THE RELATIONSHIP OF SOCIAL PROBLEM SOLIVING TO HIGH SCHOOL PREFORMANCE

This research project is undertaken to explore the relationship between social problem solving and the academic and social performance of high school youth. Psychologists have stated that social problem solving is an important component of psychosocial competence. Similarly, both theory and research have linked social problem solving to overt behavioral adjustment to school in populations of young children. In this study, the broader notion that academic and social perfornance in high school will be strongly related to social problem solving was tested. A total sample of 128 male and female students from grades 9 12, from selected classes and guidance groups participated in the study. The site was a suburban high school located in southern New England. The students were administered the survey in small groups in one 50 minute sitting. Social problem solving was operationally defined as / means-ends thinking {an interpersonal cognitive problem solving measure developed by Platt and Spivack). School performance was operationally defined as academic performance {grade point average, number of semesters on honor roll, and post high school training goals) and social performance {a measure of social network developed by Mitchell and extracurricular school activities). Hollingshead 1 s two-factor method was used to measure socioeconomic status. Measures of connunity social network and community involvement {Mitchell's social network measure, comr,unity clubs, and number of jobs) allowed for tlie comparison of school and community social perfornance profiles. The foll owing hypotheses were tested: 1. School performance, . community performance, and problem solving do not vary by the der.iographic variables of grade level, sex, or, socioeconomic status. 2. High school students with more effective social problem solving skills will show higher social and academic performance in school. Results of a series of two-way ANOVA's indicated that the demographic variables of grade level and socioeconomic status differentially affect certain school and community performance indices. Upper class students had more extensively developed school social networks than middle and lower class students. Problem solving ability varied by grade level: sophomores, juniors, and seniors had greater social problem solving scores than freshmen. Further, high problem solvers showed more effective academic and social performance in school and better comunity performance than their counterparts with middle and lower problem solving scores. Multiple regression analyses indicated strong positive relationships between school and community predictor variables and problem solving skills. The results of this study suggest that social probler.i solving is significantly related to school and community perforr.iance. Notions of problem solving as an active coping strategy for environmental mastery are supported. Schools are postulated to play an active role in the facilitation of problem solving skills and developing the overall social skills of youth. Implications of these results for psychological theory, the practice of school psychology, and the directions for future research are discussed.


Abstract
This research project is undertaken to explore the relationship between social problem solving and the academic and social performance of high school youth. Psychologists have stated that social problem solving is an important component of psychosocial competence.
Similarly, both theory and research have linked social problem solving to overt behavioral adjustment to school in populations of young children. In this study, the broader notion that academic and social perfornance in high school will be strongly related to social problem solving was tested.
A total sample of 128 male and female students from grades 9 -12, from selected classes and guidance groups participated in the study.
The site was a suburban high school located in southern New England.
The students were administered the survey in small groups in one 50 minute sitting.
Social problem solving was operationally defined as / means-ends thinking {an interpersonal cognitive problem solving measure developed by Platt and Spivack). School performance was operationally The foll owing hypotheses were tested: 1. School performance, . community performance, and problem solving do not vary by the der.iographic variables of grade level, sex, or, socioeconomic status.
2. High school students with more effective social problem solving skills will show higher social and academic performance in school.
Results of a series of two-way ANOVA's indicated that the demographic variables of grade level and socioeconomic status differentially affect certain school and community performance indices. Upper class students had more extensively developed school social networks than middle and lower class students. Problem solving ability varied by grade level: sophomores, juniors, and seniors had greater social problem solving scores than freshmen. Further, high problem solvers showed more effective academic and social performance in school and better comunity performance than their counterparts with middle and lower problem solving scores. Multiple regression analyses indicated strong positive relationships between school and community predictor variables and problem solving skills.
The results of this study suggest that social probler.i solving is significantly related to school and community perforr.iance. Notions of problem solving as an active coping strategy for environmental mastery are supported. Schools are postulated to play an active role in the facilitation of problem solving skills and developing the overall social skills of youth. Implications of these results for psychological theory, the practice of school psychology, and the directions for future research are discussed.

Acknowledger:1ents
This research project could not have been completed without the invaluable assistance and support of many people. First, I would 1 ike to thank Michael Nover, a fellow graduate student, who aided in the collection of data during the pilot testing and has provided encouragement all along the way. I would also like to thank Clarice Coler:1an who has patiently typed "rough" drafts, prepared nur:1erous tables, and created order out of my scrambled writing. Appreciation is also extended to my brother Kevin who is also a fellow graduate student and Victor Moniz for their assistance in scoring the problem solving data and putting up with long hours of training.
My doctoral connittee, Ors. Vosburgh, Kulberg, Lott, and Bumpus, has been a constant source of support not just during the dissertation, but throughout my doctoral program. Each has contributed in his/her own way, balancing constructive criticism with guidance and encouragement.
I would like to extend a special thank you to Bill Vosburgh for his professional advice and friendship during my five years at URI.
As major professor, Dr. Daniel J. Hurley has set an exar:1ple for me to strive for both in scholarship and professional excellence. He has demanded thorough research and scholarship in this dissertation, but has been there each step of the way to encourage my progress. I thank him for his personal and professional support, providing optimism and injecting humor into our discussions whenever possible.
Finally, I would like to thank my family and r.iy future wife, Nancy,   This research is undertaken to explore the relationships between high school students' performance in academic and social areas and their social problem solving skills. Social problem solving is chosen as the central construct for study because of its link to psychosocial competence. The many theories that discuss social problem solving as an important component of psychological adjustment also conceptualize adjustment as an active coping process, whereby the individual must constantly modify his/her problem solving strategy. This 11 striving 11 for self improvement and adaptation to the environment needs to be examined in terms of its relevance to developmental processes of high school youth and to specific behavior settings. It is the intent of this study to examine how students perform academically and socially in the adolescent phase of their school career; and relate this perfomance to their social problem solving skills.
The importance of high school in the socialization of youth and in the preparation for later conmunity (life) roles continues to be an important research topic. Youths' perfomance in high school and their social problem solving skills, an index of active coping skills, need further detailed examination. This study explores the relationship between students' social problem solving ability and students' social network, involvement in high school activities, and grades.
Comparisons are also made to community performance. Performance in high school is chosen for study because its task demands preview later comunity and job-related task requirements. In this sense, high school experiences are a testing ground for later cor:imunity adjustr.lent. This study first examines the relationship of demographic characteristics to high school performance and social problem solving.
Next, the study explores the relationships between various aspects of high school performance with social problem solving skills. Finally, the study looks at the relationship between cor.u:,unity performance, high school performance, and social problem solving skills.
This dissertation is divided into the following sections. The first section discusses the importance of social problem solving in theories of psychosocial competence. The second section delineates specific social problem solving theories. In particular, Spivack and compliance to the norms of society (Smith, 1959;Tyler, 1978).
Social problem solving skills have recently been identified as an important part of an individual 1 s active coping with everyday 1 ife.
Other social science theorists have considered active coping as part of the overall personality configuration of an individual. A brief review of major personality theories focusing on psychological competence traces the development of the major tenets of social problem solving thinking. From the different schools of thought a similar pattern emerges: active coping and striving toward mastery are identified as integral components of psychosocial competence (see Table 1).
Sense of ego identity; relatively conflict-free ego functioning; notion of positive reality: intrinsic motivation for 11 eff ectance 11 • Adequacy in love, work, play, and interpersonal · relations; efficiency in meeting situational demands; social problem solving process is important.
Self-actualization; active striving for self-improvement; self as a rational coper; self as an active knower.
Fae il ity in the use of personal constructs; effective hypothesis testing of environmental situations.
Social problem solving is a means of active coping; consideration of alternative solutions and consequences of one's behavior. Involvement in environment; exploratory behavior is index of coping; use of social problem solving skills; person-environment interaction.

