Health Education Research, Vol. 18, No. 4, 461-476,
August 2003
© 2003 Oxford University Press
Development of a youth survey to measure risk behaviors, attitudes and assets: examining multiple influences
School of Public Health, Regional Campus at Brownsville, Brownsville, University of Texas–Houston, TX 78520, and 1 Department of Health Promotion, Education and Behavior, Norman J. Arnold School of Public Health, University of South Carolina, Columbia, SC 29208, USA
*To whom correspondence should be addressed E-mail: evansae{at}gwm.sc.edu
 |
Abstract |
|---|
 Top Abstract IntroductionMethodsDescription of instrumentResultsDiscussionReferences |
|---|
Currently, most surveys assessing adolescent health concerns focus primarily on risk behaviors and negative influences rather than positive influences such as assets. The purpose of this paper is to describe the development and validation of the Adolescent Health Attitude and Behavior Survey (AHABS). This instrument was developed to measure the prevalence of youth health risk behaviors, attitudes towards adolescent sexual behavior and youth assets in a statewide evaluation effort. The questionnaire was completed by 4368 public high school students in Grades 9–12. Content validity was established through an extensive review of literature, a group process and factor analyses. Reliability was established through Cronbach’s
coefficients. Factor loadings ranged from 0.48 to 0.84 for scales measuring attitudes towards adolescent sexual behavior and coefficients ranged from 0.61 to 0.81. Factor loadings ranged from 0.34 to 0.90 for scales measuring youth assets and coefficients ranged from 0.69 to 0.85. Because of several limitations (e.g. construct validity was not measured), additional development work is needed. Therefore, the AHABS is still in a developing, but promising, state. Additional psychometric work will provide program practitioners and evaluators with a psychometrically sound tool to measure behaviors, attitudes and assets. |
Introduction |
|---|
|
TopAbstractIntroductionMethodsDescription of instrumentResultsDiscussionReferences |
|---|
Adolescence is the period in life characterized by significant change. Biological, psychological changes and social learning changes occur at an astonishing rate (Lerner, 1980
). Adolescents are more capable of making decisions (Hoffman et al., 1994) and are often given more freedom to do so. It is also a time when youth are particularly impressionable and vulnerable to many environmental factors (Fullerton, 1994) that positively or negatively influence their future health behaviors.
Traditionally, researchers have conducted studies that supported society’s wish to recognize the signs and symptoms of predisposing or developing risk traits and behaviors in children, so that problem behaviors could be lessened or avoided. Adolescent risk behaviors and choices tend to occur in a social context and may be synergistic. For example, evidence suggests that teenage substance abuse is correlated with numerous risk behaviors including delinquency, conduct disorders at school, school dropout, violent and aggressive behaviors, and unplanned and unprotected sexual intercourse (Jessor and Jessor, 1977; Zabin et al., 1986; Elliott et al., 1989; Richter et al., 1993; Coker et al., 1994; Baker et al., 1995; Valois et al., 1995). According to Hawkins et al. (Hawkins et al., 1992), the more risk factors a child has in his/her life, the more likely he/she is to become involved in ‘problem’ behaviors. As a result of this focus on risk factors, the majority of current surveillance efforts assessing adolescent health issues focus primarily on negative behaviors and influences (Kolbe, 1990; Garrison et al., 1993; Kolbe et al., 1993; Valois et al., 1995, 1999).
However, an increasingly popular approach to youth prevention involves investigation beyond risk factors to include identifying and establishing the prevalence of protective factors among adolescents. Protective factors are positive characteristics, predispositions and influences in adolescents’ lives that can buffer them from negative influences (Benard, 1991). Over time, protective factors can help an adolescent become more resilient and more able to resist negative influences (Benard, 1991; Rak and Patterson, 1996; Benson, 1998). Jessor et al. (Jessor et al., 1995) determined that there is an inverse relation between protective and risk factors in the prediction of problem behavior. They found that the greater number of protective factors present in the lives of adolescents, the lower engagement in problem behaviors.
Researchers who have studied resiliency have identified certain characteristics that make adolescents more resilient. Examples of resiliency factors include involvement in structured activities, parental boundary setting, religious commitment and adult mentors (Jessor et al., 1995; Greene, 1998). This relatively new focus on resiliency and protective factors has offered researchers an alternative from a pathology model that tends to be overly problem-focused (Pittman and Wright, 1991).
