Social support for adolescents at risk of school failure.

Richman, Jack M. ; Rosenfeld, Lawrence B. ; Bowen, Gary L. 等

The positive relationship between social support and an individual's physical and mental well-being (Ganster & Victor, 1988; Hardy, Richman, & Rosenfeld, 1991) has provided the impetus for a great deal of research on the clinical utility of social support for individuals and groups. For example, support has been used for purposes such as decreasing morbidity (House & Kahn, 1985), reducing stress (Cutrona & Suhr, 1994; Richman & Rosenfeld, 1987) and feelings of loss (Hobfoll & Stephens, 1990), combating burnout (Etzion, 1984; Pines, Aronson, & Kafry, 1981), increasing feelings of well-being (Ganster & Victor, 1988), increasing job performance and work innovation (Albrecht & Hall, 1991), improving performance on academic examinations (Goldsmith & Albrecht, 1993; Sarason & Sarason, 1986), reducing loneliness (Jones & Moore, 1987), and providing information and support for rural residents with AIDS (Rounds, Galinsky, & Stevens, 1991).

Social support also has been widely studied as a variable specifically designed to promote the development and adaptation of children and adolescents; for example, support has been indicated in research as useful for working with adolescent depression (Barrera & Garrison-Jones, 1992), improving academic and behavioral adjustment (Dubow, Tisak, Causey, & Hryshko, 1991; Ford & Sutphen, 1996), supporting high-risk youths and their families (Tracy, Whittaker, Boylan, Neitman, & Overstreet, 1995), and reducing delinquent behaviors that correlate highly with poor school performance (Zigler, Taussig, & Black, 1992). Furthermore, the literature on risk and protective factors and educational resilience clearly endorses the primacy of the supportive role provided by the family, the peer group, the school, and the community in predicting positive outcomes for students (Benard, 1991; Bogenschneider, 1996; Richman & Bowen, 1997; Wang, Haertel, & Walberg, 1994).

Social support is often less present in the lives of children and youths who are at risk of school failure (Coie et al., 1993; Richman & Bowen, 1997). The purpose of the present study was to explore how the self-perceived social support of adolescents at risk of school failure varied by the type and provider of support and to examine the effects of particular types of social support on school performance outcomes, such as attendance, grades, time studying, and school self-efficacy. By understanding provider networks, students' support patterns, and the effect of support on school performance outcomes, implications may be drawn for the use of social support as an intervention strategy for children and youths at risk of school failure.

Social support is a multidimensional concept that needs to be defined and measured accordingly (Milardo, 1992; Norbeck, Lindsey, & Carrieri, 1981; Sarason, Levine, Basham, & Sarason, 1983; Streeter & Franklin, 1992). Each of three broad types of social support - tangible, informational, and emotional (Cobb, 1976; House, 1981) - are communicated by support providers when they behave in ways that are perceived by recipients as enhancing the recipients' well-being (see Shumaker & Brownell, 1984). These perceptions of others' communication behaviors may take eight distinguishable forms:

1. listening support - the perception that an other is listening without giving advice or being judgmental

2. emotional support - the perception that an other is providing comfort and caring and indicating that she or he is on the support recipient's side

3. emotional challenge - the perception that an other is challenging the support recipient to evaluate his or her attitudes, values, and feelings

4. reality confirmation support - the perception that an other, who is similar to and who see things the same way the support recipient does, is helping to confirm the support recipient's perspective of the world

5. task appreciation support - the perception that an other is acknowledging the support recipient's efforts and is expressing appreciation for the work she or he does

6. task challenge support - the perception that an other is challenging the support recipient's way of thinking about a task or an activity in order to stretch, motivate and lead the support recipient to greater creativity, excitement, and involvement

7. tangible assistance support - the perception that an other is providing the support recipient with either financial assistance, products, and/or gifts

8. personal assistance support - the perception that an other is providing services or help, such as running an errand or driving the support recipient somewhere (Richman, Rosenfeld, & Hardy, 1993).

The eight types of social support are useful for defining and explaining the interaction between individuals and groups as social support is given and received in the environmental context. This multidimensional conceptualization provides a perspective of support that practitioners and clients may find useful in assessing social support and in planning appropriate intervention strategies in school settings.

Many students experience difficulty adjusting to school and acquiring the social and academic skills necessary for pursuing advanced education and training. As a result, their opportunity for functioning successfully as adults in roles associated with work and family is jeopardized. With limited means and opportunities to achieve self-sufficiency through employment, many of these young adults are unable either to participate meaningfully in society or to find personal satisfaction and purpose. The future may be even more dismal for some; for example, school failure has been associated with higher mortality rates, incidence of suicide, and admissions to state mental hospitals (Brenner, 1976; Rumberger, 1987).