THEORETICAL PERSPECTIVE
Devel opr:1ental / Social 5 An important attempt to describe and predict psychosocial competence arises out of psychoanalytic theory. The psychoanalytic perspective posits a healthy, well-differentiated ego capable of rational thought and resiliency under stress as the prime indicator of psychosocial competence (Murphy, 1962;Barron, 1952;Freud, 1952;Loevinger;. Loevinger, (1966)  Robert White (1959) and Heinz Hartmann (1951) offer two perspectives from ego psychology on the nature of pSjlchosocial competence. These authors address the individual's inner motivation toward competence and rational behavior. This competency motivation arises out of healthy ego functioning and enables the individual to strive toward relative autonomy from entironmental (physical and social) dependence. Independence of environment means relative stability in the face of stress and frustration (Hartmann, 1951). Hartmann (1951;1964) considers a conflict-free ego capable of resisting stress and producing useful behavior for the individual, indicative of psychosocial cor.ipetence. Further, he suggests that the ego r.iay not always r.iake rational decisions, but the overall utility of the behavior and flexibility of the ego to shift to meet environr.1ental pressures are prime criteria for environmental mastery. Hartmann (1951;1964) and White (1959; think similarly with respect to the ego' s involvement in psychosocial competence. However, the latter has continued to focus on intrinsic motivation and explained its composition and functioning in great detail. White (1959) (Jahoda, 1958).
Ego psychology's theory on psychosocial competence has enabled psychoanalytic theory to depart from classic Freudian notions of man as an irrational creature by nature, with ego functioning motivated by inner conflicts. White (1959) and Hartr.iann (1951) have modified traditional psychoanalytic notions to suggest that the ego strives toward competence, autonomy, and conflict-free functioning.
Consistent with ego psychology's notions of active coping and motivation toward cor.ipetence is Jahoda's theory on environmental r:iastery (Jahoda, 1958). Two themes er.ierge from the literature within the area of environmental mastery: 1) success; and 2) adaptation.
Adaptation, according to Jahoda (1958) has several dimensions including: 1) ability to love; 2) adequacy in love, work, and play; 3) adequacy in interpersonal relations; 4) efficiency in meeting situational requirements; 5) capacity for adaptation and adjustment; and 6) efficiency in social problem solving. Thus, the combination of adaptation to situational demands, social problem solving skills, and success in environmental endeavors seem to be the primary criteria for mastery of environment. This particular conceptualization represents an attempt to further define psychosocial competence by offering a variety of behavioral demensions. Jahoda's (1958) definition of mastery implies that the individual must be aware of situational variables through experience and social learning and that each situation represents a novel set of problems, resources, and solutions. A general trait of psychosocial competency would theoretically enable an individual to adapt more quickly to a novel environment than his/her incompetent peer, but mastery of the environment suggests that experience is the critical variable.
Humanistic theorists, Rogers (1950), Allport (1955), Maslow (1955), Fromr.i (1947 view psychosocial competence as .. striving toward self-improver.ient. These theorists assert that individuals actively strive to grow and become psychologically healthy. This process of active improvement is exemplified in the theory of Allport (1955) where he postulates stages of propriate functioning. Specifically, he states that the 11 self as a rational coper 11 is an active problem solving stage. The final stage of self growth he calls "self as a knower".
Here the individual evaluates his/her performance and reflects (Allport, 1955). Humanistic theories represent the view that individuals are actively striving to master their environment to improve themselves.
As a phenomenological theorist, George Kelly (1955; (Monte, 1977;Bischof, 1970 Behavioral theorists Anderson and Messick (1974), 0jemann (1967), and D'Zurilla and Goldfried (1971 have examined psychosocial competence in terms of adjustnent to various environmental settings. 10 In particular, they postulate that social problem solving is an important behavioral skill that enables the individual to actively cope with his/her environment. These theorists assert that social problem solving requires a number of steps including the recognition of a problem, generation of alternative solutions and knowledge of the consequences of one 1 s actions. The behavioral approach to adjustnent has further developed the social problem solving link to adjustnent. James Kelly, another proponent of the positive mental health movement has developed an ecological theory that departs from traditional views of individual functioning. Kelly 1 s major premise is that individuals seek and actively explore situations that will further their mastery of the surrounding environment (Kelly, 1970;1979a;1979b). This exploration/mastery notion is similar to current notions of social problem solving. Conversely, Hunt (1968) argues that psychosocial competence develops almost exclusively out of environmental experience. Successful experiences increase the probab i1 ity of future successes whereas, negative experiences decrease the chances of success. Both authors agree that a critical component of psychosocial competence is social problem solving. Smith (1968) and Hunt (1968) define social problem solving as the ability to recognize a social problem, generate a number of alternative solutions, and verify the efficacy of the solutions. Both authors assert that this ability is a critical predictor of adjustment to school and other societal institutions.
The development of notions of active coping and environmental mastery have paved the way for specific theories of social problem solving to be delineated. The relationship has been theoretically established between social problem solving and psychosocial competence.
Namely, social problem solving is an important skill in the overall make up of a competent individual and allows the individual to actively cope with his/her environment.
The next section will review specific theories of social problem solving, discuss the range of intervention studies aimed at training social problem solving skills, and examine the empirical evidence 1 inking social problem solving to overt behavioral adjustment.

THEORIES OF SOCIAL PROBLEM SOLVING
Proble□ solving has been widely discussed in the psychological 1 iterature for many years and reviewed periodically (Duncan, 1959;Davis, 1966;Simon & Newell, 1970). For the most part, psychological research in the area of problem solving has dealt with cognitive styles and abilities when an individual is presented with impersonal tasks such as puzzles, anagrams, and syllogisms. Proble□ solving has been viewed by the Gestalt theorists (Asher, 1963;& Sheerer, 1963) as an insightful process that is a basic component of an individual 1 s personality, by inforr.iation processing theorists (Posner, 1965;Sternberg, 1979) who analyze the types of cognitive skills required to solve different problems, and by intelligence theorists (Merrifield, Guilford, Christensen, and Frick, 1962) who view the skin as an essential intellectual ability indicative of general intelligence.
There are distinct theories of social problem solving that delineate the problem solving process (see Table .2). These theories will be summarized in this section of the literature review.   (1968) conceptualizes probler:t solving as a critical coping strategy that can predict future adjustment to a variety of societal institutions. He describes social probler.i solving as the ability to recognize a problem and formulate viable alternative solutions. Further, he states that the relatively 11 plastic 11 child can be exposed to a variety of experiences that will increase problem solving abilities and increase his/her control over his/her behavior. Similarly, M. Brewster Smith (1968) describes social problem solving as an active solution-generation process by which individuals cope with typical social problems.
They will be reviewed briefly.
Causal thinking. Ojeraann (1955;1967) examined the relationship between causal thinking in social problems to classroom adjustment in young children. He was able to train young children to recognize the underlying dynamics of a variety of classroom problem situations and demonstrate a significant improvement in classroom adjustment as a result •. Ojemann asserts that social problem solving is a teachable skill that allows children to understand the meaning behind surface behavior and apply the techniques to everyday situations.
Through his research, Ojemann et al., (1955) was able to empirically verify the relationship between social problem solving and overt classroom adjustment in young children. Limitations in his efforts included use of a li □ ited number of problems, short-term training, teacher ratings of behavioral adjustment, and al imited age range of subjects.