Building on the research of resiliency and protective factors are youth development interventions. Recent research provides evidence that deficit-only-focused strategies are not comprehensive and lack the valuable empowering components of approaches grounded in youth development theory (Roth et al., 1998). The emergence of the youth development approach (with its focus on positive adolescent competencies, protective factors and resources) has shown promise in adolescent pregnancy prevention (Kirby, 1999) and other adolescent health issues. The youth development approach considers the common underpinnings of multiple problem behaviors such as teen pregnancy, substance abuse, delinquency and school dropout. By simultaneously addressing multiple risk behaviors and building resiliency, youth development interventions are comprehensive and possibly more effective (Barton et al., 1997). Furthermore, while youth development programs tend to focus on building competencies and empowering responsible behavior, they naturally address personal deficits. Programs that enhance protective factors and take into account risk factors (i.e. deficits) appear to be promising, particularly for substance abuse prevention (Hawkins et al., 1992).
A core framework for the youth development approach to prevention programming is the Developmental Assets Framework suggested by the Minneapolis-based Search Institute (Leffert et al., 1998; Scales and Leffert, 1999). Benson et al. (Benson et al., 1998) noted that the Framework blends risk factors, resiliency and protective factors that precede health outcomes. The Developmental Assets Framework suggests 40 assets that can be enhanced (when present) or established (when initially absent) in youth. Half of these assets are suggested as internal and are labeled as the following domains: educational, community, values, social competency and positive identity. The other 20 assets are external, suggesting that they support resources available to adolescents and are labeled as the following domains: support, empowerment, boundaries and expectations, and time. As with protective factors, the more assets a youth has, the more likely he/she is to avoid problems such as substance abuse, teen pregnancy or delinquency (Scales and Leffert, 1999).
Based on the current varied emphases in adolescent health research, there is a need for psychometrically sound instruments to measure the multiple influences in adolescents’ life including risk behaviors, attitudes and assets or protective factors. A better understanding of the various psychosocial and behavioral influences on adolescents could assist efforts in school and community settings to promote life-long health. Currently, few instruments with strong psychometric properties exist that assess both risk and protective factors. The Search Institute’s framework and instrument provide a partial foundation for psychometrically sound instrumentation (Benson et al., 1998); however, further subscale development based on factor analysis results and strong reliability estimates are still needed. Therefore, the purpose of this paper is to describe the development, validity and reliability of the Adolescent Health Attitude and Behavior Survey (AHABS) that measures risk behaviors, attitudes towards adolescent sexual behavior and youth developmental assets. This instrument was developed as an impact evaluation instrument for a statewide evaluation examining the effectiveness of county-based teen pregnancy prevention programs. As an impact evaluation instrument, the survey was designed to assess intervention effectiveness in producing change in knowledge, attitudes, beliefs and behaviors (Windsor et al., 1994) in the counties. This survey also includes two other subscales that measure sexual knowledge and other psychosocial variables related to adolescent sexual behaviors. These two subscales are not described in this paper because they are less relevant to broad-based youth prevention activities.
For the purpose of this paper, we will describe three sections of the survey: (1) health risk behavior section, (2) the attitudes towards adolescent sexual behavior subscales and (3) the youth developmental assets subscales. Each section will be described individually and in relationship to one another. The results of correlating one of the measures of risk taking, the level of sexual activity, to the attitudes and assets subscales also will be presented.
|
Methods |
|---|
|
TopAbstractIntroductionMethodsDescription of instrumentResultsDiscussionReferences |
|---|
Questionnaire development and pilot testing
The development of the AHABS instrument was conducted in phases. First, we searched the literature for other instruments measuring adolescent risk-taking behaviors, attitudes and adolescent development, including youth developmental assets. Second, based on the literature review, we identified the following instruments: CDC Youth Risk Behavior Survey (YRBS) (Kolbe, 1990
), the Adolescent Curriculum Evaluation (ACE) Questionnaire (Locke and Vincent, 1995), the Youth Sensitive Survey (McLeroy et al., unpublished), and the Survey of Student Resources and Assets by the Search Institute (Leffert et al., 1998). Two of the instruments have been used nationally (CDC-YRBS and Search Institute Survey), the ACE survey had been used with teens living in South Carolina and the Youth Sensitive Survey was used in a teen pregnancy prevention project evaluation. Therefore, these four surveys provided either well-tested items or concepts relevant to teen pregnancy prevention evaluation that made them useful in the development of another tool. However, none of these instruments provided the full compilation of measures needed for our evaluation efforts.
