INCREASING FAMILY EMPOWERMENT AND HOPEFULNESS IN CREATING POSITIVE PARENTING ALTERNATIVES FOR HIGH-RISK FAMILIES OF CHILDREN WITH DISABILITIES. Michaela Farber1, Ravita Maharaj1, 1Catholic University of America, Washington, DC United States
Family empowerment and positive hopeful outlook at life have been linked with desirable parenting outcomes and have been deemed of particular importance to family well-being when parents cope with the adversity of parenting children with disabilities. Vulnerabilities in the family environment and structural characteristics have been known, on the other hand, to promote risks for poor family outcomes. To increase the empowerment and hopeful outlook of high-risk families raising children with disabilities in urban areas of Washington, D.C., a preventive family strengthening intervention was implemented between 2000 and 2002. The intervention was based on the theoretical practice model for social development which emphasizes strengths-based approach in altering opportunities afforded by the environment, strengthening skills in solving life´s problems, and balancing information in promoting prosocial behavior. The manualized parent education curriculum has never before been evaluated with African American families. The project recruited a sample of 51 voluntary parenting participants. The evaluation used a mixed group design with a follow up booster session. The contents of the program were applied in 13 weekly sessions including various incentives. Parent outcome data were derived from three cohort implementations. Adapted standardized measures examined attitudinal changes in participants´ views of family empowerment, hopeful outlook, aggression, parent-child interaction, and community involvement. The quantitative findings evidence beneficial outcomes within the limitations posed by the pre-experimental design, attitudinal measures, and motivational interests of the participants. The program experienced a modest 23% attrition, graduating 39 participants. Paired t-tests documented statistically significant changes in participants´ reduction in verbal aggressiveness and improvement in parent-child interaction, hopefulness, and community involvement. Repeated measures of analysis of variance demonstrated a significant positive increase in the participants´ family empowerment. Participants´ hopeful outlook outcome contributed 22% and their pre-test empowerment 18% of the variance in their post-test family empowerment. Because children with developmental challenges exhibit a wide range of problem behaviors that often do not diminish with age and parents´ management of these problem behaviors affect their own and their children´s well-being, this program´s attempt at addressing parents´ hopefulness along with their empowerment contributed to parents´ strengthening of their overall childrearing abilities.
A COMPARISON OF INTRA-CLASS CORRELATION COEFFICIENTS OF DRUG-RELATED BEHAVIORS BETWEEN URBAN AND RURAL SETTINGS: PROJECT NORTHLAND. Kari Kugler1, Melissa Stigler1, Kianoosh Farbakhsh1, Kelli Komro2, Cheryl Perry1, 1University of Minnesota, Minneapolis, MN United States; 2University of Florida, Gainesville, FL United States
Interest in group-randomized trials has emerged in recent years and many studies have been published using this design. Traditional analysis techniques, using only fixed effects, are not appropriate for group-randomized trials because of the clustering of observations within a group. Intra-class correlation coefficients (ICCs), the proportion of variance between groups and total variance, become an important component during analyses. Typically, an ICC of 0.10 is considered quite high, 0.05 moderate and 0.01, low. However, any ICC can have an affect on the interpretation of the results. Furthermore, ICCs for various outcomes and settings are not widely available. Having access to empirically-based estimates would be helpful when planning for the design and analysis of these kinds of trials. The goal of this paper is to explore the ICCs of various drug-related behaviors in the context of an alcohol-use preventive intervention. Comparisons will be made between Project Northland (1991-1994), implemented in a rural setting (northeastern Minnesota), and Project Northland for urban youth (2002-2005), implemented in an urban setting (Chicago). Briefly, Project Northland (1991-1994) was a community-wide, multiple-component intervention program designed to reduce and prevent alcohol use among youth living in rural and primarily white communities. Twenty-four school districts were recruited and randomized to treatment or control condition; a cohort of 1,901 students was followed, starting in 6th grade and ending in 8th grade. Project Northland for urban youth is an adaptation of the original intervention program and is currently being implemented in Chicago among ethnically heterogeneous communities. In 2001, sixty schools were randomized to treatment condition; a cohort of 2,759 students has been followed, starting in 6th grade and ending in 7th grade. The analyses will include ICC estimates for rural versus urban schools and trends over time (6th and 7th grade) for each of the following drug-related behaviors: past year alcohol use, past month alcohol use, ever smoked cigarette, past year marijuana use and past month marijuana use. The results of these analyses will be described and implications for future research discussed.