17
Environmental mastery. As previously alluded to, Marie Jahoda (1953; has stressed the importance of social problem solving as a critical indicator of psychosocial competence. Jahoda 1 s conceptual framework is central to the present research in that Spivack and Shure (1974;1976) have integrated her model into their theory and research. Jahoda (1953) discusses two perspectives on social problem solving.
She speaks of problem solving as an ability to reach an end product successfully. But, success in social problem solving is often dependent upon circumstances beyond the individual 1 s control. Hence, this circumstantial determinant renders social problem solving as a somewhat ambiguous attribute of mental health. Jahoda asserts that conscious awareness, long-term or short-term, of a social problem, understanding the means to reach a solution, and the intention to act on the problem are more important attributes of problem solving than the success of the solution or the number of solutions generated. Jahoda (1953) provides further insight into the process of social problem solving by distinguishing between three dimensions: 1) awareness of a social problem; 2) consideration of the relevant means toward a solution; and 3) choice of the most efficacious means. This conceptualization of the problem solving process seems to parallel several formal descriptions of the thinking process where the individual must recognize the impending problem, think of steps to solve the problem, and choose the best route to the solution (Duncker, 1945). An important consideration in Jahoda's problem solving theory is delay of gratification (e.g., the ability to consider all the steps in the problem solving process before acting). Jahoda states that the maximum degree of problem solving effectiveness occurs when an individual combines: 1) the tendency to complete all the stages of the problem solving process; 2) maintenance of an appropriate feeling tone that will facilitate the process; and 3) direct attack on the problem.
Jahoda has suggested that research should verify whether people who are mentally healthy are morel ikely to be successful as problem solvers.
Behavioral analysis. D'Zurilla & Goldfried (1971) offer a related theory on problem solving. They begin by defining a problem: The term problem will refer here to a specific situation or set of situations to which a person must respond in order to function effectively in his environment. To point out this situational emphasis (as opposed to the traditional II intrapsychic 11 connotation of the word 11 problem 11 in clinical psychology), the term problenatic situation will be used in most instances in place of 11 problem 11 • In the present context, a situation is considered problematic if no effective response alternative is ir:1r:1ediately available to the individual confronted with the situation. This stage involves recognition of the problem and the realization that actively -------~ pursuing a solution is the best way to resolve it.
2. Problem definition and formulation. Here, the specific aspects of the problem are 11 stated 11 and goals for problem resolution are set.
3. Generation of alternatives. This step involves the formulation of possible solutions in order of their probable successes. The consideration of □any alternatives is stressed.
4. Decision making. Here the consequences of each prospective solution must be weighed and the alternative that comes closest to the original goals should be implemented.

Verification.
This final step involves the assessment of the actual outcome in light of original predictions. This serves as a feedback mechanism for future problem solving attempts. D'Zurilla & Goldfried (1971) have specifically defined interpersonal problem solving as a behavioral process which creates a number of viable response alternatives and increases the 1 ikelihood of selecting the most effective response for dealing with the social problem. This approach to the social problem solving process is similar to the theory offered _by Spivack and Shure (1974). Both theories draw heavily on the original model of problem solving described by Jahoda (1953Jahoda ( , 1958. The important advantage to the theory of Spivack and his co-workers is the emphasis on an interpersonal problem solving skill called means-end thinking. Spivack and Shure's (1974) theory on interpersonal cognitive problem solving will be described next and the relevance of means-ends thinking will be presented and verified empirically.
Interpersonal Cognitive Problem Solving. Little evidence exists to ✓ suggest that solving a pencil and paper maze or puzzle will predict how well an individual will handle an interpersonal conflict.
In their empirical investigation into the relationship between social problen solving and adjustment, Spivack and Shure state that the domains of impersonal and social problem solving are disparate and require distinct and separate theoretical analyses (Spivack & Shure, 1974;Spivack, Platt, & Shure, 1976;Platt, Spivack, Altnan, Altman, & Peizer, 1974). Spivack, Shure, and their co-workers have identified a number of interpersonal cognitive problem solving skills that underlie the problem solving process. Among these abilities, the following have been identified as the most important indicators of overt social adjustment in children when the effect of intelligence is controlled.
l. Alternative thinking is defined as the ability to generate many possible alternative solutions to a specific interpersonal problem situation.
2. Consequential thinking is described as the ability to conceptualize the potential consequences of a particular alternative and to base the problem solving decision on this information.
3. Means-ends thinking is defined as the ability to plan the steps (means) necessary for a particular goal to be reached, to recognize the potential obstacles impeding the probler:i solving process, and to utilize an appropriate tirae framework as a guideline toward meeting the goal (e.g., knowing when to wait before initiating action).
The latter ability has been identified as the most important of the three. The means or ends thinking process is similar to the concept that Jahoda (1953) defined as a more important skill than generating alternatives and reaching a solution.
As the primary exponents of an interpersonal-cognitive probler:i solving approach to children's adjustnent, Spivack and Shure and their collaborators have conducted extensive research in this area. Fron~ their data, they conclude that: 1) some children are significantly more capable than others to think through and solve typical interpersonal problems; 2) effective social problem solvers display significantly better adjusted behavior in school as compared with their more deficient problem solving peers; 3) exposure to interpersonal cognitive problem solving training significantly increases children's ability to solve social problems with peers and adults; and 4) the effects of training these skills occur within a wide intelligence range (70 -120+) and seem to have relatively enduring results.

Spivack and
Shure contend that social problem solving is relatively independent of general intelligence and mere verbal production, as well as socio-~ economic status (Hopper & Kirschenbaum, Note 1). These authors contend that individuals who can think through and solve social problems, judge the valence of potential outcomes, and appreciate the underlying causal dynamics _ of an interpersonal solution will: 1) be less likely to make impulsive mistakes; and 2) suffer less frustration from failure actions leading to a decrease in naladaptive functioning. Further, Spivack and Shure (1974) have found the interpersonal cognitive problem solving process of means-ends thinking to discriminate between aberrant populations and 11 normal 11 populations across many groups (e.g., delinquents, psychiatric patients, drug addicts). Means-ends thinking has also been positively related to development of social networks in groups of adult psychiatric outpatients (Mitchell,Note 11). Patients with better social problem solving abilities (means-ends thinking) tended to have nore effectively developed support networks and manifested better overall social adjustr:lent to their environment (Mitchell,Note 11).
Summary. Each of the authors discussed in this section consider social problem solving as a critical component of an individual's active coping ability. In addition, a particular problem solving process (means-ends thinking) has been proposed and studied as more important than the ability to consider consequences and the generation of r.1any alternatives.
Spivack & Shure have considered this specific problem solving process in their intervention programs for children.
Their research and others have 1 inked means-ends thinking to overt . 24 behavioral adjustment to school and home environments. However, certain areas of the relationship between social problem solving and environmental performance have not been explored. Since the link between problem solving and performance has been established, it is necessary to consider broader notions of environmental performance and different developmental groups in order to further examine the relationship. Perfonnance in any environment requires more than just overt behavioral adjustment. As a sole index of environmental performance, even in school, overt behavioral adjustment is not comprehensive enough to provide a full explanation of what dimensions of the environment a person is adjusting to effectively (e.g., social; participation; task demands). Social problem solving has been studied in aberrant populations of children, adolescents, and adults. Although problem solving training programs have been implemented on 11 normal 11 populations with positive results, the age range that has been studied has included only children. Normal adolescent populations have not been studied.
In this research project the focus is on addressing these empirical and theoretical gaps in the social problem solving 1 iterature.

RATIONALE AND HYPOTHESES
This research project is designed to examine the relationship between social problem solving and adolescents' high school performance. Social problem solving has been traced to notions of active coping and environmental mastery. It is considered by many psychologists to be a critical component in the cognitive-behavioral repetoire of a competent person. Specific theories of social problem solving were reviewed and the process of problem solving as opposed to the generation of many alternatives was highlighted as the most important consideration. Thus, a theoretical link between the social problem solving process and the psychosocial competence constructs of active coping and environmental mastery has been established. Also, the domains of social and abstract (impersonal) problem solving have been discussed and examined as separate processes. Although both processes require a significant cognitive component, abstract problem solving is more strongly related to traditional notions of intelligence whereas, social problem solving is somewhat independent of verbal intelligence.
Empirical investigations into the relationship between social problem solving and environmental performance have shown that the two are significantly positively related. Intervention studies that train social problem solving skills have demonstrated that more effective social problem solvers are better adjusted to school and home environments. Problem solving has been positively related to the development of social support networks in the community (involving mostly family members). A particular conceptualization of problem solving, means-ends thinking, has shown the strongest relationship to overt behavioral adjustment to the school environment. Means-ends thinking has been shown to discriminate between aberrant and nomal populations across a wide age range. Also, this interpersonal cognitive problem solving skill called means-ends thinking has not been shown to vary significantly by the demographic variables of race, sex, and socioeconomic status.