Next, through a group process, we identified the need for five different sections on the AHABS instrument. A similar process of examining which assets and adolescent attitudes would be most affected by community-based teen pregnancy prevention projects was suggested by McLeroy et al. (pers. commun.) in the development of their evaluation tools. The five sections of the AHABS instrument also reflect individual, environmental and behavioral factors described in Bandura’s Social Cognitive Theory (Bandura, 1986). These sections include: demographics, health risk behaviors, attitudes towards adolescent sexual behavior, youth assets, and another section measuring psychosocial and knowledge variables related to sexuality behaviors. [Original survey available upon request from the fifth author and Principal Investigator of the evaluation study (M. L. V.).]
During the last phase of development, we pilot tested the original version of the instrument in a classroom setting with 755 high school students (composed of students in six classes, primarily in Grades 9–12, from five schools) demographically similar to the study population for the validation study. Based on the results from the pilot test, the wording for several items was altered to clarify the meaning of the questions. Reliability and validity estimates were calculated for the scales measuring attitudes towards adolescent sexual behavior and youth assets. Based on those calculations, several items were dropped for the final version of the instrument.
|
Description of instrument |
|---|
|
TopAbstractIntroductionMethodsDescription of instrumentResultsDiscussionReferences |
|---|
Demographics
Based on the CDC-YRBS instrument (Kolbe, 1990
), items in the demographic section of the AHABS were included to capture essential data on age, gender, grade level and race/ethnicity. In addition, one item measuring family structure ‘Who are the primary adults who live in your home?’ and one item serving as a proxy measure for socioeconomic status (SES) ‘At school, are you eligible for free or reduced price lunch?’ were also included in this section.
Health risk behaviors
Because this survey had an evaluation purpose of measuring effectiveness of county-based teen pregnancy prevention initiatives, this section contained 12 items on sexual risk-taking and 14 items measuring other risk behaviors including intentional/unintentional injury (n = 6), tobacco use (n = 1), alcohol use (n = 2), and marijuana and other illicit drug use (n = 3). A variety of risk behaviors were included because of their documented association with risk for unintended pregnancy (Richter et al., 1993; Coker et al., 1994; Baker et al., 1995; Valois et al., 1999).
The majority of items measuring risk behaviors were selected from the CDC-YRBS (Kolbe, 1990). The CDC-YRBS behavioral items have shown no significant difference in responses from Time 1 to Time 2, indicating good reliability especially for high school students (Brener et al., 1995). Response options for the risk questions were designed so that lower numbers represented lower risks.
One risk behavior item, the level of sexual activity (based on number of lifetime sexual partners), was used to examine validity with the scales measuring attitudes toward adolescent sexual behavior and youth assets. Past research has suggested that risks and assets are inversely correlated (Jessor et al., 1995; Scales and Leffert, 1999). Level of sexual activity among the respondents was measured through an item asking how many people the survey respondent has had sex with in their lifetime. The seven response options ranged from ‘I have never had sex’ to ‘six or more people’.
Attitudes towards adolescent sexual behavior
Eleven items measuring attitudes towards sexual intercourse during adolescence were included in the survey. Ideas for items specifically focused on sexual attitudes were based on the ACE Questionnaire (Locke and Vincent, 1995) and the Youth Sensitive Survey (McLeroy et al., unpublished), but no specific items were taken from either of these surveys. Instead, the investigators created new items to measure sexual attitudes. These items measured attitudes regarding self and peer sexual behavior. The items were on five-point Likert scales with choices such as ‘strongly agree’ to ‘strongly disagree’, ‘only after marriage’ to ‘on a first date if the girl/boy agrees’ and ‘none of them’ to ‘all of them’. Higher scores indicated higher risks. Because the attitude items were new, psychometric evaluation was needed along with analyses of the associations between risks, assets and attitudes.
Youth assets
Items in this section of AHABS were chosen because they measured assets that a teen pregnancy prevention project in a community setting would be most likely to change. Based on the desire to measure certain assets, concepts or actual items from the survey of Student Resources and Assets by the Search Institute (Leffert et al., 1998) were used. Several items, with permission, were taken from the Search Institute survey; some modified based on formatting or content. Other items were developed based on the description of the assets provided by the Search Institute. Past reporting of the psychometric properties of the Search Institute surveys does not specifically delineate which survey items are measuring which assets and some assets are measured by single items (Leffert et al., 1998). Moreover, the Search Institute’s measurement of assets is driven primarily by literature review, expert understanding and ease of understanding for the public. Additionally, the individual asset categories are not scored in any analyses.
The categories (or the asset subscales) represent the way in which the assets are used for public communication and education purposes; they are generally too multidimensional to ‘hang together’ psychometrically within the categories. [(Leffert et al., 1988
With the primary purpose of this study being to develop a psychometrically sound instrument measuring risks, attitudes and assets, the AHABS instrument selected or created multiple items to measure each asset in the hope of creating psychometrically based subscales that would create useful subscale scores.
