officials, thus introduc ing another possible factor in interpreting these results. Peer assessment measures such as the P eer Assessment Inventory (e.g., Embry, 2002) and peer nomination surveys (e.g., Fox & Boulton, 20 05) also are commonly used in violence prevention research. In the current study, a p eer nomination method was utilized to better
121 understand the students soci al status. Data from the peer nom ination survey were not used in the final analysis, however results indicate that st udents social status re mained relatively stable from preto postintervention. This finding is consistent with othe r literature (Dodge, 2001). Apparently, the intervention lacked sufficient power and effectiveness to change the social status of those students who were deemed aggressive an d socially awkward. More time may have been needed to change and for student s to notice change. Peer percep tions and thus peer reports are likely to remain stable over time as students get older (Nangle et al., 2002). However, peer nomination has been selected and used by researchers to illustrate changes in social behavior over time (Fox & Boulton, 2005). In this study, peer nomination did not appear to be a technique which was sensitive to changes in social behaviors, if such change occurred. Intervention Fidelity Presentation of PATHS in the curriculum a nd curriculum plus video groups are thought to have been consistent among the groups. Howeve r, the quality of the video making instruction may have detracted from the in tervention goals as suggested by the curriculum groups higher performance despite the additiona l lessons in the curri culum plus video groups. The process by which the students created the sc ript may have been improved to strengthen the effect of the intervention. As mentioned earlier, experien tial learning requires guidance. The script writing phase of the video portion of the inte rvention was structured. A t opic was provided, students were encouraged to envision a positive and negative ou tcome, and they were directed to use the problem solving steps to write di alogue. Students were also gi ven feedback and support during this process. However, students struggled to write dialogue pertinent to the problem solving steps and wrote dialogue which did not lead to their imagined positive conclusion. The current study did not employ specific techniques to e licit written responses commensurate with the
122 curriculum. An alteration of the video production activities may have pr ovided students with more guidance and substantially tied the curricul um to the video activity. The activity in the present study did generate inte rest in the students and guida nce counselors and the primary investigator provided information and feedback when the students were unable to move ahead. However, script development proved to be quite challenging for the Fourth graders. In fact, many of them may have found the task too difficu lt and became frustrated and disinterested. When students become frustrated and disinter ested, experiential learning does not readily occur (Wurdinger, 2005). A possible solution to combat frustration and disinterest is to implement writing techniques which support the student s further. The goal of a more structured writing activity is to engage stude nts through the use of adult dire cted scaffolding. The goal in scaffolding is learning for understanding, rath er than learning to produce a product. Additionally, the goal is for lear ners to achieve independently what was once only possible with assistance (Warwick & Malock, 2003). Using the problem solving steps as skeleton, students could be provided a time line to inform them how long they have to discuss each step within their group. Adult facilitators would serve as time keepers and redir ect students as needed. Followi ng discussion, students could then be provided with a visual repres entation of the problem solving st ep for each session along with a map for where the chosen problem solving step falls on the continuum. They might also have access to a computer to record their ideas to re duce anxiety about the writing task. The adult facilitators would then introdu ce standard questions from the curriculum as well as other situation specific que stions to elicit dialogue pertinent to each problem solving step. In effect, students would be provided with fewer intimidati ng writing tasks, and th e adults would provide
123 guided structure to elicit responses (Murphy & Dudley-Mar ling, 2003; Soderman, Gregory, & ONeill, 1999). After students are successful in responding to the prompts for each problem solving step, they could then be provided with a print out of their dialogue corresponding to each step. The print out could include a map for each step, follo wed by the dialogue the students provided. The next session would include transf erring the ideas into a script fo rmat. Students could be given a formatted worksheet for this activity to increase structure. Student s would then be encouraged to act out the parts for portions of the script for the class and get feedback regarding the problem solving steps and the relevance of their dialogue. This process could continue until the steps are covered. The goal of this process would be fo r each group to become more independent in determining the appropriateness of their dialogue. Scaffolding th e script production in this way could promote learning how to apply the prob lem solving steps in social situations. Benefits to Students Participat ing in the Program/Hypothesis 2 Hypothesis 2 stated that intervention program involvement level (i.e., curriculum only, and curriculum plus video, and no trea tment) was expected to predic t students' responses during an actual problem solving situation as measured by class wide behavioral observations. More specifically, compared to students in the curric ulum only group, students in the curriculum plus video group were expected to demonstrate incr eases in appropriate social behaviors and decreases in aggression. Additionally, compared to students in the no treatment group, students in the curriculum only group were expected to demonstrate incr eases in appropriate social behaviors and decreases in aggr ession. Researchers have employed the use of behavioral observations to measure the impact of violence prevention interventions. Typically, students are chosen for observation randomly or are each observe d for a 15 to 90 second interval (e.g., Frey et al., 2000 and Taub, 2001). Student observations are less comm only utilized in violence
124 prevention research presumably due to their high cost. In the current study, observations were utilized to provide an objective measure of behaviors in each group. The intervention seemingly had no impact on students actual problem solving ability during observations. It was not possible to predict whether a student would respond prosocially or aggressively based on their involvement in th e intervention. Results regarding Hypothesis 2 may have occurred due to the students preexisting risk factors discussed previously in this section. That is, the students behaviors as measured by the observations may have been influenced more by environmental factors than by participation in the violence prevention curriculum. Teacher assignment and school attend ed did have an impact on appropriate problem solving. This investigation did not include a m easure of school or classroom climate; however, teacher assignment and school enrollment may c onstitute two components of school climate. In this study, school climate may have influenced ho w the students responded in a problem solving situation. Positive school climate has been cited as a protective factor of violence and can reduce aggressive behaviors in school s (Furlong et al., 2000; 2003; Mo rrison et al., 1994; Sugai, 2007). Additionally, the result s regarding hypotheses 2 may have occurred due to methodological deficits in the observation process. The beha vioral observations may not have captured the quality and or quantity of student interactions. Potential Contributions and Areas of Development. When faced with a choice of programming, sc hool district personnel want to implement effective programming. This study investigated the effectiveness of a universal, classroom based curriculum with an expe riential add-on component (i.e., vide o production). The findings do not support the effectivenes s of the intervention. In addi tion, evidence suggests that video production may not provide a level of be nefit commensurate with the cost.