Rationale
The relationship between social problem solving and environmental performance has been explored from a limited age and developmental perspective and with restricted not ions of what constitutes environmental performance. More specifically, the development of social problem solving in normal high school adolescent populations has been ignored. When social problem solving has been related to school performance only the overt behavioral adjustment of the individual was measured. Overt behavioral adjustment is not an adequate conceptual ization of school performance. Without careful examination of the tasks required by public schools, one cannot fully explore the relationship between problem solving and school performance. Since the majority of children in the mainstream of schools manifest adequate behavioral adjustment, a closer inspection of school perfornance is necessary.
A more comprehensive definition of school performance includes the major task demands.in the academic and social areas. In particular, the high school environment allows a student to participate in activities (e.g., sports; clubs), relate with peers and adults, and achieve grades for classroom learning. Social problem solving has not been related to a complete index of high school performance in the literature to date. This study will examine adolescents' social problem solving processes (means-ends thinking) and school performance in academic and social areas. A study of relatively normal adolescents will contribute to the examination of the role social problem solving plays in environmental performance (see Figure 1).
Previous studies of the social problem solving process have shown limited differentiation by demographic groups. The present study will attempt to confirm these findings in a normal adolescent population.

Hypotheses and Predictions
The first question addressed by this project is whether or not discernable patterns exist in the social problem solving abilities and school performance of students in the high school environment. Thus, the first hypothesis is stated: First hypothesis. It is hypothesized that social problem solving skills, academic perfomance, and social performance do not vary by sex, grade level, or SES.

29
This hypothesis leads to the following predictions about social problem solving and high school performance: 1.
It is predicted that there will be no differences by grade level in: (a) the social problem solving skills of high school students; (b) the academic perfomance of high school students; (c) the social performance of high school students.

2.
It is predicted that there will be no differences by sex in: (a) the social problem solving skills of high school students; (b) the academic performance of high school students; (c) the social performance of high school students.

3.
It is predicted that there will be no difference in SES in: (a) the social problem solving skills of high school students; (b) the academic performance of high school students; (c) the social perfomance of high school students.

4.
It is predicted that there will be no interaction effects in: (a) grade level by sex; (b) grade level by SES; (c) SES by sex.
The second question addressed by the research project is the relationship between high school perfonnance (academic and social) and social problem solving. Thus, the second hypothesis is stated: Second hypothesis. It is hypothesized that high school students with more effective social problem solving skills will exhibit higher academic and social perfomance.
This hypothesis leads to the following predictions: The raore effective the social problem solving skills the: (a) higher the acader.iic perfonnanc· e; and ( b) greater the social perfonnance. 30 From these hypotheses, the following variables are operationally defined for use in th is study. The demographic or student attribute variables of~, grade level and SES are included in order to reflect the potentially different socialization experiences attributable to different group membership. Grade level and sex are reported by the student. The socioeconomic status of the student is operationalized by the education level and occupation the student reports for each parent.
Social problem solving, the first najor construct of this study, is defined as interpersonal cognitive probler., solving (means-ends thinking). Here, students respond to ten problem stories and this score is used as the operational definition of social probler., solving.
High school perfonnance is the second major construct of this study.
It is divided into: 1. academic perfonnance, operationally defined as students' self-report of grades; teachers' ratings of students' academic perfonnance; and number of semesters on the honor roll the previous year; and 2. social perfonnance, operationally defined as students' report of their social support network (size, diversity, and reciprocity) in school; participation in school clubs and activities; and post high school training goals.

Subjects
One hundred-thirty seven (N = 137) students were surveyed from grades 9 through 12 at a medium-sized high school located in southern New England. In order to include equal numbers of male and female students and equivalent groups at each grade level, 16 male and 16 female students' questionnaires were randomly selected from each grade level. Thus, a final sample of 128 students was created for inclusion in the present study. More specifically, there were 32 students (25% of the sample) at each grade level and a total of 64 male (50% of the sample) and 64 fer.1ale students (50% of the sample). The socioeconomic measure utilized in the present study (described later in this section) yielded the following breakdown by group: there were 44 students (34%) in the upper group, 26 (21%) in the middle group, and 58 (45%) in the 1 ower group.
The suburban town in which the high school in this present study is located has a population of approximately 35,000 and a relatively large working and professional class distribution (1980 census). There are relatively equal numbers of working class and professional class families.
All procedures for subject sampling and participation were negotiated with the superintendent of school's office and with the principal of the high school. An attempt \vas made to acquire IQ data for the sampled students from their school records. These negotiations were unsuccessful because IQ data were not available for all students and because of potential violations of student confidentiality. The 9th and 10th grade students were from randomly selected guidance groups. These guidance groups were required of all students during their freshman and sophomore years in high school as part of a standard affective education curriculum; therefore, there were no significant pre-selection criteria that might bias the study. Guidance groups were selected from a master list of counselors and their respective groups that was given to the experimenter by the high school principal.
Students were randomly assigned to guidance groups previous to the start of the school year by the Director of Guidance. The experimenter picked groups randomly from the master list and attempted to include an equivalent number of groups for each counselor. In total, nine guidance groups were sampled from four counselors.
By arrangement with the high school principal, the 11th and 12th grade students were selected by classes from the social studies department. Two teachers were designated by the principal and the experimenter sought their voluntary cooperation. After the teachers volunteered their classes and cooperation, five class groups were chosen by the experimenter. The selection process was not random, but was done to include groups of varying levels of abilities as noted by the teachers. Social studies courses are offered to all 11th and 12th grade students and they are expected to take one before graduation.
Although students from business, vocational, and precollege tracts were represented, the majority of students were in the latter category.
This sample is representative of the high school population in the present study and there is no reason to believe that any pre-selection criteria biased the sample.