The AHABS asset items had response options based on a five-point Likert scale with choices such as ‘strongly agree’ to strongly disagree’, ‘delighted’ to ‘terrible’, ‘none’ to ‘all’, ‘0’ to ‘11 or more’. Higher scores indicated lower assets.
Data collection
Participants were recruited from public high schools in 45 South Carolina counties that received funding for the Adolescent Pregnancy Prevention Initiative. In counties that had three or fewer high schools, all of the high schools were contacted regarding participation. In counties with four or more high schools, three high schools that ‘best represented’ the total population aged 14–17 years were asked to participate. Best representation was based on race/ethnicity and school size. Once a school agreed to participate, a number of second period classes were randomly selected. All students in those classes received a passive parental consent form. Questionnaires were administered by trained evaluation staff during Period 2. Teachers were asked to remain present for survey administration; however, staff handled all aspects of data collection as one method for assisting in ensuring student anonymity. Most important in assuring student anonymity was that students were asked not to provide their name or any other identifying information on the answer sheets. Completion of the questionnaire took 30–40 min. These procedures were deemed appropriate by the referent university’s review board for the rights of human subjects in research.
Analysis
All survey items measuring youth assets and attitudes towards sexual behavior included in the analysis used Likert-type response options. Examples of these response options included ‘strongly agree’ to ‘strongly disagree’, ‘very much like me’ to ‘not at all like me’ and ‘terrible’ to ‘delighted’ with mid-point options such as ‘not sure’, ‘somewhat like me’ or ‘equally satisified or disatisfied’. Data were analyzed descriptively such that mean scores, standard deviations, frequencies and ranges were calculated. Exploratory factor analysis identified two subscales measuring Attitudes Towards Adolescent Sexual Behavior and seven subscales measuring Youth Assets. All factor analyses utilized principal axis with promax rotation. Factor analysis was initially performed on the entire sample (n = 4368). However, due to the large sample size, we also divided the sample into two randomly assigned groups and re-ran the factor analysis to examine if the initial factor pattern was maintained with the subsamples.
An item was assigned to a factor when its loading was at least 0.34 or higher and had no other loadings at 0.30 or higher on any other factor. Analysis of the eigenvalues in the scree plot was also used to confirm the number of factors identified by the factor loadings (Cattell, 1965a,b; Kim and Mueller, 1978; Hatcher, 1996). We created subscale scores by averaging the item scores for each scale. To assess internal consistency of the items, Cronbach coefficients were calculated for each subscale. To further validate the scales, relationships between each subscale and one risk item were explored. The relationships between the continuous subscales representing attitudes towards youth sexual behavior and youth assets was explored using Pearson correlations (significance level was set at 0.01). Relationships between the continuous scales and one ordinal risk item, the level of sexual activity of the respondents, was explored using Spearman correlations.
|
Results |
|---|
|
TopAbstractIntroductionMethodsDescription of instrumentResultsDiscussionReferences |
|---|
Sample population
Our sample included 4368 public high school students with a median age of 15 and approximately 53% of the participants (n = 2299) being female. Grade 9 and 10 students made up 60.7% of the respondent population. The students represented 43 high schools. Most participants were African-American (47%) and white (46%). Other ethnic groups included Hispanic (2%), American Indian (1%) and Asian (1%). All grades were evenly distributed by race/ethnicity. Approximately 41% of the participants were eligible for free or reduced lunch, and 48% reported living with both mother and father (real or adoptive). Nineteen percent reported living with mother only, and 15% reported living with mother and stepfather. Males and females were equally represented in all grades except Grade 9, with 67% of the respondents being female and 33% being male. The student response rate for schools that agreed to participate in the survey was approximately 87%. However, only 38% of schools asked to participate in the study did so. Schools in the upstate counties of the state, typically more conservative with a higher proportion of Caucasians, declined to participate in the study more often than other schools in the state. This was the only school/county characteristic that discerned school participation rates. Reasons for declining included: (1) school policies do not allow for ‘outside’ surveys, (2) lack of time to do the surveys due to other educational priorities and (3) prior agreement to participate in other health surveys.