125 This investigation also supported other res earch which finds little correlation between teacher and student report of social skills. Some researchers are currently investigating social responding of students with social skills problems and whether they are competent to reliably report their own social skills deficits and stre ngths. According to Crick and Dodge (1994) aggressive students possess social processing deficits, which cause them to interpret interpersonal information differently than their p eers. These deficits ar e proposed to interfere with the perceiving, decoding, and interpreting of social contexts. They propose that in comparison to teacher and peer ratings, aggres sive students overesti mate their cognitive competence and peer relationships, even though they are judged as being le ss preferred than their peers. Also, they propose that non-aggressive students actually tend to underestimate their cognitive and social abilities in comparison to teacher reports and peer nominations (Crick & Dodge, 1994). Some researchers are calling fo r measurement of cognitive processes which underlie desirable social behavior This was out of the scope of the current investigation; however may account for the differences in teacher and student perceptions. Limitations The data collected in the course of this study did not reveal di fferences in social behaviors between the students in the curriculum plus video, curriculum only, and no intervention groups. These findings normally may not be published, theref ore we may be unaware of the frequency of similar findings. Many scholars discuss the diffi culty of conducting rigo rous evaluations of prevention programs in school settings (Alkon et al. 2001; Casella, 2002; Cunningham, 2000; Gorman, 2002; Greenberg, 2004; Guerra, 2003; Hunt er et al., 2001; Miller-Johnson, Sullivan, & Simon, 2004; Multisite Violence Prevention Project, 2004; Spoth, Greenberg, Bierman, Redmond; 2004; Farrell et al., 2001; Vernberg & Gamm 2003). In this investigation, difficulties encountered in data collection may have impacted the final results (i.e., difficulty acquiring other
126 no treatment groups and difficulty collecting teacher data ). Increased sample size may have reduced error and provided a more precision to determine the effect of the intervention. Although the experimental design includes preand post-measures to account for preexisting differences in experi mental groups, lack of random a ssignment to intervention groups may affect the findings. Random assignment was not possible due to administrator directives. The study also may have been limited by the absen ce of data on participat ing students risk and protective factors of violence. They have been found to have a profound affect on individuals. The collection of data on school climate, academic levels, undesirable parental characteristics and skills, and perception of few pr ogram benefits may have been he lpful to determine if the risk factors corresponded with deficits in social behaviors and resistan ce to the intervention (Embry, 1997; Kerns & Prinz, 2002). Further limitations may have arisen from personnel issues. Two trained guidance counselors, assigned to each school, delivered the curriculum along with the principal investigator. Guidance counselor assignment ma y have affected the findings. Additionally, teachers were not required to use the violence pr evention curriculum content in their classrooms. Some teachers may have discussed curriculum c ontent (e.g. using the problem solving steps to solve a student problem) outside the designate d lesson time, thus affecting the findings. Teachers also were not blind to theFinally, teachers and administra tors elected to participate in the study. Thus, level of support may vary among these professionals and thus may affect the findings. Data were not collected on st udents special education or general education membership, retention status or school discipline records (e.g., Rollin, Kaiser-Ulrey, Potts, & Creason, 2003,
127 Sugai, Sprague, Horner, & Walker, 2000). Data collection in these ar eas may have provided valuable information regarding group differences.
128 APPENDIX A PATHS KNOWLEDGE SURVEY Name:___________________ __________Age:________T eacher:____ ________ 1. What is the first step in problem solving? a. call your mom b. stop and calm down c. cry d. walk away 2. What does the word identify mean? a. using a magnifying glass b. doing something c. to know what something is d. to fix something 3. What questions do you need to ask yourself to identify a problem? a. What is the problem? b. Is there a problem? c. Who owns the problem? d. You should ask all of these questions 4. What is the third st ep of problem solving? a. identify feelings b. go home c. stop and calm down d. identify the problem 5. What can you look at to identify what another person is feeling? a. How their hair and clothes look b. How their faces and bodies look c. What they are saying and doing d. Choices b and c 6. What is a goal? a. Something we want to happen b. A term used in baseball c. Something we hope will happen d. Choices a and c 7. Goals can be categorized as a. Goals that will happen soon b. Goals that will happen in a long time
129 c. Goals that will never happen d. Choices a and b 8. What does the word choice mean? a. to make a decision b. to choose something c. to pick one thing from many d. all of these 9. What is the fifth step in problem solving? a. give up b. stop trying c. think of solutions d. think of what you want to do after school 10. What is a good way to come up w ith lots of ways to solve problems? a. talk to friends and ask their advice b. come up with ideas of your own c. ask adults for ideas d. all of these 11. How many ideas should y ou have before you decide how to solve a problem? a. One idea b. No ideas c. Many ideas 12. What does evaluate mean? a. to find out the price of something b. to pick between go od and bad solutions c. to lift something up 13. What should you do when you are problem-solving? a. Freeze b. Leave town c. Ignore the problem d. Think about the consequences What step of the problem solving step s should be used in these situations? 14. John tried to figure out why Sarah was crying a. stop and calm down b. identify the problem c. identify the feelings d. decide on a goal