Measures
This discussion will focus on how the following variables included in this research project will be measured. The various measures selected for this survey instrument are the standard measures that have been used in studies of social problem solving skills, social competence, and social network. As the measures to be utilized by each variable are presented, the studies fron which they were taken will be cited and any adaptations made for this study will be described.
Reliability and validity data are offered.
Perfomance in high school. As previously discussed, this variable is divided into the components of academic and social perfomance in school.
l. Academic performance. Students were asked to pick the grade, A+, A, A-, B+, B, 8-, C+, C, C-, D+, D, D-, or Fail that best describes their overall work in school. Each rating, student and teacher/counselor, was converted to a score from 1-13 depending on the grade circled (e.g., A+= 1, A= 2, A-= 3, ••. Fail = 13). The student and teacher/counselor ratings were compared using a Pearson product-moment correlation procedure. A coefficient of r = .92 ..e. L .001 was obtained between these b10 ratings.
The teacher/counselor rating was selected as the index of grade point average for subsequent data analysis for the following reasons: (1) its high correlation with the student rating; (2) a simpler more parsimonious data analysis procedure; (3) its referent to teacher ' s/ counselor's familiarity with students• actual performance; and (4) the inclusion of only the past two years academic performance, whereas students were asked to rate their overall grades.
Students were also asked to report the number of marking periods they were on the honor roll last year (all students had an honor roll last year). Finally, students were asked if they planned to enter some post high school training program (e.g., college, vocational program, beautician's school). They answer yes, no, or undecided.
2. Social performance. A measure of social network reported by Mitchell (Note 10,11); Hirsch (Note 12); and Sandler and Barrera (Note 15) was used as an index of social performance in school. This measure required students to identify by initials, the four persons whom they felt were closest or most important to them in school. Then, students indicated their relationship to each support person listed (e.g., friend, teacher, counselor) and whether they relied on that person for material assistance, emotional support, and/or companionship and recreation. Finally, the student was asked whether each support person listed relied on the student for material assistance, emotional support, and/or companionship/ recreation. Students checked 11 yes 11 or 11 no 11 for the support categories across each support person listed.
Each 11 yes 11 answer received a score of l, and 11 no 11 answers were scored 0.
The social network measure yielded the following five separate comprised of the number of 11 yes 11 answers checked for support received from others; {4) Instrumentality (score 0-12), comprised of the number of "yes" answers checked for support given (others' reliance on the student); and (5) Total network score (0-24), which consisted of the total number of 11 yes 11 answers checked across all support categories for all support persons listed. Students' social networks were also scored for reciprocity and type of relationships. An index of reciprocity was calculated for each support relationship listed by assigning a one (1) to each relationship that had an equal nur.iber of 11 yes 11 answers for support received versus support given and a two (2) for relationships that had an unequal amount of support given versus received. Thus, relationships fell into the categories reciprocal or nonreciprocal.
The reciprocity score was a measure of the balance between material assistance, emotional support, and companionship/recreation that was given to support people and received from ther.1. Relationships were also categorized according to peer or adult affiliation for further analysis.
The multidimensionality and instrur.1ental ity of support networks has been found to distinguish: (1) the degree of psychological effectiveness between 11 cl inical 11 and 11 normal 11 populations (Tolsdorf, 1976;Mitchell, Note 10, 11); and (2)  and (3)   The MEPS test has been found to discriminate between psychiatric and normal populations along the dimension of psychological adjustment across many age groups (Spivack & Shure, 1974;Spivack, Platt, & Shure, 1976). A similar procedure has been shown to predict behavioral adjustment to school in populations of younger children (Shure, 1979).
The present study utilized the entire standarized MEPS procedure that presented students with ten short stories, each with a beginning and an ending section. The student was asked to write the middle of the story. Spec if ical ly, they were asked to write a few sentences that would connect the beginning of the story (problem description) to the end of the story. There were both male and female versions to all ten stories, differing only by the gender of the names and pronouns used.
Consequently, female students received stories with the female gender and male students received the stories with the male gender.
The MEPS instrument has extended reliability and validity data (Spivack & Shure, 1974;Spivack, Platt, & Shure, 1976;Platt & Spivack, 1974a;1974b;1974c;Note 13;Note 6). The authors report: (1) discriminant validity with college students and delinquent adolescents; (2) content validity through a series of three factor analytic studies that identify one general problem solving factor suggesting that all the stories measure the same quality of thinking; (3) construct • validity using college students and other measures of problem solving; and (4) predictive validity with a population of youthful heroin offenders (Platt & Spivack, 1975). In addition, the factor analytic studies of the MEPS test yielded the general problem solving factor just mentioned and three specific factors: Factor I, problem production and solution; Factor II, interpersonal facilitation; and Factor III, consideration of consequences. These three factors accounted for 70% of the total variance (Platt & Spivack,Mote 13).
Reliability data on the MEPS tests were obtained through a series of five studies using mostly psychiatric populations (Platt & Spivack, 1975 Students with questions were encouraged to ask the experimenter for assistance. Items were administered one at a time until the questionnaire was completed. The experimenter(s) monitored each student carefully to ensure that students understood the instructions and \·tere able to read the ite171s. All students completed each section before the group was allowed to continue. 43 After the questionnaires were completed each student wrote his/her name and the questionnaire identification number on a piece of paper.
The experimenter explained that the students would be rated on academic performance by the teacher or counselor in a manner similar to their own self report. Students that did not wish to be rated by the teacher or counselor were told not to put their name on the paper. Also, the experimenter explained that when the teacher or counselor rated the students• academic performance the list with the names and numbers on it was destroyed. The rer.iaining ratings identified the student only by nur.iber.
Teachers and counselors rated each student's academic performance using the 13 point rating system. The experimenter was given the ratings for each student and the corresponding identification numbers the next day.
Students, teachers, and counselors were thanked for their participation. Individual feedback regarding the results of the study was given both verbally and in writing to the school administration and the students, teachers, and counselors involved.

RESULTS
In this study it is predicted that: (1) social problem solving skills and academic and social perfomance do not vary by sex, grade level or SES.; and (2) high school students with more effective social problem solving skills exhibit higher academic and social performance.
In order to test these hypotheses the data were organized into the cate- In general, the data supported the hypothesi s that there were no demographic differences in problem solving ability and school and cornr:,unity performance. However, grade level and SES. differences were found in certain variables. Similarly, the data supported the hypothesis that effective problem solvers showed higher high school performance than less effective problem solvers. High problem solvers had higher grades and more effective performance on selected social indices. Also, problem solving ability was significantly related to selected community performance indices. Regression equations indicated that there was a highly significant relationship between problem solving skills and school and community performance indices.
In the following discussion of these results, the relationship between the der.iographic variables and school and cor:n:,unity perfon:iance will be presented first. Next, dernograph ic differences in total probler.i solving will be presented, followed by problem solving differences across the various school and community performance variables. Finally, the results of the multiple regression analyses and intercorrelations between solving and performance variables will be presented.
Demographic results. were suppressed. · The factorial ANO VA allows for a more pars irnon ious data analysis and limits the chance of obtaining an F score by chance.
The present section will discuss just the demographic main and . interaction effects (see Table 3). Problem solving results will be discussed in a later section.
The discontinuous school and community perforr.tance variables were analyzed with a Chi-square statistic.
Specific sex, SES., and grade level results are presented (see Table 3).
Academic performance. The category includes grade point average rating, nunber of senesters on the honor roll, and post high school training goals. The results for each variable will be presented separately.    Total number of support people Number of school support people Numb · er of comnunity support people Multidimensionalityschool network lnstrumenfality (school) Reliance on others (school) Total school network score Multidimensionalitycommunity network Instrumentality (community) Reliance on others (community) Total community network score Total reciprocity of school and community networks Type of school relationships Type of community relationships Type of relationships in total network F = 5.14** F = 4.16** F = 3.37* F = 3.70* F = 6.92** F = 3.85* F = 6.30**  A significant two-way interaction for grade level by sex f. (3,127) = 2.91, .E. L.05 was also obtained. The interaction effect accounted for 4.5% of the total variance for the g.p.a. scores A simple effects test for this interaction found that sophonore males had significantly higher g.p.a. ratings than sophomore females or any other group at any grade level (see Table 4). Sophomore females had significantly higher g.p.a. ratings than freshmen males, junior males and females, and senior males and females.  . 05 \'las obtained. The interact ion effect accounted for 4. 3% of the total variance for this variable. A simple effects test found that sophomore males were on the honor roll significantly more often than sophomore females or any other group at any other grade level (see Table 5). Also, sophomore females were on the honor roll significantly more frequently than freshman males.  Table 6 for mean scores . ). l.
Clubs. This variable was divided into total number of clubs, number of school clubs, and number of community clubs for analysis.