The impact of school failure has consequences for society as well as for the individual, including a waste of human capital, loss of national income, loss of tax revenues, higher risk of sexually transmitted disease, increased use and demand for social services, increased crime, reduced political participation, and higher health care costs (Carnahan, 1994; Santelli & Beilenson, 1992). In addition to these concerns, business leaders have noted that many students either graduate or leave before graduation without the basic competencies to perform even rudimentary tasks in industry (Slavin, Karweit, & Madden, 1988). Supporting students and enhancing academic success are critical steps toward promoting more competent adult role performance. These aims have important implications for the individual, the family system, the economy, and the general well-being of society.

The concept of resilience has been introduced into the literature as part of the search for protective factors that buffer the potential effects of risks and promote the positive adaptation of individuals (Garmezy, 1993; Wang et al., 1994; Werner & Smith, 1982). Why are some children more resilient than others when confronted with similar stressors and risk factors? Social support - provided by the family, the peer group, the school, and the community - has been identified as an important characteristic of resilient children (Benard, t 991; Bogenschneider, 1996; Coie et al., 1993; Wang et al., 1994). Because of its association with resilience, social support is likely to promote academic resilience and positive educational outcomes for at-risk students.

Feeling socially supported in one's environmental contexts seems to encompass a large part of building academic resilience and encouraging positive educational outcomes for at risk students. Social support, in all likelihood, serves two major functions: It contributes to adjustment and development, and it provides a buffer against stress (Clark, 1991). Two research questions, based on known and surmised relationships among social support, resilience, and positive school outcomes, guided the present investigation: (1) Who do at-risk middle and high school students perceive as their providers of each of the eight types of social support? (2) What are the school outcomes associated with the receipt of each of the eight types of social support for at-risk middle and high school students?

Method

Respondents

Respondents were 808 students enrolled in Communities In Schools (CIS) programs in 17 middle schools and high schools in North Carolina and Florida. Of the 808 students, 525 (65 percent) completed usable surveys; the bulk of the remaining 35 percent either failed to return signed parent permission forms or failed to be present the day of testing. All students who returned signed parent permission forms and were present on the day of testing completed the survey.

The CIS program, having more than 137 programs in 30 states, is the largest U.S. public-private partnership in the nation for dropout prevention. Spanning the school years from kindergarten to high school, CIS serves more than 300,000 students. Students who participate in CIS have been identified by school officials or human services professionals as at risk of school failure. Many of the students come from lower socioeconomic households, racial or ethnic minority groups, and single-parent families-profile characteristics that are highly confounded (Rumberger, 1987).

Using a school-based and family-based approach to intervention, CIS has as its goal to mobilize and integrate community services and supports for both the at-risk youths and their families. Each local CIS organization develops a program that follows the guidelines provided by the National CIS office and works in partnership with the local social services, business, and educational communities to develop a total academic and support program that is responsive to the resources and direction provided by the local community. CIS facilitates the receipt of necessary support and aid for clients either by directly providing necessary services, by coordinating with other providers (for example, public health services, social services, mental health services, public school personnel, or volunteers from local businesses) or by referring clients to community agencies for service. Specially designed educational programs, tutoring, mentoring, health screening, and family support almost always are a part of the core CIS plan for working with students at risk of school failure.

For the purposes of this investigation, respondents were divided into two groups: middle school students (grades 6 to 8, n = 296) and high school students (grades 9 to 12, n = 229). The National Governor's Association (1989) concluded in its report that 6th, 7th, and 8th grade children have unique developmental needs; developmentally, middle-school-age children are not just smaller versions of high school-age adolescents. Because of differences in psychosocial development, middle school and high school students may identify a different set of support providers, perceive their support environments somewhat differently, and, therefore, have distinct school performance outcomes.

Demographically, the two groups were quite similar with respect to their racial or ethnic composition (that is, predominantly African American and non-Hispanic white), and family structure (that is, predominantly two-parent and single-parent configurations) (Table 1). The two groups differed, however, with respect to the percentages receiving free school lunch: whereas 70.1 percent of middle school students received free lunches, fewer than one-half (41.2 percent) in high school received free lunch [[[Chi].sup.2](1, N = 525) = 61.7, p [less than] .001]. According to Charles Heise, Child Nutrition Division, Food and Consumer Service, U.S. Department of Agriculture (personal communication, March 23, 1996), federal guidelines mandate that students are eligible for a subsidized school lunch when their family income is below 180 percent of the poverty level.