130 15. John talks to his friends to come up with ideas on how to cheer Sarah up a. identify the feelings b. decide on a goal c. think of solutions d. choose the best solution 16. What kind/s of questions should you ask yourself wh en you come up with a good plan? a. What time is it? b. Who? What? When? Where? and How? c. Why should I make a plan? d. How long is this going to take? 17. What is an obstacle? a. a type of exercise to get strong b. a rock c. a tree d. things that get in the way 18. What should you do if a plan doesnt work the first time? a. dont try it ever again b. give up c. try not to have any more problems d. try a different plan 19. Why is joining a new group a problem? a. because you dont know them b. because you dont feel comfortable c. because you may be shy d. all of these are problems 20. What are some reasons good plans dont work? a. not enough planning b. you didnt pick a positive goal c. you picked a wrong ti me to try your plan d all of these are reasons why plans dont work 21. What can we do if a plan does fail? a. come up with a new plan b. dont give up c. change your goals d. do all of these
131 APPENDIX B BEST ACTING REPORT OF SOCIAL BEHAVIORS Name:__________________ Age:__________ Circle: If you are a Boy or a Girl Circle if you have done these things at school during th e past two weeks 1. Call a student bad names Many Times Often Sometimes Never 2. Push or kick a student Many Times Often Sometimes Never 3. Bully (gang up on) a student Many Times Often Sometimes Never 4. Dont let a student play with you Many Times Often Sometimes Never 5. Say rude things about another student Many Times Often Sometimes Never 6. Hit a student Many Times Often Sometimes Never Circle if you have done these thin gs at school during th e past two weeks 7. Help another students Many Times Often Sometimes Never 8. Stick up for a friend Many Times Often Sometimes Never 9. Say something nice to a classmate Many Times Often Sometimes Never 10. Let someone join in your group Many Times Often Sometimes Never 11. Say you are sorry Many Times Often Sometimes Never 12. Solve a problem in a positive way Many Times Often Sometimes Never 13. Help a friend solve a problem Many Times Often Sometimes Never Circle if anyone in your class has done these things at school during the past two weeks 14. Calls you a bad name Many Times Often Sometimes Never 15. Pushes or kicks you Many Times Often Sometimes Never 16. Bullies or gangs up against you Many Times Often Sometimes Never 17. Refuses to let you play with them Many Times Often Sometimes Never 18. Says rude things about you Many Times Often Sometimes Never 19. Hits you Many Times Often Sometimes Never Circle if you have seen these things happen to your classmates at school during the past two weeks
132 20. Being called a bad name Many Times Often Sometimes Never 21. Being pushed or kicked Many Times Often Sometimes Never 22. Being bullied or ganged up against Many Times Often Sometimes Never 23. Not let another person play with them Many Times Often Sometimes Never 24. Say rude things to another person Many Times Often Sometimes Never 25. Being hit Many Times Often Sometimes Never 26. A student helping another student Many Times Often Sometimes Never 27. A student stick up for a friend Many Times Often Sometimes Never 28. Let a student join in your group Many Times Often Sometimes Never 29. A student saying they are sorry Many Times Often Sometimes Never 30. A student solving a problem in a positive way Many Times Often Sometimes Never
133 APPENDIX C PEER NOMINATION QUESTIONAIRE Name three children wh o you like the most. 1. 2. 3. Name three children who fight the most. 1. 2. 3. Name three children who start fights the most. 1. 2. 3. Name three children who dont know how to join in a group. 1. 2. 3. Name three children who sa y mean things a lot. 1. 2. 3. Name three children who sh are their things a lot. 1. 2. 3. Name three children who help others a lot. 1. 2. 3.
134 APPENDIX D BEST ACTING BEHAVIORAL OBSERVATION CHECKLIST Time in Time out Classroom Event # Setting Classroom Transition Lunch Other Initiator M F W B H Other Respondent M F W B H Other Aggressive Physical Aggressive Verbal Aggressive Relational Prosocial Physical Prosocial Verbal Ignore Initiator Behavior Respondent Behavior Description of initiation Description of response Time in Time out Classroom Event # Setting Classroom Transition Lunch Other Initiator M F W B H Other Respondent M F W B H Other Aggressive Physical Aggressive Verbal Aggressive Relational Prosocial Physical Prosocial Verbal Ignore Initiator Behavior Respondent Behavior Description of initiation Description of response Time in Time out Classroom Event # Setting Classroom Transition Lunch Other Initiator M F W B H Other Respondent M F W B H Other Aggressive Physical Aggressive Verbal Aggressive Relational Prosocial Physical Prosocial Verbal Ignore Initiator Behavior Respondent Behavior Description of initiation Description of response
135 LIST OF REFERENCES Aber, L. J., Brown, J. L., & Jones, S. (2003). De velopmental trajectories toward violence in middle childhood: Course, demographic diffe rences, and response to school-based intervention. Developmental Psychology, 39 324-348. Acosta, O. M., Albus, K., Reynolds, M. W., Spri ggs, D., & Weist, M. D. (2001). Assessing the status of research on viol ence related problems among youth. Journal of Clinical Child Psychology, 30 152-160. Adelman, H. S., & Taylor, L. (2003). Rethi nking school psychology (commentary on public health framework series). Journal of School Psychology, 41 83-90. Alkon, A., Tschann, J. M., Ruane, S. H., Wolff, M ., & Hittner, A. (2001). A violence prevention and evaluation project with et hnically diverse populations. American Journal of Preventive Medicine, 20 48-55. Bandura, A. (1963). Social learning and personality development. New York: Holt Rinehart and Winston. Barrios, L. C. (2001). Pr eventing school violence. The Western Journal of Medicine, 174 88-91. Batsche, G. M. (1999). Bullying. In A. S. Canter & C. A. Servio (Eds.), Crisis Prevention and Response A Collection of NASP Resources (pp. 23-33). Bethesda, MD: The National Association of School Psychologists. Bellanti, C. J., & Bierman, K. L. (2000). Disent angling the impact of low cognitive ability and inattention on social behavi or and peer relationships. Journal of Clinical Child Psychology, 29, 66-75 Benes, K., Furlong, M., & Mitchell, K. (2002). Mental Measurement Yearbook (12). Retrieved July 5, 2006, from http://gateway.ut.ovi d.com.lp.hscl.ufl.edu/gw2/ ovidweb.cgi. Bosworth, K., Espelange, D., & DuBay, T. ( 1998). A computer-based violence prevention intervention for young adolescents: Pilot study. Adolescence, 33 785-795. Boyd, R. C., Cooley, M. R., Lambert, S., & Ialo ngo, N. (2003). First grad e child risk behaviors for community violence e xposure in middle school. Journal of Community Psychology, 31 297-314. Bronfenbrenner, U. (1979). The Ecology of Human Development: Experiments by nature and design Cambridge, MA: Harvard University Press. Bronfenbrenner, U. (1988). Interacting systems in human development. Research paradigms: Present and future. In N. Bolger, A. Caspi, G. Downey, & M. Moorehouse (Eds.), Persons in context: Developmental processes (pp. 25-49). New York, NY: Cambridge University Press.