Each of these variables was used as a dependent variable in separate
AMOVAs to determine whether there were der:iographic differences. There were no significant ma in or two-way interact ion effects for total nur:iber of clubs, number of school clubs, or number of comr.,unity clubs.
In general, students were more involved in school than corrmunity clubs and tended to be involved in at least one club somewhere. For this sample, club participation does not appear to be related to demographic variables.
2. Jobs. The continuous dependent variable, total number of jobs, v,as examined using a series of two-way AtJOVAS. The discontinuous variables, past jobs, and present jobs each included the categories: no job; sur.111er job only; part-time/year-round-only; and both surrr.1er and part-time/year-round and \·1ere analyzed with a Chi square statistic.
There were no significant relationships between any of the demographic variables (sex, SES., grade level) and the present job categories.
Also, there were no significant relationships beb,een sex and SES., and the past job categories. Finally, no main effects for sex or SES. and no two-way interaction effects were obtained with total number of jobs.
A significant ma1n effect for grade level was obtained for total number of jobs f. (3,127)     A significant effect for type of relationship and SES. was obtained ?:: 2 (1) = 7. 22, .e_ L . 01. Upper SES. students had significant 1 y more adult relationships than lower SES. students. There were no significant differences between middle and upper and middle and lower SES. students.
Also of note, was the fact that most of the support relationships (89%) fell into the category of peer.
To summarize, community networks did not vary by denographic variables. These results are consistent with original predictions. There were limited denographic differences in the school network indices.
In general, school networks varied by socioeconanic status with upper SES. students having more network properties than either middle or lower SES. students. Specifically, the socioeconanic differences can be stJTimarized as follows: a) upper SES. students had more support people than middle SES. students, but not more than lower SES. students; b) upper SES. students had greater multidimensionality to their network than middle or lower SES. students; and c) upper SES. students had greater instrtJTientality (others' reliance on them), reliance on others, and total school network scores than lower SES. students, but not greater than middle SES. students. These socioeconanic differences are contrary to initial predicti6ns in the first hypothesis.
Grade level differences in school social networks were found along the dimensions of reliance on others and total network score. Sophomores, juniors and seniors consistently had more school network properties and relied more heavily on others for m .ateri al, enoti onal and recreational support than freshmen students. This is contrary to the predictions offered in the first hypothesis.
The ccmnunity network had overall mean scores similar to the school network scores yet the former had almost no differences by sex, SES., problem solving, or grade level. Only the number of ccmnunity support people listed showed any variance by demographic variables; upper SES. students had more support people in the ccmnunity than middle SES. students, but not more than lower SES. students. It should be noted that both school and ccmnunity networks did not include family members.

Problem Solving Results
As a further test of the first hypothesis, that there will be no demographic differences in problem solving and school perfonnance, a a series of two-way ANOVAs (sex by grade level, sex by SES., SES. by grade level) were c011puted with problem solving (MEPS score) as the dependent variable. There were no significant main effects for sex or SES. and no significant two-way interaction effects. There was a significant main effect for grade level f. (3,127) Table 6). These limited demographic differences offer further support for the first hypothesis.
The prediction that effective problem solvers will perfonn better in school, was tested wi.th a series of two-way ANOVAs and Chi-square analyses. As previously described, the problem solving score was trichot011ized to create high, medium, and low problem solving groups.
Problen solving was then included as an independent variable with the denographi c variables of sex, SES. and grade 1 evel as each school and perfonnance variable was analyzed. The problen solving differences in school and community perfonnance variables will be presented next.

63
There were no significant main or two-way interaction effects for problen solving and the following variables: (1) number of clubs (total, school, carmunity); (2) Table 7). The problen solving effect accounted for 39.7% of the variance in g.p.a. and 33% of the variance in number of senesters on the honor roll. Further analysis of the probl en .
A significant two-way interaction effect of SES. by problem solving F (3,127) = 2.91, .e. L .05 was obtained for g.p.a. ratings. The interaction effect accounted for 4.3% of the total variance for g.p.a. A simple effects test found that upper SES., . high problem-sol·vers (!i = 3. 88) had significantly higher g.p.a. ratings than low SES., high problem solvers (!i = 4.68) or any other group (see Table 8). A significant two-way interaction effect of sex by problem solving F (2,127) = 3.51, .e_L .05 was obtained for number of semesters on the honor roll. The interaction effect accounted for 5.2% of the total variance for this variable. A simple effects test indicated that both male and female high problem solvers were on the honor roll significantly more often than male and female middle and low problem solvers. Also, male, middle problem solvers were on the honor roll significantly more often than female, low problen solvers (see Table 9).  of the total variance in the reliance-community variable. A Tukey A procedure indicated that high problem solvers had significantly higher reliance scores -(school network,~= 10.33; community network,~= 9.71) than middle problem solvers (school network,~= 9.00; cCJTITlunity network, network, _!i = 8.24). Middle and low problen solvers did not differ significantly from each other in their mean reliance scores. Thus, high problen solvers seened to rely more on their peers for support than did their middle and low problen solving counterparts.
A significant relationship was found bet~een problen solving and reciprocity of support relationships -X, 2 (1) = 11.46, .e_L .01. Further analysis of this result revealed that high problen solvers had significantly more reciprocal relationships than did middle and low problem solvers.
Middle and low problen solvers did not differ significantly on this dimension.
As a further analysis of the reciprocity variable the school and carmunity network rel ati ans hips were examined separately. A significant associationship was found between problen solving and reciprocity for both school X 2 (1) = 12.47, £.L .01 and community x, 2 (1) = 11.94, .e_L .01 relationships.
In both networks high problen solvers had significantly more reciprocal relationships than did middle and low problen solvers. Middle and low problen solvers did not differ from each other in either the school or community reciprocity analyses. The total reciprocity frequency table was divided into total reciprocal relationships versus nonreciprocal relationships for further comparison. There were significantly more reciprocal relationships (72%) than nonreciprocal relationships (28%) ')(.. 2 (1) = 262.11, .e_L. .001 (Total N = 970).
In summary, problen solving differences in acadenic performance and social performance offer support for the second hypothesis. H-igh problen solvers: . (1) received higher g.p.a. ratings; (2) were _ on the honor roll more often last year; (3) had more year-round jobs in the past; (4) relied more heavily on peers for enoti anal, material, and recreati anal support, and (5) had more reciprocal support relationships than middle or low problem solvers. For the most part, middle and low problem solvers did not differ si gnifi cantl y in acadeni c and social perf onnance.
The next sections offer a further test to the second hypothesis and discuss the more general relationships that exist between the continuous school and community perfonnance variables and problem solving scores.
First, Pearson product-manent correlation coefficients are presented between problem solving skills, school perfonnance, and community perfonnance. Second, the i nter-correl ati ans between the total problem solving score and the ten stories that comprise the MEPS test are Pearson product-manent correlation coefficients were also computed for each of the 10 MEPS pro bl ems and the total score (N = 128) • .

70
Correlations between each of the stories and the total MEPS score ranged from r = .36 tor= .64. Stories one U: = .64), three (J: = .59), and four (r = • 54) had the highest i ntercorrel ati ans with the total MEPS score. Each story seems to have a relatively independent contribution to the total score according to the intercorrelation matrix presented in Table 11. Individual story intercorrelations ranged fran a high of r = to.!= .26 (see Table 10 for specific correlations). Although the correlations r = .15, .! = .18, .! = .26 are statistically significant they account for such a small portion of the total variance that they are practically insignificant. These data suggest that there is a relatively strong relationship between academic performance and school and coomunity perf onnance as measured by the devel opnent of social networks.
Nunber of clubs, an index of student involvement, was found to be significantly related to all the school and community network variables.! = .16 tor= .29 (see Table 10 for specific correlations). However, the correlations in the r = .16 to_!= .29 range are practi ·cally insignificant because of the smal 1 portion of variance they account for. Thus, number of clubs can be considered as a related component to social  networks. By contrast, number of jobs, an index of coomunity involvement, (or an econanic reality) was not significantly related to any of the school or coomunity perfonnance variables.
The school and community network indices have significantly high intercorrelations with each other r = .42 tor= .57. Within the 73 school network, the measures of multidimensionality, instrumentality, reliance on others and total network score carrel ate highly with each other.!:= .79 to..!:= .95 (see Table 10). The same trend exists in the canmunity network scale, multidimensionality, instrunentality, reliance on others, and total network score i ntercorrel ate highly.!: = . 83 to..!: = .95. These data suggest that the school network indices are largely measuring the same thing. A similar statenent may be offered for the canmunity network indices. The intercorrelations between the school and community networks are significantly high as well,.!:= .35 tor= • 61 ( see Table 10). A 1 though there appears to be a . . strong overlap in infonnation provided by both the school and community networks, they are also measuring a distinct amount of separate infonnation. That is, it cannot be assumed that if a particular student had a strong network in one area that the other area would be equally well developed.
Problen solving and school perfonnance. A stepwise multiple regression analysis was used to examine the relationships between problen solving skills and school perfonnance and as a further test of the second hypothesis, that higher problen solvers will perfonn more effectively in school. The total MEPS score, an index of problen solving, was used ,as the dependent variable.  Table 12). The average contribution of each predictor was calculated by entering the variables into the regression equation in all possible orders (See Table 13).