THE RELIABILITY AND VALIDITY OF SELF-REPORTED MEASURES TO EVALUATE ADOLESCENT HIV/AIDS PREVENTION PROGRAMS. Kari Kugler1, K. Mnyika2, Kelli Komro3, M Masatu2, An Astrom4, Aklei Kessy2, Melkizedeck Leshabari2, Knut-Inge Klepp4, 1University of Minnesota, Minneapolis, MN United States; 2University of Dar es Salaam, Dar es Salaam, Tanzania; 3University of Florida, Gainesville, FL United States; 4University of Oslo, Oslo, Norway
Reliability and validity of measures used to describe concepts central to behavioral theories are critical to our understanding of these concepts, theories and their application to behavioral interventions. However, quite often, the psychometric properties of survey instruments designed to measure the concepts common to behavioral theories are not extensively published and limit our comprehensive understanding. The purpose of this paper is to present the results of psychometric analyses on measures used to evaluate the Tanzanian Health Systems Research and Health Promotion program, a reproductive health promotion intervention program for young adolescents. More specifically, the results from analyses focusing on measures used to evaluate risky sexual behavior and its determinants, as it pertains to HIV transmission, will be presented. The survey instrument was modeled after previous research conducted with adolescents in the Kahe region of Tanzania. The survey instrument was tested using different adolescent samples for each stage of its development. The 246-item self-administered questionnaire was first pilot-tested and revised. A test-retest sample of 67 students was used to assess reliability of the instrument. Finally, the main scale development sample included 6th and 7th graders in 11 villages throughout the Kahe region of Tanzania (n=524). Theoretically-driven scales of determinants of risky sexual behavior included intrapersonal (e.g. communication skills, decision making skills, ability to resist pressure to have sex), social (e.g. overall parental connection and perceived subjective norms related to sexual activity) and cultural-environmental influences (e.g. HIV knowledge, perceived HIV severity and attitudes towards people with AIDS). Assessment of the reliability of scales was determined by test-retest and internal consistency, e.g. Chronbach´s alpha. Assessment of the validity of scales was determined by face, discriminant and predictive validities. The results indicate that the scales constructed are reliable and valid measures of young adolescents´ risky sexual behavior, as well as intrapersonal, social, and cultural-environmental factors related to that behavior.
LOGISTIC GROWTH CURVE ESTIMATION OF SUBSTANCE INITIATION CONTAGION MODEL PARAMETERS. Cleve Redmond1, Chungyeol Shin1, Changwan Kang2, Richard Spoth1, 1Iowa State University, Ames, IA United States; 2Dongeui University, Seoul, South Korea
When measured over a sufficient length in time, growth in population substance initiation rates typically exhibit an S-shaped curve. This finding is consistent with a diffusion or epidemic/contagion model of substance use adoption. Within these models, a key factor determining the rate at which substance use behaviors will be adopted in a population is the degree to which members of that population are exposed to others who have already adopted the behavior (exposure would be associated with peer influence factors, behavior modeling, opportunity, etc.). From a more mathematical standpoint, the expected proportion of a population that will initiate use between one time point and the next can be expressed as the product of (a) the proportion of the population that had initiated use at the initial time point, (b) the proportion of the population that had not yet initiated use at that time point (i.e., the proportion who could subsequently initiate), and (c) a growth rate adjustment factor, referred to here as g (included because not all substances are as readily adopted and adoption rates may vary across different populations). Unfortunately, direct estimation and testing of g in the discrete-time contagion model is difficult. An alternative model that can be used to describe S-shaped initiation curves is the logistic growth curve (LGC) model. LGC models can be estimated using widely available software, but model parameters are not easily interpreted from a substantive perspective. However, it can be mathematically demonstrated that the LGC model is the continuous-time equivalent of the discrete-time contagion model. Further, the discrete-time model parameter g, can be expressed as a function of LGC parameters. The substantive interpretation of g is straightforward (an adjustment to the probability of initiation per unit of time), and is particularly useful when describing differences in initiation rates across groups. For example, in a preventive intervention efficacy trial, the difference in gs between intervention and control groups would be interpreted as the intervention-induced change in the probability of initiation from one time point to the next. The mathematical relationship between the discrete-time contagion model and the continuous-time LGC model is described, including the calculation of g from LGC parameters. The utility of g as an indicator of intervention effect on initiation of drunkenness is then illustrated using data from a preventive intervention efficacy trial involving 446 students enrolled in 22 schools. Data were collected from students beginning in the 6th grade and concluding in the 12th grade. A brief universal family skills program was offered to students and their parents in 11 of the 22 schools during the 6th grade.