136 Carmona, R. H., & Hoffman, J. S. (2005). Hea lth Professional Traini ng in Youth Violence Prevention: A commentary by the surgeon general. Journal of Preventative Medicine, 29 137-174. Casella, R. (2002). Where policy meets the pavement: Stages of public involvement in the prevention of school violence. Qualitative Studies in Education, 15 349-372. Centers For Disease Control. (2005). Behavior and social environment. In America's children in brief: Key national indica tors of well-being, 2006. Retrieved from http://www.childstats.gov/americaschildren/beh.asp on April 26, 2007. Chapin, J., & Coleman, G. (2003). Unrealis tic optimism and school violence prevention programs. North American Journal of Psychology, 5 (2), 193-203. Chesney-Lind, M., & Belknap, J. (2004). Trends in deliquent girls' aggression and violent behavior: A review of the evidence. In M. Putallaz & K. L. Bierman (Eds.), Aggression, antisocial behavior, and violence among girls: A developmental perspective. (pp. 203220). New York: Guilford Publications. Cirillo, K. J., Pruitt, B. E., Colwell, B. ( 1998). School violence: Prevalence and intervention strategies for at risk adolescents. Adolescence, 33 317-328. Clayton, C. J., Ballif-Spanvill, B., & Hunsaker M. (2001). Preventing violence and teaching peace: A review of promising and effective antiviolence, conflict resolution, and peace programs for elementary school children. Applied & Preventive Psychology, 10 1-35. Coie, J. D., Dodge, K. A. (1983). Continuities an d changes in children's social status: A fiveyear longitudinal study. Merrill-Palmer Quarterly, 29, 261-282. Conduct Problems Prevention Research Group (1999). In itial impact of the fast track prevention trial for conduct problems: II. Classroom effects. Journal of Consulting and Clinical Psychology, 67 648-657. Crick, N. R.;& Dodge, K. A. (1994). Review and re formulation of social information-processing mechanisms in children's social adjustment. Psychological Bulletin, 11 74-101. Cunningham, N. J. (2000). A comprehensive approach to school-community violence prevention. Professional School Counseling, 4 126-133. Dahlberg, L. L. (1998). Youth vi olence in the United States: Majo r trends, risk factors, and prevention approaches. American Journal of Preventive Medicine, 14 259-272. Dahlberg, L. L., & Potter, L. B. (2001). Y outh violence: Developmental pathways and prevention challenges. American Journal of Preventive Medicine, 20 3-14.
137 Daiute, C., Stern, R., & Lelutiu-Weinberger C. (2003). Negotiating violence prevention. Journal of Social Issues, 59 83-101. Daley, C. E. & Onwuegbuzie, A. J. (2004). Attri butions toward violence of male juvenile delinquents: A concurrent mixed-methodological analysis. Journal of Social Psychology, 144 549-570. Dawson, G. (1994). Frontal electoenchalographic corre lates of individual differences in emotion expression in infants: A brain systems perspective on emotion. In N.Fox (Ed.), Monographs of the society for research in child development: Vol. 59, The development of emotional regulation: Biologi cal and behavioral considerations (pp. 135-151). Boston: Blackwell. Derzon, J. H., Wilson, S. J., & Cunningham, M. A. (1999). The effectiveness of school-based interventions for preventing and reducing violence. Center for Evaluation Research and Methodology: Vanderbilt Institute for Public Policy Studies. Diperna, J.C. & Volpe, R.J. (2005). Self-report on the social skills rati ng system: Analysis of reliability and validity for an elementary sample. Psychology in the Schools, 42, 345-356. Dodge, K. A. (2001). The science of youth violen ce prevention: Progressi ng from developmental epidemiology to efficacy to effectiveness to public policy. American Journal of Preventive Medicine, 20 63-70. Dodge, K. A, Lansford, J. E., Burks, V. S., Bates, J. E., Petit, G. S., Fontaine, R., & Price, J. M. (2003). Peer rejection and social informati on processing factors in the development of aggressive behavior problems in children. Child Development, 74 374-393. Dodge, K. A. & Rabiner, D. L. (2004). Returning to roots: On social informationprocessing and moral development. Child Development, 75 1003-1008. Domitrovich, C. E. & Greenberg, M. T. (2000). The study of im plementation: Current findings from effective programs that prevent ment al disorders in sc hool-aged children. Journal of Educational Psychology Consultation, 11 193. Duffey, J.B., Salvia, J., Tucker, J., & Ysseldyke, J. E. (1981). Nonbiased assessment: A need for operationalism. Exceptional Children 47 427-434. DuRant, R. H., Barkin S., & Krowchuk, D. P. (2001). Evaluation of a peaceful conflict resolution and violence prevention curriculum for sixth-grade students. Journal of Adolescent Health 28, 386-393. Dwyer, K. P. (1999). Children killing children. In A. S. Canter & C. A. Servio (Eds.), Crisis Prevention and Response: A Collection of NASP Resources (pp. 1-6). Bethesda, MD: The National Association of School Psychologists.