Order
Note: Both predictors seemed to be providing relatively independent sources of infonnation to the overall prediction of social problem solving skills.

77
Social problem solving can be predicted by community perfonnance variables of social network and number of jobs. Social problem solving can also be predicted by academic and social perfonnance variables; g.p.a., honor roll, number of clubs and social network. The relationship between school perfonnance and social problem solving is much stronger than the relationship between cOOTT1unity perfonnance and problem solving.

Summary
The results of the present study offer support for both the first and second hypotheses. Specifically, limited demographic differences were found with school perfonnance variables and problem solving.
Grade level differences were found in academic perfonnance with sophomores exhibiting higher perfonnance than other grade levels. A sex by grade level interaction was found with sophanore males and females functioning higher than other groups at differing grade levels. There were also limited demographic differences in social perfonnance, both in school and in the community. Differential perfonnance by SES. was found across the school social network i _v . properties.
Usually, upper SES. students perfonned higher than middle and lower SES. students. However, upper SES. students were not consistently higher than both lower and middle SES. groups, suggesting that the differentiation between SES. groups on the school network variables is not stable. Grade level differences were noted on one dimension of the school network (reliance on others) with sophanores, juniors and seniors functioning higher than freshTien. Similarly, grade level differences were found in job participation; seniors had more jobs and more past, year-round jobs than their fresh'nen, sophomore, and junior counterparts. Grade 1 evel differences were al so found in social problen solving; fresh'nen achieved significantly lower problen solving scores than their upper grade level counterparts. Thus, significant support for the first hypothesis was obtained. No differences by sex were found across school and ccmnunity perfonnance indices. Sane SES.
and grade level differences were found in selected perfonnance variables.
The second hypothesis was strongly supported by the results of the present study. Problen solving differences were noted in grade point rating, number of senesters on the honor roll, past jobs held, school social network (reliance on others), canmunity social network (reliance on others), and number of reciprocal support relationships (both in school and canmunity). Consistently, high problem solvers perfonned more effectively than middle and low problen solvers. Specifically high problen solvers had significantly stronger acadenic grade ratings and were more able to rely on others for support than lower problen solvers. Further, the support relationships tended to be reciprocal in nature; equal anounts of recreational, ernoti onal, and physical support were given and received.
Multiple regression analyses indicated that problen solving can be predicted effectively by academic and social perfonnance in high school and by ccmnunity performance indicators. The strongest single predictor of problan solving is acadanic perfonnance (__r: =· .64). -However, social network variables both in school and in the carmunity were strong predictors of problan solving ability. Also, number of jobs held accounted for a significant portion of the variance. Club participation (school and community) was not a significant predictor of probl an solving ability. These results also offer strong support for the second hypothesis.

DISCUSSION
The present research has explored the rel ati onshi p between social problen solving skills and adolescents' acadenic and social perfonnance. In general, the results confinned the hypothesis that social problen solving skills are significantly relatep to high school acadenic and social perfonnance. Also, the results indicate that 80 pro bl en solving ski 11 s are si gnifi cantl y related to cannuni ty perfonnance. Perfonnance in school and community was found to be positively related. These findings are consistent with past studies that link social problen solving skills to notions of school adjustment and personal canpetence.
The data also indicated that certain denographic differences existed in social problen solving skills and school perfonnance, contrary to the first hypothesis. While no sex differences were found socioeconanic and grade level differences were obtained.
The discussion section will be organized in the following manner.
First, the results of this study will be canpared with past theory and research.
In partic~lar, the following themes will be discussed: (1) the relationship between social problen solving skills and broader notions of school and environmental perfonnance; (2) the relationship between social problen solving skills and perfonnance in different environments; the "transfer" of social problen solving as a construct; and (3) the developnent of social problen solving skills in adolescence as a function of experience and cognitive factors. Next, th~ implications for further research will be discussed with an enphasis on longitudinal methodology and cross-situational experiments to verify sane of the theoretical notions just postulated. Finally, the present findings will be discussed in tenns of implications for psychological service delivery in the schools, especially within a primary preventive framework.

Denographic Differences
The data indicate that denographic differences existed in the develoJJ11ent of school social support networks. Specifically, upper SES. students had developed more extensive social support relationships in school than middle and lower SES. groups. However, the differences between upper, middle, and 10\'/er SES. groups were not consistent across all school social network variables. That is, upper SES.
students perfonned more effectively than middle SES. students, but not more effectively than lower SES. students on variables measuring network size in the school and community. On variables measuring support given and received, upper SES. students were usually higher than both middle and 10\'/er SES. students. The more effective perfonnance of upper SES. students in developing social support networks in school was not found along most of the community social network dimensions. Even though both school and ccmnunity networks for the students consisted mostly of peers, upper SES. students did not develop more extensive support systens in the ccmnunity than their middle and 1 O\'/er SES. counterparts. Actua 11 y, most students sampled had an effective school and ccmnunity support network. The. SES. differences found in the present study have not been clearly identified in past social network research (Mitchell & Trickett, 1981).
Social Pro bl em Solving and Environment al Perf onnance The major finding in this study is that social problem solving ski 11 s are si gnifi cantl y related to broader notions of academic and social performance in school. The link between social problem solving skills and limited notions of school perfonnance has been established (e.g., Spivack & Shure, 1974;Spivack, Platt, & Shure, 1976;Hopper & Kirschenbaum, Note 1). This study provides additional support to Spivack and Shure's model of the interpersonal cognitive problem solving abi 1 ity of means-ends thinking as a mediator of environmental performance. Specifically, no sex or SES. differences in social 82 problem solving skills were found, consistent with past findings by Spivack, Platt, & Shure (1976). More importantly, the findings that social problem solving skill~ in adolescence relate strongly to school and camtunity performance suggests that problem solving is a critical skill that enables the individual to negotiate everyday problems and structure support systems across settings. Similarly, Platt and Spivack (1975) state: " ••• the investigations on real-life problem solving thinking clearly indicate an area of thought process that bears a direct relationship to human adjustment and of which means-ends thinking is an important part". (p. 11) Platt and Spivack (1975) al so state that these cogni ti ve-behavi oral processes involve a sensitivity to the presence of hunan problems, the ability to generate alternative courses of action, the ability to conceptualize the means to solve the problem, and the awareness of the consequences of one's actions. These authors assert that human problemsolving skills have relevance to envi ronnental performance across a wide age span, a broad range of human adaptive ability, and across disparate socioeconanic groups (Platt & Spivack, 1975). The findings in the present study strongly reinforce this model, especially with regard to problem solving skills and school performance and problem solving as a skill that is transferable across situations and envi ronTlents.