Table 1

Characteristics of Middle and High School Students at Risk of
School Failure

 Grades Grades
 6-8 (%) 9-12 (%)
Characteristic (n = 296) (n = 229)

Gender
Female 49.3 48.4
Male 50.7 51.6

Racial or ethnic group
Non-Hispanic white 39.3 38.1
African American 49.3 56.2
Multiracial 3.4 1.8
Asian/Pacific Islander 2.1 0
Other 5.9 3.9

Family structure
Single parent 33.9 36.1
Two parent 54.1 53.7
Other 12.0 10.2

Free school lunch
No 26.5 54.8
Yes 70.1 41.2
Unknown 3.4 4.0
Age (mean) 12.2 15.6

Measures

Measures used to assess whether a student received each of the eight types of social support, who the support providers are, and the school outcomes of support, are in a comprehensive instrument, the School Success Profile (SSP). During fall 1995 CIS students completed the SSP as part of a larger study assessing the instrument's validity and reliability (Bowen & Chapman, 1996; Richman & Bowen, 1997). The SSP is a self-report survey questionnaire that is based on an ecological approach that assesses several dimensions of students' four most important microsystems (Bronfenbrenner, 1979; Garbarino, 1992, 1995). Specifically, 92 variables are measured with 265 close-ended items that assess students' individual adaptation in four areas of their social environment - neighborhood, school, friends, and family - and a fifth area that assesses self-perceptions of their health and well-being. For example, the section devoted to "neighborhood" contains items related to students' perceptions of neighborhood safety and satisfaction; the section devoted to "school" contains items related to grades, school self-efficacy, and teacher support; the section devoted to "friends" contains items related to how well the student gets along with her or his friends, and the extent to which he or she trusts them; and the section devoted to "family" contains items related to adult caretaker support and family interaction. The "health and well-being" section of the SSP assesses students' feelings of confidence, self-image, and self-esteem, and whether they receive various types of social support.

The SSP is designed to measure student perceptions. From a social work perspective, a student's perceptions may represent the most important aspect of reality. For example, students who do not perceive their attendance or grades as a problem are unlikely to participate in an intervention designed to positively affect their actual attendance or grades. A potential problem with self-report data, however, is social desirability bias. Rubin and Babbie (1997) cautioned researchers to be careful because "whenever you ask people for information, they answer through a filter of what will make them look good" (p. 163). For this study, the SSP asked students to report on their school outcomes and behavior - questions that may elicit socially desirable responses. The bias, however, is easily dealt with (when it occurs) by the practitioner who has access to school archival data on grades, attendance, and behavior problems. Where a discrepancy exists between a student's self-report and documented school outcomes, the practitioner can use this information when planning interventions in concert with the student.

Confirmation for the use of student self-report as the basis for data collection is available from several sources. For example, Garbarino, Stott, and Faculty of the Erikson Institute (1989) argued that "school age children are frequently the best informants regarding their own behavior and feelings, particularly as they get older" (p. 37). Edelbrock and his colleagues (1985) suggested that as a child's age increases (10 and over), the reliability of his or her self-report "appear[s] reliable enough to serve as descriptive aids in clinical research and epidemiology" (p. 223). Also, self-reported grades have been supported in prior research as a valid measure of academic performance (Dornbusch, Ritter, Leiderman, Roberts, & Fraleigh, 1987; Paulson, 1996).

Each of the SSP items and scales received extensive field testing to ensure content validity, including reviews by child development experts, school officials and teachers, social intervention experts in the schools, and CIS student representatives. Because of the low reading levels among many students at risk of school failure, special attention was given to both the wording of items and the ease with which the response format could be used. Items and measures have shown good concurrent validity in prior analyses (Bowen & Chapman, 1996).

An earlier field test of the SSP with 87 at-risk middle and high school students demonstrated a significant level of convergence between students' reports of performance and behavior at school and information available from school records. The kinds of grades the students reported on their most recent report card, as well as the number of Ds and Fs they reported receiving, were moderately to highly correlated with their average math and English grades for the academic year reported on their student records (correlations ranged from .42 to .66, which, according to Rubin & Babbie, 1997, are between medium and high effect sizes). Students who reported lower grades also were more likely than those reporting higher grades to be coded by school officials as a retention risk in their record at the end of the academic year. In addition, the number of absences that were officially recorded in the students' records were moderately correlated with their self-reports of cutting school or missing school because of illness during the last seven days (r = .33 and .28, respectively, both medium effect sizes) (Rubin & Babbie, 1997). Last, students' reports of school suspension or expulsion during the past 30 days were moderately to highly correlated with both the number of disciplinary actions and the number of suspensions that they received during the academic year (including the previous 30 days), as reflected in the school records (r = .47 and .37, respectively, both medium effect sizes).

Social Support. Students' receipt of social support was assessed with eight SSP items, each referencing a different type of social support (Richman & Rosenfeld, 1987; Richman et al., 1993; Rosenfeld, Richman, & Hardy, 1989). Each item was evaluated with a dichotomous no-yes response format. Each student was asked, Are there people you talk to at least weekly,

* who listen to you without giving you advice or judging you - listening support

* who tell you that they appreciate your efforts-technical appreciation support

* who encourage you to do well - technical challenge support

* who comfort you and tell you that they are on your side - emotional support

* who get you to think about your values and feelings - emotional challenge support

* who are similar to you and see things the way you do - reality confirmation support

* who help you by giving or loaning you money - tangible support

* who provide you with help, such as giving you a ride somewhere or helping you with your homework - personal support

Items are based on the Social Support Survey (SSS), developed by Richman et al. (1993).