138 Eddy, M. J., Reid, J. B., & Fetrow, R. A. (2000). An elementary school-based prevention program targeting modifiable anticedents of youth delinquency and violence: Linking the interests of families and teachers (LIFT). Journal of Emotional & Behavioral Disorders, 8 165-176. Ellickson, P. L., & McGuigan, K. A. (2000). Early predictors of adolescent violence. American Journal of Public Health, 90 566-572. Embry, D. D. (1997). Does your school have a peaceful environment? Intervention in School & Clinic, 32 217-223. Embry, D. D. (2002). The good behavior game: A best practice candidate as a universal behavioral vaccine. Clinical Child and Family Psychology Review, 5 273-297. Fagan, A. A., & Mihalic, S. (2003). Strategies for enhancing the a doption of school-based prevention programs: Lessons learned from the Blueprints for violence prevention replications of the Life Skills Training Program. Journal of Community Psychology, 31 235-253. Farrell, A. D., Meyer, A. L., Sullivan, T. N., & Kung, E. M. (2003). Evaluation of the Responding in Peaceful and Positive Ways (R IPP) seventh grade violence prevention curriculum. Journal of Child and Family Studies, 12 101-120. Farrell, A. D., Meyer, A. L., & White, K. S. (2001). Evaluation of Res ponding in Peaceful and Positive Ways (RIPP): A school-based prev ention program for reducing violence among urban adolescents. Journal of Clinical Child Psychology, 30 451-463. Flanagan, K. S., Bierman, K. L., & Kam, C. M. (2003). Identifying at-r isk children at school entry: The usefulness of multibehavioral problem profiles. Journal of Clinical Child and Adolescent Psychology, 32 396-407. Flannery, D. J. (2003). Improving school violence prevention programs through meaningful evaluation ERIC Clearinghouse on Urban Education. NY. Retrieved from http://ericweb.tc.columbia.edu/dige st/dig132.asp on May 22, 2003. Florida Department of Education (2005). FCAT scores Tallahassee, FL. Retrieved from http://www.fldoe.org/default.asp?bhcp=1 on July 15, 2005. Florida Department of Education (2007). Free and reduced lunch statistics Tallahassee, FL. Retrieved from http://www.fldoe.org/FN M/natlschoollunch/ on February 5, 2007. Florida Department of Law Enforcement (2005). Florida crime rates Tallahassee, FL. Retrieved from http://www.fdle.state.fl.us/fsac/Crime_ Trends/total_Index/i ndex.asp on July 15, 2005.
139 Florida State Department of Health (2003). 2003 Florida Youth Risk Behavior Survey Tallahassee, FL. Retrieved from www.doh.state.fl.us on May 1, 2005. Fields, S., & McNamara, J. (2003). The prev ention of child and a dolescent violence. Aggression and Violent Behavior, 8 61-91. Forness, S. (2003). Barriers to evidence-base d treatment: Developmental psychopathology and the interdisciplinary di sconnect in school mental health practice. Journal of School Psychology, 41 61-67. Fox, C. L., & Boulton, M. J. (2005). The social skills problems of bullying: Self, peer and teacher perceptions. British Journal of Educational Psychology, 75 313-328. Frederickson, N. L., & Furnham, A. F. (2004). Peer-assessed behavioral characteristics and sociometric rejection: Differences betw een pupils who have moderate learning difficulties and their mainstream peers. British Journal of Educ ational Psychology, 74, 391-410. Frey, K. S., Hirschstein, M. K., & Guzzo, B. A. (2000). Second Step: Preventing aggression by promoting social competence. Journal of Emotional & Behavioral Disorders, 8 102-112. Furlong, M., & Morrison, G. (2000). The school in school violence: Definitions and facts. Journal of Emotional & Behavioral Disorders, 8 71-80. Furlong, M., Morrison, G., Chung, A., Bates, M., & Morrison, R. L. (1999). School violence. In A. S. Canter & C. A. Servio (Eds.), Crisis Prevention and Response: A Collection of NASP Resources (pp. 7-21). Bethesda, MD: The National Association of School Psychologists. Furlong, M., Morrison, G., & Pavelski, R. (2000) Trends in school psychology for the 21st Century: Influences of school violence on professional change. Psychology in the Schools, 37 81-90. Furlong, M., Paige, L., & Osher, D. (2003). The sa fe schools/healthy studen ts (SS/HS) initiative: Lessons learned from implementing comprehensive youth development programs. Psychology in the Schools, 40 447-456. Goldstein, A., & Conoley, J. (2004). Student Aggr ession: Current Status. In J. Conoley & A. Goldstein (Eds.), School violence interven tion: A practical handbook (2nd ed., pp. 3-20). New York, NY: Guilford Press. Gorman, D. M. (2002). Defining and operationaliz ing 'research-based' prevention: A critique (with case studies) of the US Department of Education's safe, disciplined and drug-free schools exemplary programs. Education and Program Planning, 25 295-302.
140 Greenberg, M. T. (2004). Current and future challenges in school-based prevention: The researchers perspective. Prevention Science, 5 5-13. Greenberg, M. T., Domitrovick, C ., & Bumbarger, B. (2001). The prevention of mental disorders in school-aged children: Curre nt state of the field. Prevention & Treatment, 4 1-62. Greenberg, M. T. & Kusch, C. A. (1998). Preventive interventions for school-age deaf children: The PATHS curriculum. Journal of Deaf Studies and Deaf Education, 3 49-63. Greenberg, M. T., & Kusch, C. A. (2006). Bu ilding social and emotional competence: The PATHS curriculum. In: S. R. Jimerson & M. Furlong (Eds.z0 Handbook of school violence and school safety: From research to practice. Lawrence Erlbaum Associates Publishers. Mahwah, NJ, pp. 395-412. Greenberg, M. T., Kusch, C. A.; Cook, E. T ., Quamma, J. P. (1995). Promoting emotional competence in school-aged children: Th e effects of the PATHS curriculum. Development and Psychopathology, 7, 117-136. Greenberg, M. T., Kusch, C. A., & Mihalic, S. F. (1998). Blueprints for Violence Prevention, Book 10: Promoting Alternative Thinking Strategies Retrieved from http://www.nimh.nih.gov/publ icat/nimhblueprint.pdf Greenhoot, A. F., McCloskey, L., & Glisky, E. (2005). A longitudinal study of adolescents' recollections of family violence. Applied Cognitive Psychology, 19 719-743. Gresham, F. M., & Elliot, S. N. (1990). Social Skills Rating System American Guidance Service Inc. Circle Pines: MN. Grossman, D. C., Neckerman, H. J., Koepsell, T. D., Liu, P. Y., Asher, K. N., Beland, K., Frey, K., & Rivara, F. P. (1997). Effectivene ss of a violence prevention curriculum among children in elementary school. A randomized controlled trial. The Journal of the American Medical Association, 277 983-988. Guerra, N. G. (2003). Preventing school violence by promoting wellness. Journal of Applied Psychoanalytic Studies, 5 (2), 139-154. Guerra, N. G., & Williams, K. R. (2006). Et hnicity, Youth Violence and the Ecology of Development. In Nancy Guerra & Emilie Smith (Eds.), Preventing youth violence in a multicultural society (pp. 17-45). Washington, DC: American Psychological Association. Hall, L. J., & Strickett, T. ( 2002). Peer relations of preadolescen t students with disabilities who attend a separate school. Education and Training in Mental Retardation and Developmental Disabilities, 37, 399-409.