METHODOLOGICAL CHALLENGES IN EVALUATING ABSTINENCE EDUCATION PROGRAMS: HOW DO WE KNOW IF THEY´RE NOT REALLY DOING IT?. Patricia Thompson1, Joanne Jenson1, Barbara Cohen1, Leslie Raneri1, Allison Roper1, Alma Golden1, 1U.S. Department of Health and Human Services, Rockville, MD United States
Abstinence education programs as a public health approach for addressing teenage pregnancy and the spread of HIV/AIDS and other sexually-transmitted infections had an early start with the 1981 enactment of Adolescent Family Life legislation (also known as Title XX of the Public Health Service Act). Additional Federal abstinence education programs were launched in the 1990s. All such programs have in common the goal of promoting abstinence from sexual activity until marriage as the most effective prevention approach to these public health problems. Abstinence curricula teach the knowledge, values, attitudes and skills that help undergird behavioral commitment to sexual abstinence among young people. To determine the effectiveness of abstinence programs, the Federal Government has given increasing attention to the evaluation of abstinence programs. The Title XX prevention programs, featured in this poster presentation, are legislatively mandated to be demonstration/evaluation programs from their outset.
The poster presentation briefly describes some representative Title XX prevention projects, which reveal a range of strategies for teaching sexual abstinence to young people from a variety of population settings. The focus of the poster presentation, however, is on articulating methodological challenges faced by Title XX evaluators.
The following challenges are developed, using particular examples and proposing given resolutions: 1) determining appropriate definitions for key concepts such as “sexual activity” and “abstinence”; 2) obtaining valid measures of behaviors which are inherently private and subject to self-report error; 3) addressing social, cultural and political difficulties in asking questions about adolescent sexual behavior: and 4) establishing valid control or comparison groups for gauging abstinence program effects as distinct from other factors that could account for observed program changes in clients pre and post program intervention.
MEASURING TEACHERS' DEPTH OF UNDERSTANDING OF PREVENTION CONCEPTS. Linda Dusenbury1, William Hansen2, Julia Jackson-Newsom1, Costella Green3, 1Tanglewood Research, Greensboro, NC United States; 2Tanglewood Research, Inc., Greensboro, NC United States; 3Prevention First, Chicago, IL United States
As prevention programs become widely disseminated, the most serious threat to effectiveness is maintaining the quality of implementation that was intended by developers. There is a current emphasis within the field on promoting adherence to teaching, meaning that teachers should follow the curriculum as outlined. However, nearly all teachers adapt programs and many in ways that are not consistent with the goals of the program. In order to adapt successfully, it is likely that teachers need to have an understanding of the conceptual goals of the program. To date, depth of understanding of prevention concepts has not been assessed or examined in research. This paper describes the pilot test of a measure for assessing teachers´ depth of understanding conducted with 16 teachers in Chicago public schools. These teachers had all just completed teaching All Stars, a substance abuse prevention program for middle school students. All Stars emphasizes norm setting, idealism, and positive commitments. Because norm setting is often misunderstood by teachers, this project focused specifically on depth of understanding of norm setting. All teachers were interviewed over the telephone by a Ph.D. level researcher who recorded responses to five detailed questions. Specifically, teachers were asked how they would define norms, why they thought it was important for students to have positive norms, which sessions (out of 12) of All Stars addressed norms, how sessions that addressed norms accomplished this goal, and what other efforts they made (outside of program) to promote positive norms. Responses to each question were scored on a four point scale (from -1 to +2), with -1 being awarded when teachers failed to mention a session that was designed by the developer to address norms, 0 indicating a fundamentally flawed response, 1 indicating a response that hinted at a normative topic but failed to define how norms were addressed, and 2 indicating a very strong response. Teachers' scores were calculated by summing across all ratings. Two senior researchers independently rated responses. Raters´ scores correlated highly (r=.97). There was agreement about the mention of normative issues (k=.812) and overall score (k=.732). For both raters, minimum ratings were -1 and maximum ratings were 17 on this scale. Rater 1 gave teachers an average rating of 5.78, rater 2 gave average ratings of 6.38, indicating there is room for improvement in depth of understanding among most teachers.