Descriptive results
Table I displays the descriptive characteristics of the subscales measuring attitudes towards adolescent sexual behavior and youth assets. Mean scores on the subscales ranged from 2.21 to 3.44 on a five-point Likert scale. Descriptive results also indicated respondents used all response options available, thus decreasing the likeliness of subscales being skewed. Standard deviations ranged from 0.81 to 1.10 and variance ranged from 0.69 to 1.20.
|
Behavioral results
Items to measure risky health behaviors such as sexual intercourse, intentional/unintentional injury, tobacco use, and alcohol and drug use were assessed. Although not reported here, the prevalences of these behaviors in this study population are similar to other published studies of national and statewide adolescent health risk behaviors (Kolbe, 1990
; Valois et al., 1999). One health behavior that was chosen to include in this psychometric study was level of sexual activity measured by the number of lifetime sexual partners. Thirty-nine percent of the respondents had never had sexual intercourse. Approximately 18% reported having had sexual intercourse with only one sexual partner, while 10.9% reported having had two sex partners. Additionally, 19% of the respondents reported having had three to five sex partners during their lifetime and an additional 13% reported having sexual intercourse with six or more people during their lifetime.
Factor analysis results
Attitudes towards adolescent sexual behavior subscales
The attitudes towards adolescent sexual behavior section of the survey emerged with seven items (11 originally) which created two subscales. Two items were discarded because they did not load high enough on any one factor and two other items created a content cluster. The two subscales are Perceived Sexual Norms and Perceived Birth Control Use. Factor loadings for these two scales ranged from 0.48 to 0.84 (Table II). The coefficient for Perceived Sexual Norms was 0.81 and the coefficient for Perceived Birth Control Use was 0.61. Significant correlations were observed among the two attitudes towards adolescent sexual behavior subscales of the survey (r = 0.32), indicating that youth who perceive sex as less normative are also more likely to agree that they would use birth control. This correlation provide some evidence for discriminate validity (Hatcher, 1996) in that the scales are associated. However, the correlations are not so strong to indicate that they are measuring the same construct.
|
Two items clustered together creating a content cluster labeled Perceptions of Others Sexual Involvement. This content cluster had only two items and a reliability of 0.73 (Table II), but did show a distinct and consistent conceptual arena. Therefore, the content cluster is presented to show an area of promise for future scale development.
Youth assets subscales
The assets section of the survey initially contained 43 items. In the end, 33 items were retained and yielded seven subscales (Table III). Factor loadings ranged from 0.34 to 0.90 with over half of the items in each scale having factor loadings of at least 0.50. Four items were dropped either because their loadings were below 0.34 or their loadings did not discriminate between two different factors. All remaining items loaded at 0.34 or above on one of the factors and no higher than 0.30 on any other factor (Table IV). The factor pattern was consistent for the factor analysis using the entire sample and for the two randomly selected subsamples. The scree plot also showed an elbow at the seven-factor solution. Additionally, these seven factors all had eigenvalues greater than 1. Subscale items were also examined to determine if they factored together due to similarity in wording or placement in the survey. It was determined that three subscales, Youth Accountability to Parents and Other Adults, Quantity of other Adult Support, and Youth Empathetic Relationship may be loading together because they were presented as a group in the survey and because the response options were the same. Therefore, more examination of these subscales will be done in the future to assess their limitations. coefficients for the seven youth assets subscales ranged from 0.69 to 0.85. Significant correlations were found among each of the asset scales. The correlations were not so strong to indicate that the subscales were measuring the same constructs, thus providing discriminant validity (Hatcher, 1996).
|
|
Six items clustered together creating three content areas that did not meet the criteria to be considered a subscale because of the number of items and their reliability; however their factor patterns were distinct and consistent conceptually. Thus, the content clusters are presented in Table III as areas that show promise of possibly being developed into subscales in the future.
Relationship among attitude subscales, assets subscales and level of sexual activity
The Pearson correlation between the attitudes towards adolescent sexual behavior subscales and the assets subscales yielded primarily significant correlations (P < 0.01). The Perception of Others Sexual Involvement content cluster was not significantly correlated with two other content clusters: Youth’s Responsibility and Youth Planning. All correlations between attitudes towards adolescent sexual behavior subscales and assets subscales were in a positive direction, indicating that as assets increased, so did healthy attitudes towards adolescent sexual behavior. The strongest correlation (r = 0.56) was between the Youth Perceived Support by Parents and Other Adults Subscale and Youth’s Accountability to Parents and Other Adults Subscale, indicating that students perceive more support from parents who also hold them accountable for their actions. The correlation between Self Peer Values Regarding Risk Behaviors Subscale and Perceived Sexual Norms Subscale also was strong (r = 0.48), indicating that if youth report norms for adolescents to have sexual intercourse then they are more likely to also report that their closest friends are engaging in risky behaviors.