Correlations among the eight social support items were positive; however, they were generally modest, indicating their relative independence. The range was from. 16 (listening support and tangible support) to .53 (technical challenge support and emotional challenge support), with most of the correlations falling between .21 and .34. The average correlation was .32, indicating slightly more than an average of 10 percent common variance between pairs of items.

Social Support Providers. In addition to items measuring the eight social support types, SSP items also assess four potential sources of support: (1) parents and adult caretakers, (2) teachers, (3) friends, and (4) neighbors. Support from parents, referred to as "the adults in your home," was measured by 20 summative items presented with a three-point response continuum: 1 = never, 2 = sometimes, and 3 = often (for example, during the past 30 days, the adults in your home "made you feel appreciated"). Teacher support was measured by nine true-false summative statements that describe various feelings about teachers at their school (for example, "My teachers really care about me"). Friend support was assessed by five true-false summative items that describe the level of trust and closeness students feel with their peers (for example, "I am satisfied that I can turn to my friends when something is bothering me"). Neighbor support was measured by four agree-disagree items that assess students' perceptions of the cohesion and encouragement of people in their neighborhood, such as, "If I had a problem, there are neighbors who could help me" (Small & Kerns, 1993). Each measure was computed to range from low to high support.

Alpha reliability for the parent support measure was .97, for teacher support it was .84, for friend support it was .81, and for neighbor support, the measure with the fewest items, it was .69. Correlations among the four provider measures of social support were moderately positive. They ranged from .25 (teacher support and parent support) to .31 (teacher support and neighbor support), with an average of .27. Variance in common among the four measures, therefore, was low, ranging from approximately 6.0 percent to 9.6 percent, with an average of 7.0 percent.

School Performance and Outcome Measures. SSP items of primary interest in this investigation concerned students' satisfaction with school and their performance and behavior as students. Specifically, data were collected on the following variables: attendance (five items, for example, "cut at least one class" during the previous seven days); avoidance of problem behavior (five items, for example, during the past 30 days, "I was sent out of class because I misbehaved"); grades (one item concerning the percentage of As and Bs on the last report card); prosocial behavior (two items, for example, during the past 30 days "I was helpful to a teacher or another student"); school satisfaction (nine items, for example, "I enjoy going to this school"); school self-efficacy (five items, for example, how difficult would it be to "have your ideas listened to in class"?); school sense of coherence (nine items, for example, "I often feel mixed up and confused while at school"); and time spent studying (one item referencing the number of hours studying each school night).

Alpha reliability for each of the multi-item measures was: attendance, .58; avoidance of problem behavior, .66; prosocial behavior, .49; school satisfaction, .65; school self-efficacy, .61; and school sense of coherence, .68. Correlations among the eight school outcome measures indicated their relative independence. They ranged (disregarding sign) from .02 (attendance and prosocial behavior) to .47 (school satisfaction and school sense of coherence), with half the correlations falling below .2, and only 3 above .35. The average correlation was .20.

Data Analysis

To respond to the two research questions, 16 discriminant analyses (DAs) were conducted for each group of students, middle and high school. Discriminant analysis is a statistical procedure that separates two or more groups and provides for differentiating the groups on the basis of relations with several discriminating variables (Williams, 1991). It is most useful when there is a relatively large set of potential discriminating variables (for example, grades, attendance, study habits) that are best considered simultaneously because of their possible interdependence. DA provides a linear combination of the variables that best distinguishes the groups (for example, students receiving listening support versus those not receiving this support). Also, by considering all potential discriminating variables simultaneously, DA is a parsimonious procedure that reduces the high Type I error associated with conducting multiple tests.

First, to determine sources of support for each of the eight types of social support for middle school students, eight DAs were conducted. For each type of support, students were divided into two groups: those who indicated receiving the support and those who indicated not receiving the support. Then, the four sources of support - parents and adult caretakers, teachers, friends, and neighbors - were entered into each DA as potential discriminating variables.

Second, to determine school outcomes and behavior of middle school students receiving-not receiving each of the eight social support types, eight additional DAs were conducted. Again, for each type of support, students were divided into two groups: those who indicated receiving the support and those who indicated not receiving the support. Then the eight school outcome and behavior variables - attendance, avoidance of problem behavior, grades, prosocial behavior, school satisfaction, school self-efficacy, school sense of coherence, and time spent studying - were entered into each DA as potential discriminating variables. This process was repeated for the high school students, resulting in a total of 32 DAs conducted.