PREVENTING DRUG USE AND ABUSE WITHIN A DEVELOPMENTAL CONTEXT:UNDERSTANDING THE THEORY BEHIND THE CONSEQUENCES. Michael D. Newcomb1, Thomas Locke2, 1University of Southern California, Alhambra, CA United States; 2University of California, Los Angeles, Los Angeles, CA United States
Several epidemiological researchers have begun a theoretical integration of the rather rich drug use and abuse etiology literature and the far less developed and somewhat paltry drug use and abuse consequences research findings. This integration has numerous advantages as well as theoretical and methodological challenges (Newcomb, 2004). Overall, drug use consequences can be thought of as relating to three main processes: Direct, indirect, moderating, and mediating effects.
Direct effects vary across different developmental periods and are related to the specific influence of a drug on later biopsychosocial functioning. Indirect effects are related to the influence of another person´s drug use (parents, peers, partners) on the individual´s life (e.g., psychological distress due to a partner´s drug use). Indirect effects emphasize the importance of a systemic or holistic point of view. Indeed, drug use can serve as a factor that alters other processes that may already be in play, as a moderator or a mediator.
Few people have studied drug use as a moderator. Studying drug use as a moderator investigates how it “affects the relationship between two variables, so that the nature of the impact of the predictor on the criterion varies according to the level or value of the moderator Further, a moderator interacts with a predictor variable in such a way as to have an impact on the level of a dependent variable” (Holmbeck, 1997, p. 600). This can be represented in questions such as Can drug use exacerbate the effects of a poor upbringing?
As a mediator, drug use is viewed as a “generative mechanism through which the focal independent variable is able to influence the dependent variable of interest” (Baron & Kenny, 1986, P. 1173). This process can be represented in questions such as Does drug use influence the relationship between early experience and later development? As mediators (drug use and abuse) are predicted by various risk and protective factors. Also, they themselves are the predictors of later consequences and outcomes.
Studying drug use as a moderator or mediator of different domains is an essential and often overlooked component in consequence-focused research. Too often drug use is studied in relative isolation, ignoring the possibility that it is intertwined with other norm-violating behaviors and attitudes (McGee & Newcomb, 1992). Drug use is only one discrete component of the myriad biopsychosocial forces that forge a human being.