Level of sexual activity was significantly correlated with attitudes towards adolescent sexual behavior subscales and youth assets subscales (Table IV). The strongest correlation (r = 0.52) was observed between the item measuring level of sexual activity and the Perceived Sexual Norms Subscale, indicating that youth who perceived that it is okay for youth to have sex reported higher numbers of lifetime sexual partners.
|
Discussion |
|---|
|
TopAbstractIntroductionMethodsDescription of instrumentResultsDiscussionReferences |
|---|
The AHABS was created for use with high-school-aged youth from different ethnic backgrounds and SES status. Assessed in this survey are several health risk behaviors, attitudes towards other adolescents’ sexual behaviors and protective factors or assets.
The psychometric analysis yielded interpretable subscales for the attitudes towards sexual behaviors and assets sections of the survey, most with acceptable evidence for validity and internal consistency. The reliability and validity obtained within the asset scales are better or comparable to those obtained in similar scales for adolescents (Leffert et al., 1998).
More specifically, the attitudes towards adolescent sexual behavior items were identified and refined. Factor analysis identified two subscales: Perceived Sexual Norms ( = 0.81) and Perceived Birth Control Use ( = 0.61). The Perceived Birth Control Use subscale and the content cluster labeled Perceptions of Others Sexual Involvement ( = 0.73) will require more development in the future to obtain better conceptual representation and internal consistency.
Asset items examined in this study yielded seven subscales through exploratory factor analysis. These scales had coefficients ranging from 0.69 to 0.85. Three of the seven subscales, Youth Accountability to Parents and other Adults, Quantity of other Adult Support, and Youth Empathetic Relationship, need further study to ensure that they are behaving as true subscales rather than appearing to due so because of survey format. Additionally, psychometric development should be done with the three content cluster areas: Youth’s Responsibility, Youth’s Planning and Youth’s Satisfaction with Life. These content clusters require further development to enhance their conceptual representativeness and internal consistency.
In comparison to the Search Institute’s psychometric analysis results on their asset items (Leffert et al., 1998), the AHABS asset subscales tend to be broader than any given asset and narrower than their domains. However, each of the eight domains suggested by the Search Institute are measured to some extent by the AHABS. Additionally, the AHABS subscales cover more of the external factors affecting youth than the internal factors, although with further development of the content clusters more internal factors would be measured accurately. Finally, the topics measured by each of the AHABS subscales are easy to understand and can be scored by subscale, providing useful information to practitioners for program planning purposes.
Beyond the psychometric findings of our study, we also found significant correlations between the two subscales representing attitudes towards adolescent sexual behavior and all of the subscales representing youth assets. This finding suggests that the more assets an adolescent reports, the less likely they are to report attitudes reflecting support for their peers engaging in sexual behaviors. In addition, the significant correlations (P < 0.01) between the attitudes towards adolescent sexual behavior subscales and the item measuring level of sexual activity as well as the significant correlations (P < 0.01) between the youth assets subscales and the item measuring level of sexual activity indicate a relationship between behaviors, attitudes and assets that had been suggested by past research on adolescent behaviors (Jessor and Jessor, 1977; Leffert, 1998). Therefore, youth with more assets and more protective attitudes about sex, engage in fewer risk behaviors.
Some limitations of this study are in need of discussion. First, all information collected in this study is self-reported and although multiple procedures were used to ensure confidentiality, it is possible that the bias of providing socially desirable answers is present. Two other limitations, which have been previously mentioned, are the need for further exploration of the influence of item wording and response options on factor analysis results, and the need for further development of survey items labeled as ‘content clusters’ or that fall in subscales with low reliability so as to enhance psychometric soundness. While the large sample size and the large number of items included in the analysis present limitations, the procedure of randomly selecting two subsamples and running the analysis on these samples helps to provide stronger evidence that the factor pattern that emerged was truly representative of the underlying structure. Finally, another limitation of this study is the lack of examination of construct validity. The survey results were not compared to measures of actual behavior. Additionally, the instrument was not validated against other established instruments so as to examine convergent and divergent validity. In addition, future studies may want to examine the stability of the constructs measured with this instrument using test–retest analysis. Therefore, caution should be taken when using this instrument in the future, especially until further studies can provide additional evidence of appropriate validity and reliability of the various subscales.