141 The Hamilton Fish Institute (2003) Effective violence prevention programs Washington, D.C. Retrieved from www.hamfish.org/pub/evpp.html on May 29, 2003. Heider, F. (1959). On perception, event structure, and ps ychological environment, selected papers. New York: International Universities Press. Hermann, M. A. & Finn, A. (2002). An ethical an d legal perspective on the role of school counselors in preventing violence in schools. Professional School Counseling, 6 46-54. Herrenkohl, T., Maguin, E., & Hill, G. (2000). Deve lopmental risk factors for youth violence. Journal of Adolescent Health, 26 176-186. Hintze, J. M. (2002). Interventions for Fears an d Anxiety Problems. In M. R. Shinn & H. M. Walker & G. Stoner (Eds.), Interventions for academic and behavior problems II: Preventive and remedial approaches Washington DC: Nationa l Association of School Psychologists. Hintze, J. M.; Shapiro, E. S. (1999). School Vi olence. In W.K. Silverman, T. H. Ollendick (Eds.), Developmental issues in the clinical treatment of children (pp. 156-170). Needham Heights, MA: Allyn & Bacon. Hudley, C., Britsch, B., Wakefield, W., Smit h, T., DeMorat, M., & Cho, S. (1998). An attribution retraining program to reduce a ggression in elementary school students. Psychology in the Schools, 35 271-282. Hunter, L., Elias, M. J., & Norris, J. (2001). School-based violence prevention: Challenges and lessons learned from an action research project. Journal of School Psychology, 39 (2), 161-175. Ikeda, R. M., Simon, T. R., & Swahn, M. (2001). The prevention of youth violence: The rationale for and characteristic of four evaluation projects. American Journal of Preventive Medicine, 20 15-21. Johannes, E. M. (2004). Effects of PATHS after school program on children's social environment and behavior. Dissertation Abstracts International Section A:Humanities and Social Sciences, 64 4643. Johnson, J. C., Ironsmith, M., Poteat, G. M., (1 994). Assessing children' s sociometric status: Issues and the application of social network analysis. Journal of Group Psychotherapy, Psychodrama & Sociometry, 47, 36-48. Jollivette, K., Stitchner, J. P., Sibilsky, S., Scott, T. M., & Ridgely, R. (2002). Naturally occurring opportunities for pr eschool children with or w ithout disabilities to make choices. Education and Treatment of Children, 25 396-414.
142 Kam, C., Greenberg, M. T., & Kusch, C. A. ( 2004). Sustained effects of the PATHS curriculum on the social and psychological adjustment of children in special education. Journal of Emotional and Behavioral Disorders, 12 66-78. Kam, C., Greenberg, M.T., & Walls, C.T. (2003). Examining the role of implementation quality in school-based prevention using the PATHS curriculum. Prevention Science 4 55-63. Kamphaus, R. W., & Frick, P. J. (2002). Pee r-Referenced Assessment. In R. Pascal (Ed.), Clinical Assessment of Child and Adolescent Personality and Behavior (pp. 218-229). Boston: Allyn & Bacon. Kelly, B., Longbottom, J., Potts, F., & Williamson, J. (2004). Applying emotional Intelligence: Exploring the Promoting Alternativ e Thinking Strategies curriculum. Educational Psychology in Practice, 20 221-240. Kerns, S. E. U., & Prinz, R. J. (2002). Critical issues in the preven tion of violence-related behavior in youth. Clinical Child & Family Psychology Review, 5 133-160. Knoff, H. M., & Batsche, G. M. (1995). Projec t ACHIEVE: Analyzing a school reform process for at-risk and underachieving students. School Psychology Review, 24 579-603. Kratochwill, T. R., Donald, L., Levin, J. L., Young Bear-Tibbetts, H., & Demaray, M. K., (2004). Familes and schools together: An exper imental analysis of parent-mediated multi-family group program for American Indian children. Journal of School Psychology, 42 359-383. Kuperminc, G. P., & Brookmeyer, K. ( 2006). Developmental psychopathology. In R. Ammerman (Ed.) Comprehensive handbook of personal ity and psychopathology, Vol. 3. (pp. 100-113). Hoboken, NJ: John Wiley & Sons, Inc, Kusch, C. A. (2002). Psychoanalysis as prevention: Using PATHS to enhance ego ddevelopment, object relationships, & cortical integration in children. Journal of Applied Psychoanalytic Studies, 4, 283-301. Kusch, C. A.,& Greenberg, M. T.(1994). Promoting Alternative Thin king Strategies (PATHS). Channing Bete Company: South Deerfield, MA. Leech, N. L., Barrett, K. C., Morgan, G. A. (2005). SPSS for Intermediate Statistics: Use and Interpretation, Lawrence Erlbaum Associates: Mahwah, NJ Leff, S. S., Power, T. J., Manz, P. H., Costig an, T. E., & Nabors, L. A. (2001). School-based aggression prevention programs for young childre n: Current status and implications for violence prevention. School Psychology Review, 30 344-362.