Results

Social Support Providers

Who are the primary sources of the eight types of support for middle and high school students? Table 2 contains the best combination of variable(s) - and their standardized weights - in the discriminant function that separates the two groups of students, the group centroids (means useful for interpreting the discriminant function weights), the percentage of "hits" and "misses" resulting from the classification of each student using the discriminant function (which provides an indication of the usefulness of the function), and the effect size, w, of the resulting chi-square (which provides an indication of the importance of the results). The Wilks's Lambda associated with each canonical discriminant function was statistically significant, with four p values less than .05, and the remaining 12 less than .007.

Middle school students have their friends as their primary source of listening support, whereas for high school students, it is their parents or adult caretakers. Technical appreciation support is provided by friends for both middle and high school students; however, parents are an additional source for middle school students, [TABULAR DATA FOR TABLE 2 OMITTED] and parents and teachers are both providers of this support for high school students.

Parents and adult caretakers are the major source of technical challenge support and emotional challenge support for both middle and high school students. They are also the major source of emotional support for both groups, with friends serving as a second source for middle school students, and teachers serving as a second source for high school students. Similarly, parents and adult caretakers are the major source of reality confirmation support for both groups, with teachers serving as a second source for middle school students. Parents and adult caretakers also provide both middle and high school students with personal assistance support, with teachers serving as a second source of this support for high school students.

Neighbor support is perceived as the major source of tangible assistance support for middle school students, whereas parents and friends provide this support for high school students.

For middle school students, five of the effect sizes were between medium and large, one was large, and two were between small and medium. For high school students, four of the effect sizes were between medium and large, three were medium, and one was between small and medium. These results indicate that, overall, the DAs were highly successful in their discriminations.

School Outcomes and Social Support

What are the school outcomes for middle and high school students of receiving-not receiving the eight types of support? The variables making up each discriminant function, the group centroids (means), the percentage of "hits" and "misses" resulting from the classification of each student using the discriminant function, and the effect size, w, of the resulting chi-square are presented in Table 3. The Wilks's Lambda associated with each canonical discriminant function reported in Table 3 was statistically significant, with five p values less than .05, and the remaining eight less than .005 (three results were not statistically significant). Interpretation of the results of each significant DA is based on an examination of each DA's discriminant function weights in light of the group centroids.

Listening Support. Middle school students who receive listening support have a greater sense of school self-efficacy and higher grades than those not receiving this support. For high school students, no school outcome differences were revealed between students receiving this support and those not receiving this support.

Technical Appreciation Support. Both middle school and high school students receiving technical appreciation support report a greater school sense of coherence than those not receiving this support. In addition, high school students receiving this support report spending more time studying on school nights.

Technical Challenge Support. Middle school students receiving technical challenge support report better attendance than those not receiving this support. High school students receiving this support report greater avoidance of problem behavior.

Emotional Support and Emotional Challenge Support. Middle school students who receive emotional support and emotional challenge support report being more satisfied with school than those not receiving these types of support. High school students receiving these types of support report spending more time studying on school nights. Those receiving emotional support also report having greater self-efficacy, while those receiving emotional challenge support also report having better attendance.

Reality Confirmation Support. Middle school students who receive reality confirmation support report being more satisfied with school and having greater school self-efficacy than those not receiving this type of support. High school students receiving this support report better grades.

Tangible Assistance Support. For both middle and high school students, no school outcome differences were revealed between students receiving this support and those not receiving this support.

Personal Assistance Support. Middle school students who receive personal assistance support report engaging in more prosocial behaviors than those not receiving this type of support. High school students receiving this type of support report spending more time studying on school nights.

For middle school students, three of the effect sizes were between medium and large, two were approximately medium, and two were between small and medium. (One DA result was nonsignificant.) For high school students, two of the effects sizes were between medium and large, and four were approximately medium. (Two DA results were not significant.) These results indicate that, overall, the DAs were highly successful in their discriminations.

Conclusion

Parents and adult caretakers are major sources of social support for both middle and high school CIS students (Table 4). Parents and adult caretakers provided their middle school children with six of the eight types of social support (technical appreciation support, technical challenge support, emotional support, emotional challenge support, reality confirmation support, and personal assistance support). Their friends provided them with three (friends are the sole providers of listening support, and provided, along with parents and adult caretakers, technical appreciation support and emotional support). Parents and adult caretakers provided their high school children with all eight types of social support. Teachers provided three (emotional support and personal assistance support, and, along with parents and friends, technical appreciation support), and friends provided two (tangible support, and, along with parents and teachers, technical appreciation support).