At this point, even though additional psychometric work is needed, the AHABS can provide program practitioners and evaluators a promising tool with primarily acceptable psychometric properties and a range of items covering risks, attitudes and assets. This instrument contributes to the current need for psychometrically sound instrumentation measuring not only the risk behaviors of adolescents, but also their attitudes and assets. Additionally, after additional developmental work has been completed, the tool should be useful to researchers/evaluators trying to work in school settings because school administrators tend to be more comfortable asking students about assets than risks. Finally, the AHABS could prove useful in program development and evaluation because it is based on a holistic approach to youth programming. Specifically, the tool examines individual factors such as knowledge, environmental factors such as peer norms and presence of other adults, and behavioral factors such as risk taking and constructive use of time. All of these factors reciprocally affect the health of youth and thus serve as important intervention points. Just as youth intervention programs continue to recognize and intervene across multiple influences in a youth’s life, evaluation tools must also become increasingly comprehensive in their measurement of a program’s impact. The AHABS provides one example of a youth questionnaire striving to soundly measure multiple factors influencing youth today.
|
Acknowledgements |
|---|
This research was conducted as a component of a statewide adolescent pregnancy prevention initiative evaluation funded through 1998 South Carolina legislation. AHABS survey items 48–57, 59–63, 67, 69–71, 73, 85, 86 and 88–93 were taken from the Search Institute Profiles of Student Life: Attitudes and Behaviors (© 1996 Search Institute, Minneapolis, MN) and used by permission. The authors would also like to acknowledge Dr Ruth Saunders (Norman J. Arnold School of Public Health, University of South Carolina) for her guidance and assistance with analysis for this manuscript. Lastly, the authors would like to acknowledge the thoughtful comments of the blind reviewers in developing this manuscript.
|
|
References |
|---|
|
TopAbstractIntroductionMethodsDescription of instrumentResultsDiscussionReferences |
|---|
Baker, S.A., Morrison, D.M., Gillmore, M.R. and Schock, M.D. (1995) Sexual behaviors, substance use, and condom use in a sexually transmitted disease clinic. Journal of Sex Research, 32, 37–44.
Bandura, A. (1986) Social Foundations of Thought and Action: A Social Cognitive Theory. Prentice-Hall, Englewood Cliffs, NJ.
Barton, W.H., Watkins, M. and Jarjoura, R. (1997) Youth and communities: toward comprehensive strategies for youth development. Social Work, 42, 405–419.
Benard, B. (1991) Fostering Resiliency in Kids: Protective Factors in the Family, School, and Community. Western Regional Center for Drug-Free Schools and Communities, Far West Laboratory, San Francisco, CA.
Benson, P.L., Leffert, N., Scales, P.C. and Blythe, D.A. (1998) Beyond the village rhetoric: Creating healthy communities for children and adolescents. Applied Developmental Science, 2, 138–159. [CrossRef]
Brener, N., Collins, J., Kann, L., Warren, C.W., Williams, B.I. (1995) Reliability of the Youth Risk Behavior Survey Questionnaire. American Journal of Epidemiology, 141 96), 575–580.
Cattell, R.B. (1965a) Factor analysis: an introduction to essentials. I. The purpose and underlying models. Biometrics, 21, 190–215.[CrossRef][Web of Science]
Cattell, R.B. (1965b) Factor analysis: an introduction to essentials. II. The role of actor analysis in research. Biometrics, 21, 405–435.[CrossRef][Web of Science][Medline]
Coker, A., Richter, D., Valois, R., McKeown, R., Garrion, C. and Vincent, M. (1994) Correlates and consequences of early initiation of sexual intercourse. Journal of School Health, 64, 372–377.
Elliot, D.S., Huizinga, D. and Menard, S. (1989) Multiple Problem Youth: Delinquency, Substance Use, and Mental Health Problems. Springer, New York.
Fullerton, C.S. and Ursano, R.J. (1994) Preadolescent peer friendships: a critical contribution to adult social relatedness. Journal of Youth and Adolescence, 23, 43–64.
Garrison, C.Z., McKeown, R.E., Valois, R.F. and Vincent, M.L. (1993) Aggression, substance use and suicidal behavior among high school students. American Journal of Public Health, 83, 1116–1120.
Green, M.B. (1998) Youth violence in the city: the role of educational interventions. Health Education and Behavior, 25, 175–193.
Hatcher, L. (1996) A Step-by-Step Approach to Using the SAS System for Factor Analysis and Structural Equation Modeling, 2nd edn. SAS Institute, Cary, NC.
Hawkins, J.D., Catalano, R.F. and Miller, J.Y. (1992) Risk and protective factors for alcohol and other drug problems in adolescence and early childhood: Implications for substance abuse prevention. Psychology Bulletin, 112, 64–105. [CrossRef]
Hoffman, L., Paris, S. and Hall, E. (1994) Developmental Psychology Today, 6th edn. McGraw-Hill, New York.