TECHNICAL ADEQUACY OF STRENGTHS AND DIFFICULTIES QUESTIONNAIRE FOR AFRICAN AMERICAN CHILDREN. Laura Zionts1, Karla Anhalt1, 1Kent State University, Kent, OH United States
This study confirms the concurrent validity and internal consistency of the Strengths and Difficulties Questionnaire (SDQ) (Goodman, 1997) with the Child Behavior Checklist Teacher Report Form (Achenbach, 1991). Teachers of 229 African American public school children in grades k-3 completed the SDQ, and about 1 week to 1 month later, they completed the Child Behavior Checklist Teacher Report Form. The results indicate that the SDQ is useful as a brief measure of adjustment for African American children. Use of this scale for diagnostic/clinical purposes is not recommended. The SDQ appears to be an appropriate measure to evaluate constructs of adjustment within a developmental study with multiple variables. The SDQ has been validated using a five factor structure reported for British, Arab and Swedish samples. Separate confirmatory factor analyses were performed on the SDQ Parent form and SDQ teacher form in order to determine whether items on these measures fit the proposed 5-factor structure. The 5-factor structure was not confirmed for either form of the SDQ; though the SDQ-Teacher form has a slightly better fit than the SDQ-Parent form. Once Promax rotation was performed constraining to five factors, interpretation was attempted. Only two items loaded on the fifth factor and its alpha was .39. A second Promax rotation was performed constraining to four factors. These were more conceptually clear and had 4-6 items loading on each factor. Overall conclusions are that the SDQ is useful in community studies where the focus of research is NOT on following children with psychopathology. It appears helpful in prevention and developmental research.
THE NATIONAL REGISTRY OF EFFECTIVE SUICIDE PREVENTION PROGRAMS. Philip Rodgers1, Howard Sudak1, 1American Foundation for Suicide Prevention, Philadelphia, PA United States
In response to the problem of suicide, hundreds of suicide prevention programs have been created and utilized across the country. However, there has been a dearth of available evidence concerning what constitutes an effective suicide prevention program. This need was codified in Objective 10.2 of the National Strategy for Suicide Prevention, which states “By 2005, establish and maintain a registry of prevention activities with demonstrated effectiveness for suicide or suicidal behaviors”. To meet this objective, the National Registry of Effective Suicide Prevention Programs (NRESPP) was created through a collaboration between the Suicide Prevention Resource Center and the American Foundation for Suicide Prevention. This presentation will provide information related to the development and implementation of the NRESPP review process and the function of expert and community panels in the review and product-development phases. The role of NRESPP in the context of the larger evidence-based movement will also be discussed, especially in light of increasing cross-disciplinary demands for the use of rigorous methodology to determine effectiveness and provide accountability.
VALIDITY TEST OF THE ALEXSA (ASSESSMENT OF LIABILITY AND EXPOSURE TO SUBSTANCE USE AND ANTISOCIAL BEHAVIOR). Ty Ridenour1, Mark Greenberg1, Amanda Miller1, 1Pennsylvania State University, State College, PA United States
The ALEXSA is an illustrated ACASI assessment of children´s and early adolescents´ `risk and protective factors´ for substance use and antisocial behavior. ALEXSA base measures span a broad range of domains including demographic, social, family, community environment, interpersonal (academic, affective, behavioral, biological, cognitive, competency, and religiosity), and substance use-specific risks. New measures can be added to the ALEXSA. Previous studies indicate that children enjoy completing the ALEXSA, nearly all of the base measures are reliable, and that concurrent use of tobacco, alcohol, and inhalants as well as conduct disorder criteria are well-predicted. This study was conducted to examine the convergent and discriminant validity of ALEXSA base measures using three samples. The first sample included 117 nine to 12 year old students in an after-school program designed for Mexican immigrants. Correlations indicated that ALEXSA measures correlated with program intake measures of similar characteristics, except for academic achievement. The second sample consisted of a comparison between 78 nine to 12 year old psychiatric inpatients and 262 school-recruited students. Psychiatric inpatients had worse scores on average compared to school-recruited participants on measures of psychiatric symptomotology as well as tolerance of deviancy, peer pressure susceptibility, peers´ conduct disorder, alcohol & tobacco use, and gambling. The third sample consisted of 388 eight to 13 year old Pennsylvanians in a summer camp for children experiencing long-term stressors. ALEXSA measures correlated consistently with hypothesises with campers´ application data. With few exceptions, results indicated that ALEXSA base measures were valid and may be useful for preventive and etiological measurement. Current expansion of the ALEXSA includes development of three Spanish versions, creation of gender- and ethnic-specific illustrations, and measures for specific prevention and etiological studies. Development of the ALEXSA base measures is funded by NIDA (K01 00434) and the Penn State University Children, Youth, and Families Consortium.