From analyses of school outcomes data from both middle and high school students, we found that particular types of social support were associated with desirable school outcomes. For middle school students, school satisfaction was affected by the receipt of three different [TABULAR DATA FOR TABLE 3 OMITTED] [TABULAR DATA FOR TABLE 4 OMITTED] types of social support - emotional support, emotional challenge support, and reality confirmation support. Other effects were unique to each type of social support; for example, grades were affected by listening support (perhaps because students who feel listened to have a greater sense of educational self-efficacy), prosocial behavior by personal assistance support, and attendance by technical challenge support.

The pattern that emerged from the analyses of school outcomes for the high school students focused on time spent studying: This outcome was affected by four of the eight types of social support - technical appreciation support, emotional support, emotional challenge support, and personal assistance support. Other effects were unique to each type of social support; for example, grades were affected by reality confirmation support, avoidance of problem behavior was affected by technical challenge support, and attendance was affected by emotional challenge support.

In general, results of each analysis in this investigation indicate that the receipt of social support is related to some positive school outcome; in no analysis was the receipt of social support revealed to have negative consequences. Although the popular belief is that support for adolescents comes mainly from their peers, findings of this investigation indicate that with respect to school outcomes, middle and high school at-risk students perceive their parents and teachers as their primary sources of support. This outcome underscores the importance of adults in at-risk students' lives, and provides an argument for practitioners to work with parents and teachers in a continual effort to provide support that enhances school outcomes. This is especially important at a time when adolescents' separation and individuation highlights rejecting adult support.

Implications for Social Work Practice

Social workers interested in families and children, regardless of the delivery system in which they work (for example, school systems, and departments of social services, public health, and mental health) are often working with the same populations toward similar ends. Collaboration among workers in the various delivery systems as they attempt to support their clients is becoming the preferred method in service delivery. The implications of the present investigation are offered with this collaborative schema in mind.

Results of this investigation have ramifications for school social workers interested in the social support process and in augmenting the breadth and diversity of interventions already in use (Benard, 1991; Bogenschneider, 1996; Carnahan, 1994; Wang et al., 1994). Intervention strategies for affecting school outcomes for at-risk middle and high school students that use the social support process require understanding the relationships among social support providers, types of support available from the providers, and the specific school outcomes desired.

Any intervention begins with an assessment of the student and her or his ecological support system. For example, an eco-map or social support instrument, such as the SSS (Richman et al., 1993), may be useful for identifying sources of support in an individual student's environment. Once a school outcome is identified as the target of social work intervention, the social worker may use the data in Table 4 in the intervention planning stage. Of course, moving from results of investigations that are derived from group analyses to planning interventions with the individual student is always problematic. To what extent does the individual student reflect group characteristics? The practitioner has more information to draw on than a single score on a single instrument, and can make a determination about the possible usefulness of the results of this investigation. For example, the middle school social worker who wishes to affect a student's school sense of coherence (that is, help the student feel less confused while in school) and who determines that the student reasonably reflects the group characteristics of students in this investigation, has two potential points of intervention available. The social worker can attempt to enlist the student's adult caretakers or peer group to provide additional or more effective technical appreciation support. In addition, the social worker can work with the student to learn more effective ways to recognize and to access technical appreciation from her or his adult caretakers or peer group. Similarly, the high school social worker who wishes to affect a student's amount of time studying has several options for points of intervention. The social worker can develop strategies to increase or maximize the effectiveness of adult caretakers' technical appreciation support, emotional support, emotional challenge support, and personal assistance support; peer group members' technical appreciation support or teachers' technical appreciation support, emotional support, and personal assistance support.

The specific strategies associated with a particular point of intervention will depend on the circumstances of the individual student, resources available, school and school system guidelines, and the student-practitioner relationship. It may be difficult to create conditions for each individual student that reflect the provider-support-outcome relationships outlined in Table 4. For example, some parents and other adult caretakers may be incapable of or unavailable for providing students with the necessary support - regardless of the potential benefits. Also, it may be difficult to mobilize support from teachers who are overworked or from peers who may be inaccessible. In these instances, other support providers, such as religious leaders, community leaders, school administrators, tutors, and mentors may need to be enlisted.

Overall, social support may be used as an intervention strategy to positively influence both middle and high school students who are at risk of school failure. For examples of strategies that may be used to affect the receipt of specific types of social support, see Richman (1989), Richman and Rosenfeld (1987), and Richman, Hardy, Rosenfeld, and Callanan (1989). Even though some types of support and support providers were found to be less important than others in this particular study, the literature on social support and social support outcomes argues for maximizing support networks and ensuring the receipt of all types of support.

References

Albrecht, T. L., & Hall, B. J. (1991). The role of personal relationships in organizational innovation. Communication Monographs, 58, 273-288.

Barrera, M., & Garrison-Jones, C. (1992). Family and peer social support as specific correlates of depressive symptoms. Journal of Abnormal Clinical Psychology, 20(1), 1-16.