Jessor, R., Bos, J.V.D., Vanderryn, J., Costa, F.M. and Turbin, M.S. (1995) Protective factors in adolescent problem behavior: moderator effects and developmental change. Development Psychology, 31, 923–933. [CrossRef]
Jessor, R. and Jessor, S. (1977) Problem Behavior and Psychological Development: A Longitudinal Study of Youth. Academic Press. San Diego, CA.
Kim, J. and Mueller, C.W. (1978) Factor Analysis: Statistical Methods and Practical Issues, vol. 14. Sage, Newbury Park, CA.
Kirby, D. (1999) Reflections on two decades of research on teen sexual behavior and pregnancy. Journal of School Health, 69, 89–94.
Kolbe, L.J. (1990) An epidemiological surveillance system to monitor the prevalence of youth behaviors that affect health. Journal of Health Education, 21, 44–48.
Kolbe, L.J., Kann, L. and Collins, J.L. (1993) Overview of the youth risk surveillance system. Public Health Reports, 108 (Suppl. 1), 2–10. [Web of Science][Medline]
Leffert, N., Benson, P.L., Scales, P.C., Sharma, A.R., Drake, D.R. and Blyth, D.A. (1998) Developmental Assets: measurement and prediction of risk behaviors among adolescents. Applied Developmental Science, 2, 209–230.
Lerner, R. (1980) Adolescent Development. McGraw-Hill, New York.
Locke, S.E. and Vincent, M.L. (1995) Sexual decision-making among rural adolescent females. Health Values, 19, 47–58.
Montano, D.E., Kasprzyk, D. and Taplin, S.H. (1996) The Theory of Reasoned Action and The Theory of Planned Behavior. In Glanz, K., Lewis, F.M. and Rimer, B. (eds), Health Behavior and Health Education: Theory, Research and Practice, 2nd edn. Jossey-Bass, San Francisco, CA, pp. 85–112.
Pittman, K. and Wright, M. (1991) Bridging the Gap: A Rationale for Enhancing the Role of Community Organizations in Promoting Youth Development. Academy for Educational Development, Washington, DC.
Rak, C.F. and Patterson, L.E. (1996) Promoting resilience in at-risk children. Journal of Counseling and Development, 74, 368–373. [Web of Science]
Richter, D.L., Valois, R.F., McKeown, R.E. and Vincent, M.L. (1993) Correlates of condom use and number of sexual partners among high school adolescents. Journal of School Health, 63, 91–96.
Roth, J., Brooks-Gunn, J., Murray, L. and Foster, W. (1998) Promoting healthy adolescents: synthesis of youth development program evaluation. Journal of Adolescent Research, 8, 423–459.
Saunders, R.P., Pate, R.R., Felton, G., Dowda, M., Weinrich, M.C., Ward, D.S., Parsons, M.A. and Baranowski, T. (1997) Development of questionnaires to measure psychosocial influences on children’s physical activity. Preventive Medicine, 26, 241–247. [CrossRef][Web of Science][Medline]
Scales, P.C. and Leffert, N. (1999) Developmental Assets: A Synthesis of the Scientific Research on Adolescent Development. Search Institute, Minneapolis, MN.
Valois, R.F. and Fraser, J.G. (1997) The 1997 Carolina youth Risk Behavior Survey and Summary Report. South Carolina Department of Education and the University of South Carolina School of Public Health, Columbia, SC.
Valois, R.F., McKeown, R.E., Garrison, C.Z. and Vincent, M.L. (1995) Correlates of aggressive and violent behaviors among public high school adolescents. Journal of Adolescent Health, 16, 26–34. [CrossRef][Web of Science][Medline]
Valois, R.F., Oeltmann, J.E., Waller, J. and Hussey, J.R. (1999) Relationship between number of sexual intercourse partners and selected health risk behaviors among public school adolescents. Journal of Adolescent Health, 25, 328–335. [CrossRef][Web of Science][Medline]
Windsor, R., Baranowski, T., Clark, N. and Cutter, G., (1994) Evaluation of Health Promotion Education and Disease Prevention Programs, 2nd edn. Mayfield, Mountain View, CA.
Zabin, L.S., Hardy, J.B., Smith, E.A. and Hirsche, M.B. (1986) Substance use and its relation to sexual activity among inner-city adolescents. Journal of Adolescent Health Care, 7, 320–331. [CrossRef][Medline]
Received on August 14, 2001; accepted on June 11, 2002
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  K. J. Zullig, R. M. Ward, K. A. King, J. M. Patton, and K. A. Murray Testing the Feasibility of Developmental Asset Measures on College Students to Guide Health Promotion Efforts Assessment, March 1, 2009; 16(1): 31 - 42. [Abstract] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||