143 Lisak, D.,& Miller, P. (2003). Childhood trauma, posttraumatic stress disord er, substance abuse, and violence. In Paige Ouimette & Pamela Brown (Eds.), Trauma and substance abuse: Causes, consequences, and trea tment of comorbid disorders. (pp. 73-88).Washington, DC: American Psychological Association. Lochman, L. E. (1992). Cognitive-behavioral in tervention with aggressive boys: Three-year follow-up and preventive effects. Journal of Consulting an d Clinical Psychology, 60 426-432. Luria, A. R. (1976). Cognitive development: Its cu ltural and social foundations. Cambridge, MA: Harvard University Press. Lynch, M. (2006). Children exposed to community violence. In Margaret M. Feerick & Gerald Silverman (Eds.), Children exposed to violence (pp. 29-52). Baltimore, MD: Paul H Brookes Publishing. Mattila, V. M., Parkkari, J. P., & Rimpel, A. H. (20060. Risk factors for violence and violencerelated injuries among 14to 18-year-old Finns. Journal of Adolescent Health, 38 617620. Maxwell, C. D., & Maxwell, S. R. (2003). Experiencing and witnessing familial aggression and their relationship to physically aggre ssive behaviors among Filipino adolescents. Journal of Interpersonal Violence, 18, 1432-1451. McCann, J. T. (2002). Threats in schools: A practic al guide for managing violence New York: US Haworth Press, Inc. Mercy, J., Krug, E., & Dahlberg, L. (2003). Violence and Health: Th e United States in a Global Perspective. American Journal of Public Health, 93 256-261. Merrel, K. (2002). Social-emotional interventi on in schools: current status, progress, and promise. School Pschology Review, 31 143-147. Mihalic, S., Irwin, K., Elliot, D., Fagan, A. A., & Hansen, D. (2001). Blueprints for violence prevention Office of Juvenile Justic e and Delinquency Prevention. Miller A. L., Gouley K., Seifer, R., Zakriski, E. & Ve rgna, A. (2005). Emotion knowledge skills in low-income elementary school children : Associations with so cial status and peer experiences. Social Development, 14, 637-651. Miller-Johnson, S., Sullivan, T. N., & Simon, T. R. (2004). Evaluating the impact of interventions in the multis ite violence prevention study. American Journal of Preventative Medicine, 26 48-61. Molina, I. A., Dulmus, C. N., & Sowers, K. M. (2005). Secondary prevention for youth violence: A review of selected school-based programs. Brief Treatment and Crisis Intervention, 5 1-13.
144 Morrison, G. M., Furlong, M. J., & Smith, G. (19 94). Factors associated with the experience of school violence among general education, leadersh ip class, opportunity class, and special day class pupils. Education & Treatment of Children, 17 356-369. Moon, J. A. (2004). Experiential and reflective learning. Handbook of reflect ive and experiential learning: Theory and practice (pp. 69-129). London: Routeldge-Fulmer. Mu, K., Siegel, E. B., & Allinder, R. M. (2000). P eer interactions and sociometric status of high school students with moderate or severe di sabilities in general education classrooms. Journal of Association for Persons with Severe Handicaps, 25, 142-152. Multisite Violence Prevention Project (2004). Lessons learned in the multisite violence prevention project collaboration. American Journal of Preventative Medicine, 26 62-71. Murphy, S., & Dudley-Marling, C. (2003). Writing about writing. Literacy through language art s \pp. 221-344). Urbana, IL: National Council of Teachers of English. Nangle, D., Erdley, C., & Carpenter, E. (2002) Social skills traini ng as a treatment for aggressive children and adolescents: A developmental-clinical integration. Aggression and Violent Behavior, 7 169-199. National Rural Behavioral Health Center. (2003). Risk Incidents for Schools Inventory Gainesville, Florida: University of Florida. Okwumabua, J. O., Wong, S. P., Duryea, E., M ., O. T., & Howell, S. (1999). Building selfesteem through social skills training a nd cultural awareness: A community-based approach for preventing violence among African American youth. The Journal of Primary Prevention, 20 61-73. Ollendick, T. H., Weist, M. D., Borden, M. C., & Greene, R. W. (1992). Sociometric status and academic, behavioral, and psychological ad justment: A five-year longitudinal study. Journal of Consulting and Clinical Psychology, 60 80-87. Parker, J. G., & Asher, R. S. (1987). Peer re lations and later persona l adjustment: Are lowaccepted children at risk? Psychological Bulletin, 102 357-389. Pettit, G. S. (2004). Violent children in devel opmental perspective: Risk and protective factors and the mechanisms through which they (may) operate. Current Directions in Psychological Science, 13, 194-197. Pettit, G. S., & Dodge, K. A. (2003). Violent ch ildren: Bridging development, intervention, and public policy. Developmental Psychology, 39 187-188. Reiss, A. J. (1951). Delinquency as the fa ilure of personal and social control. American Sociological Review 16 196-207.