Benard, B. (1991). Fostering resiliency in kids: Protective factors in the family, school, and community (Northwest Regional Educational Laboratory). Portland, OR: Western Center for Drug-Free Schools and Communities.

Bogenschneider, K. (1996). An ecological risk/protective theory for building prevention programs, policies, and community capacity to support youth. Family Relations, 45, 127-138.

Bowen, G. L., & Chapman, M. V. (1996). Poverty, neighborhood danger, social support, and the individual adaptation among "at risk" youth in urban areas. Journal of Family Issues, 17, 641-666.

Brenner, M. H. (1976). Estimating the costs of national economic policy (Study prepared for the Joint Economic Committee, U.S. Congress, 94th Cong., 2nd Sess.). Washington, DC: U.S. Government Printing Office.

Bronfenbrenner, U. (1979). The ecology of human development: Experiments by nature and design. Cambridge, MA: Harvard University Press.

Carnahan, S. (1994). Preventing school failure and dropout. In R. J. Simeonsson (Ed.), Risk, resilience and prevention: Promoting the well-being of all children (pp. 103-123). Baltimore: Paul H. Brookes.

Clark, M. L. (1991). Social identity, peer relations, and academic competence of African-American adolescents. Education and Urban Society, 24, 41-52.

Cobb, S. (1976). Social support as a moderator of life stress. Psychosomatic Medicine, 38, 300-314.

Coie, J. D., Watt, N. F., West, S. G., Hawkins, J. D., Asarnow, J. R., Markham, H. J., Ramey, S. L., Shure, M. B., & Long, B. (1993). A conceptual framework and some directions for a national research program. American Psychologist, 48, 1013-1022.

Cutrona, C. E., & Suhr, J. A. (1994). Social support communication in the context of marriage: An analysis of couples' supportive interactions. In B. R. Burieson, T. L. Albrecht, & I. G. Sarason (Eds.), Communication of social support: Messages, interactions, relationships, and community (pp. 113-135). Thousand Oaks, CA: Sage Publications.

Dornbusch, S. M., Ritter, P. L., Leiderman, P. H., Roberts, D. F., & Fraleigh, M. J. (1987). The relation of parenting style to adolescent school performance. Child Development, 58, 1244-1257.

Dubow, E. F., Tisak, J., Causey, D., & Hryshko, A. (1991). A two-year longitudinal study of stressful life events, social support, and social problem-solving skills: Contributions to children's behavioral and academic adjustment. Child Development, 62, 583-599.

Edelbrock, C., Costello, A. J., Dulcan, M. K., Kalas, R., & Conover, N. C. (1985). Age differences in reliability of the psychiatric interview of the child. Child Development, 56, 265-275.

Etzion, D. (1984). Moderating effects of social support on the stress-burnout relationship. Journal of Applied Psychology, 69, 615-622.

Ford, J., & Sutphen, R. D. (1996). Early intervention to improve attendance in elementary school for at-risk children: A pilot program. Social Work in Education, 18, 95-102.

Ganster, D.C., & Victor, B. (1988). The impact of social support on mental and physical health. British Journal of Medical Psychology, 61, 17-36.

Garbarino, J. (1992). Child and families in the social environment. New York: Aldine de Gruyter.

Garbarino, J. (1995). Raising children in a socially toxic environment. San Francisco: Jossey-Bass.

Garbarino, J., Stott, F. M., & Faculty of the Erikson Institute. (1989). What children can tell us: Eliciting, interpreting, and evaluating information from children. San Francisco: Jossey-Bass.

Garmezy, N. (1993). Children in poverty: Resilience despite risk. Psychiatry, 56, 127-136.

Goldsmith, D., & Albrecht, T. L. (1993). The impact of supportive communication networks on test anxiety and performance. Communication Education, 42, 142-158.

Hardy, C. J., Richman, J. M., & Rosenfeld, L. B. (1991). The role of social support in the life stress/injury relationship. Sport Psychologist, 5, 128-139.

Hobfoll, S. E., & Stevens, M.A.P. (1990). Social support during extreme stress: Consequences and intervention. In B. R. Sarason, I. G. Sarason, & G. R. Pierce (Eds.), Social support: An interactional view (pp. 454-481). New York: John Wiley & Sons.

House, J. (1981). Work stress and social support. Reading, MA: Addison-Wesley.

House, J., & Kahn, R. (1985). Measures and concepts of social support. In S. Cohen & S. L. Syme (Eds.), Social support and health (pp. 83-108). New York: Academic Press.

Jones, W., & Moore, T. (1987). Loneliness and social support. Journal of Social Behavior and Personality, 2(Special issue), 145-156.

Milardo, R. M. (1992). Comparative networks for delineating social networks. Journal of Social and Personal Relationships, 9, 447-461.

National Governor's Association Task Force on Children. (1989). America in transition. Washington, DC: Author.