145 Reiss, A. J., & Roth, J. A. (1993). Understanding and Preventing Violence. Washington, D. C.: National Academy Press. Riggs, N. R., Greenberg, M. T., Kusch, C. A., & Pentz, M. A. (2006). The mediational role of neurocognition in the behavioral outcomes of a social-emotional prevention program in elementary school students: Eff ects of the PATHS curriculum. Prevention Science, 7 91102. Rollin, S. A., Kaiser-Ulrey, C, Potts, I., & Creason, A. (2003). A school-based violence prevention model for at-ri sk eighth grade youth. Psychology in the Schools, 40 403-416. Salzinger, S., Feldman, R. S., & Stockhamm er, T. (2002). An ecological framework for understanding risk for exposure to community violence and the effects of exposure on children and adolescents. Aggression & Violent Behavior, 7 423-451. Scott, T. M., Nelson, C. M., & Liaupsin, C. J. (2001). Effective instruction: The forgotten component in preventing school violence. Education & Treatment of Children, 24 309322. Sharkey, J. D., Furlong, M. J., & Yetter, G. ( 2006). An overview of measurement issues in school violence and school safety research. In Shane R. Jimerson & Michael Furlong (Eds.), Handbook of school violence and school sa fety: From research to practice. (pp. 365-382). Mahwah, NJ, US: Lawrence Erlbaum Associates Publishers. Skinner, C. H., Dittmer, K. I., & Howell, L. A. (2000). Direct Observation in School Settings. In Edward S. Shapiro & Thomas R. Kratochwill (Eds.) Behavioral Assessment in Schools (pp. 19-45). New York: The Guilford Press. Smith, D. C., Larson, J., & Nuckles, D. R. (2006) A critical analysis of school-based anger management programs for youth. In Shane R. Jimerson & Michael Furlong (Eds.), Handbook of school violence and school sa fety: From resear ch to practice. (pp. 365382). Mahwah, NJ, US: Lawrence Erlbaum Associates Publishers. Smith, E., Gorman-Smith, D., Quinn, W., Ra biner, D., Tolan, P., & Winn, D. (2004). Community-based multiple family groups to prevent and reduce violent and aggressive behavior: The GREAT Families Program. American Journal of Pr eventive Medicine, 26 39-47. Sneddon, H. (2003). The effects of maltreat ment on children's health and well-being. Child Care in Practice, 9 236-250. Soderman, A. K., Gregory, K. M., & O,Neill, L. T. (1999). Classroom applications to support emerging literacy. Scaffolding Emergent Literacy. Boston: Allyn and Bacon. Spivack, G., & Shure, M.B. (1974). Social adjustment of young children: A cognitive approach to solving real-life problems. San Francisco: Jossey-Bass.
146 Spoth, R,, Greenberg, M., Bierman, K., & Redmond, C. (2004). PROSPER communityuniversity partnership model for public ed ucation systems: Capacity-building for evidence-based, competence-building prevention. Prevention Science, 5 31-39. Stephens, R. D. (1994). Planning for safer a nd better school: School vi olence prevention and intervention strategies. School Pschology Review, 23 204-215. Strein, W., & Hoagwood, K. (2003). School ps ychology, a public hea lth perspective I: Prevention, populations, and systems change. Journal of School Psychology, 41 23-38. Sugai, G. (2007). Promoting behavioral comp etence in schools: A commentary on exemplary practices. Psychology in the Schools,44, 113-118. Sugai, G., Spraue, J. R, Horner, R. H., & Walk er, H. M. (2000). Preventing School violence: The use of office disipline referrals to as sess and monitor school-wide discipline interventions. J ournal of Emotional and Behavioral Disorders, 8, 94-101. Tabachnick, B. G., & Fidell, L. S. (2001). Using Mutlivariate Statistics, Needham Heights, MA. Allyn & Bacon. Taub, J. (2001). Evaluation of the Second St ep violence prevention program at a rural elementary school. School Pschology Review, 31 186-200. Thorton, T. N., Craft, C., Dahlberg, L. L., Lynch, B. S., & Baer, K. (2002). Best Practices of youth violence prevention: A sourcebook for community action Centers for Disease Control and Prevention. Washington, D.C. Tobin, T., & Sprague, J. R. (2000). Alternative ed ucation strategies: Reduc ing violence in school and the community. Journal of Emotional & Behavioral Disorders, 8 177-186. Tolan, P. H., & Gorman-Smith, D. (2002). What vi olence prevention resear ch can tell us about developmental psychopathology. Development and Psychopathology, 14 713-729. Tolan, P. H., & Guerra, N. G. (1994). Preventi on of delinquency: Current status and issues. Applied & Preventive Psychology, 3 251-273. United States Department of Justice (2005). School Crime and Safety Washington, D.C. Retrieved from: http://www.ojp.usdoj.gov/bj s/abstract/iscs05.htm on July 15, 2006. United States Public Health Service (2003). Youth Violence: A report of the Surgeon General Washington, D.C. Retrieved from: www.mentalhealth.org/youthviolence /surgeongeneral/SG_Site/chapter5/sec1.asp on May 22, 2003 Vernberg, E. M., & Gamm, B. K. (2003). Resi stance to violence prevention intervention in schools: Barriers and solutions. Journal of Applied Psyc hoanalytic Studies, 5 125-137.
147 Vitale, C. A. (2001). Begin at th e beginning: Violence prevention at the elementary school level. Nursing Forum, 36 25-32. Walker, H. M., & Gresham, F. M. (1997) Making schools safer and violence free. Intervention in School & Clinic, 32 199-205. Warwick, P., & Maloch, B. (2003). Scaffoldi ng speech and writing in the primary classroom. Reading: Literacy & Language, 37 54-63. Weist, M. D. (2003). Challenges and opportunities in moving toward a public health approach in school mental health. Journal of School Psychology, 41 77-82. Weist, M. D., & Cooley-Quille, M. (2001). A dvancing efforts to address youth violence involvement. Journal of Clinical Child Psychology, 30 147-151. Werner, N. E., & Crick, N. R. (2004). Maladaptiv e peer relationships a nd the development of relational and physical aggr ession during middle childhood. Social Development, 13 495-514. White, S. H. (1965). Evidence for hierarchical arrange ment of learning processe s. In L. P. Lipsett & C. C. Spiker (Eds.) Advances in child development and behavior, 2 (pp. 164-192). New York: Academic Press. Wood, J. J., Cowan, P. A., & Baker, B. L. (2002) Behavioral problems and peer rejection in preschool boys and girls. Journal of Genetic Psychology, 163, 72-88. Wurdinger, S. D. (2005). Why experiential learning works. Using experiential learning in the classroom (pp. 8-17). Lanham, MD: Scarecrow Education. Yell, M. L., & Rozalski, M. E. (2000). Searching for safe schools: Legal issues in the prevention of school violence. Journal of Emotional & Behavioral Disorders, 8 187-197.
148 BIOGRAPHICAL SKETCH Ms. Christian completed her bachelors de gree in 1995 at Marshall University in Huntington, West Virginia. She later completed her masters degree in clinical psychology at Marshall in 1995. Following her work as a mental health counselor, she began the School Psychology doctoral program at the University of Florida in 2001.