Norbeck, J., Lindsey, A., & Carrieri, V. (1981). The development of an instrument to measure social support. Nursing Research, 30, 264-269.

Paulson, S. E. (1996). Maternal employment and adolescent achievement revisited: An ecological perspective. Family Relations, 45, 201-208.

Pines, A. M., Aronson, E., & Kafry, D. (1981). Burnout. New York: Free Press.

Richman, J. M. (1989). Social support for hospice teams: Strategies for effective groups. Home Health Care Nurse, 8, 8-10.

Richman, J. M., & Bowen, G. L. (1997). School failure: An ecological-interactional-developmental perspective. In M. W. Fraser (Ed.), Risk and resilience in childhood: An ecological perspective (pp. 95-116). Washington, DC: NASW Press.

Richman, J. M., Hardy, C. J., Rosenfeld, L. B., & Callanan, R. (1989). Strategies for enhancing social support networks in sport: A brainstorming experience. Journal of Applied Sport Psychology, 1, 150-159.

Richman, J. M., & Rosenfeld, L. B. (1987). Stress reduction for hospice workers: A support group model. Hospice Journal, 3, 205-221.

Richman, J. M., Rosenfeld, L. B., & Hardy, C. J. (1993). The Social Support Survey: An initial validation study of a clinical measure and practice model of the social support process. Research on Social Work Practice, 3, 288-311.

Rosenfeld, L. B., Richman, J. M., & Hardy, C. J. (1989). Examining social support networks among athletes: Description and relationship to stress. Sport Psychologist, 3, 23-33.

Rounds, K. A., Galinsky, M. J., & Stevens, L. S. (1991). Linking people with AIDS in rural communities: The telephone group. Social Work, 36, 13-18.

Rubin, A., & Babbie, E. (1997). Research methods for social work (3rd ed.). Pacific Grove, CA: Brooks/Cole.

Rumberger, R. W. (1987). High school dropouts: A review of issues and evidence. Review of Educational Research, 57, 101-121.

Santelli, J. S., & Beilenson, P. (1992). Risk factors for adolescent sexual behavior, fertility, and sexually transmitted diseases. Journal of School Health, 62, 271-279.

Sarason, I., Levine, H., Basham, R., & Sarason, B. (1983). Concomitants of social support: The social support questionnaire. Journal of Personality and Social Psychology, 44, 127-139.

Sarason, I. G., & Sarason, B. R. (1986). Experimentally provided social support. Journal of Personality and Social Psychology, 50, 122-125.

Shumaker, S. A., & Brownell, A. (1984). Toward a theory of social support: Closing conceptual gaps. Journal of Social Issues, 40, 11-36.

Slavin, R. E., Karweit, N. L., & Madden, N. A. (1988). Effective programs for students at risk. Boston: Allyn & Bacon.

Smith, S. A., & Kerns, D. (1993). Unwanted sexual activity during early adolescence: Incidence and risk factors. Journal of Marriage and the Family, 55, 941-952.

Streeter, C., & Franklin, C. (1992). Defining and measuring social support: Guidelines for social work practitioners. Research on Social Work Practice, 2, 81-98.

Tracy, E. M., Whittaker, J. K., Boylan, F., Neitman, P., & Overstreet, E. (1995). Network interventions with high-risk youth and families throughout the continuum of care. In I. M. Schwartz & P. AuClaire (Eds.), Home-based services for troubled children (pp. 55-72). Lincoln: University of Nebraska Press.

Wang, M. C., Haertel, G. D., & Walberg, H. J. (1994). Educational resilience in inner cities. In M. C. Wang & E. W. Gordon (Eds.), Educational resilience in inner-city America: Challenges and prospects (pp. 45-72). Hillsdale, NJ: Lawrence Erlbaum.

Werner, E. E., & Smith, R. S. (1982). Vulnerable but invincible: A longitudinal study of resilient children and youth. New York: McGraw-Hill.

Williams, F. (1991). Reasoning with statistics: How to read quantitative research (4th ed.). Fort Worth, TX: Harcourt Brace Jovanovich.

Zigler, E., Taussig, C., & Black, K. (1992). Early childhood intervention. American Psychologist, 47, 997-1006.

Jack M. Richman, PhD, is associate professor of social work, Lawrence B. Rosenfeld, PhD, is professor of communication studies, and Gary L. Bowen, PhD, ACSW, is Kenan Distinguished Professor of Social Work, University of North Carolina at Chapel Hill, 301 Pittsboro Street, Chapel Hill, NC 27599-3550; e-mail: jrichman@email.unc.edu. The authors thank John Galassi, Professor of Educational Counseling at the University of North Carolina at Chapel Hill, for his helpful comments on this manuscript. This article was prepared with grant support from the BellSouth Foundation (Atlanta) and the Knight Foundation (Miami).