83
A strong association between academic perf onnance and social problem solving skills was found. Although previous studies (Hopper & Kirschenbaum, Note l;Spivack & Shure, 1974;Shure & Spivack, 1978) have found only minor associations between IQ and means-ends thinking the strength of the association between problem solving ability and academic performance in the present study suggests that both may share a similar set of cognitive-behavioral skills.
Social problem solving skills were found to be associated with social performance in school and in-the community. In particular, high pro bl em solvers farmed more reciprocal support rel ati ans hips, held more jobs in the past, and relied more on support personnel in school and in the carrnunity than their middle and lower problem solving counterparts. This finding suggests that a problan solving component is associated with school and carrnunity performance. It seems that this skill is transferable fran one situation to another (e.g., classroan performance, forming support relationships) and fran one environT1ent to another (e.g., perfonnance in school and perfonnance in the community). High problan solvers in this study functioned more effectively in the classroan and in deriving support fran peers than their lower problan solving counterparts. Similarly, high problan solvers had more past jobs and derived more support fran peers in the community than lower problan ~olvers. The notion that social problem solving as a cognitive-behavioral cluster of skills is applicable across many situations is similar to other cognitive-personality dimensions identified by Mischel (1979).
The relationship of social problen solving skills to both school and CaTTTlunity perfonnance is similar to the 11 benign 11 cycle of competence stressed by Smith (1968 ••• interpersonal problen-solving style was positively and significantly related to the ntJT1ber of intimates cited by the respondent. While most of the social network 1 i terat ure views the i ndi vi dual as a passive agent within the web of social ties, this finding suggests that the i ndi vi dual may play an active role in structuring the size of his/her network. Clients may have to exercise a variety of interpersonal skills in order to take advantage of the environment provides.
More generally, this finding suggests the need to look more broadly at the range of coping skills that individuals . use to shape their networks. (pp. [15][16]. The differential perf onnance in school and community areas for various problan solving abilities offers anpirical support for the theory of Jahoda (1953) and others who suggest that the social problan solving process is an active coping skill that enables the individual 85 to engage the environment in his/her quest for mastery and self improvement. Further, the present study suggests that social probl an solving ski 11 s transfer f ram one setting to another.
The strong association between acadeni c perf onnance and social problan solving skills found in the present study suggests that a significant behavioral-cognitive factor is influencing both of these

Developnent of Social Problem Solving in Adolescence
The results of this study suggest that there are sane danographic differences in social problan solving skills and school perfonnance.
Specifically, grade level differences were found in acadanic perfonnance, social problan solving skills, and social perfonnance. In acadenic perfonnance sophanores showed higher functioning than other grade levels. _ Although this finding is not supported by past studies, it is consistent with other data in the present study. For exc111ple, freshmen were si gnif i cantl y lower than sophomores in social pro bl an solving abilities and in the extent to which their social networks were developed. Further, sophomores and upper grade level menbers tended to rely more on the support personnel in their network for enotional, physical, and recreational assistance than frest-rnen. Thus, in acadenic perfonnance, social problan solving skills and social perfonnance in school, differential perfonnance was noted between frest-rnen and upper grade level menbers. One explanation for this difference is that upper grade level menbers are more familiar with the school environment and thus perfonn more effectively.
Another explanation is also tenable. The possibility that a shift in adaptation to school occurs as a function of the develoJJTlent of interpersonal-cognitive problen solving skills merits further exploration. The work of Piaget as discussed by Flavell (1963) (Phillips, 1975;Piaget, 1952). Fran this, it could be concluded that fonnat. This could be accomplished through modeling and discussions.
Teachers of social science courses might include a problem solving canponent in their approach to their course.
Social pro bl em solving could al so be i ncTuded in guidance counseling and psychological counseling in high school settings. This could be fonnally accomplished through role-modeling and training in groups or in one-to-one counseling situations where the counselor would review the dilenmas facing the student or create novel problens. The aim would be to facilitate the developnent of consequential thinking, means-ends conceptualization, and generation of alternative solutions.
Inf onnal ly, counsel ors could encourage students to approach pro bl ems in a systenatic manner fostering an active role rather than a passive one.
Large scale intervention of a primary preventive fonnat would ideally occur in the first years of high school using a fonnal groupguidance approach. Students and a trained counselor would follow a prescribed training course including other dimensions besides ·social problen solving. Such an approach (e.g., Spivack & Shure's model) -should await further research on problen solving skills · training with younger children and further delineation of the devel opnent of these skills in adolescents before implenentation is possible.

Sumnary
Social prohlen solving skills are significantly, positively related to school and cannunity perfonnance in the present study. High probl en solvers showed more effective acadenic perfonnance, relied more on their peers for support in school and in the cannunity, had more reciprocal support relationships, and held more jobs in the past than their lower problen solving peers. Thus, social problen solving · skills were linked to effective perfonnance in differing environments, suggesting that as a set of skills, it is transferable across envirorments. This finding awaits verification in future research.
Social problen solving skills and social performance in school were found to differ for freshnen and sophomores. The notion of a developmental period in social pro bl en solving par al 1 el i ng more advanced cognitive develoP'Tlent was posited. School experience could be a significant factor in the develoP'Tlent of these skills. Again, this finding awaits verification in future research.
Suggestions were made for including problen solving in regular education curricula and in guidance and psychological counseling.
Larger scale preventative programs should be carefully scrutinized before they are implenented on any level. Future research is needed to further delineate the acquisition of social problen solving skills in adolescence and their relationship to high school performance.
No. ----- In this procedure we are interested in your imagination. You are to make up sane stories. For each story you wi 11 be given the beginning of the story and how the story ends. Your job is to make up a story that connects the beginning that is given to Jou with the ending given you. In other words, you will make up the middle of the story.
Write at least one paragraph for each story. wanted to have friends in the neighborhood. The story ends with Mr. C. having many good friends and feeling at hone in the neighborhood. You begin the story with Mr. C. in his room immediately after arriving in the neighborhood. 106 5. During the Nazi occupation a man's wife and children were viciously tortured and killed by an SS trooper, and the man swore revenge.
The story begins one day after the war, when the man enters a rest au rant and sees the ex-SS trooper. The story ends with the man killing the SS trooper. You begin when he sees the SS trooper.
6. One day Al saw a beautiful girl he had never seen before while eating in a restaurant. He was immediately attracted to her. The story ends when they get married. You begin when Al first notices the girl in the r estaurant.
7. Bob needed money badly. The story begins one day when he notices a valuable diamond in a shop window. Bob decides to steal it. The story ends when he succeeds in stealing the dic111ond. You begin when he sees the dicmond.
8. John noticed that his friends seened to be avoiding him. John wanted to have friends and be liked. The story ends when John's friends like him again. You begin where he first notices his friends avoiding him. 9. One day George was standing around with sane other people when one of then said sanething very nasty to George. George got very mad.
George got so mad he decided to get even with the other person.
The story ends with George happy because he got even. You begin the story when George decided to get even.
10. Joe is having trouble getting along with the forenan on his job.

107
Joe is very unhappy about this. The story ends with Joe's foreman liking him. You begin the story where Joe isn't getting along with his foreman. ,.    I feel the test-retest reliability is acceptable in considering that data was obtained on aberrant population, in some cases with significant lapse in time between test and retest. I would wish it was some-what higher, and it probably would be if test-retest were done with normal population with a week or two lapsing in between.

NO
In general I am satisfied with the validity of t..r1e MEPS: there are · now numerous studies indicating it differentbates between and within groups in ways consistent with ICPS theory.
The MEPS. is the best ICPS measure available currently for adolescents.
Unfortunately, no careful study of abbreviated forms of MEPS has been ·done.
'!'here is evide.'"1-ce t.i.'lat a..,long adults there is internal consistency among the stories. My own impression is that I would feel. safer using four stories rather than three stories. Stories 1, . 2, 3, 4, and 7 seem to load higher on the common element being tapped (at least among adult . patients). However, one must also consider the appropriateness of the content of the story. Again, among adults eight and nine tend to have lower loadings.
I regret, for your sake and mine, that more is not . known about which are the best stories for teenagers. , Do I understand you correctly in that you have data on stories 2, 3, and 8? If so, how do they intercorrelate among the teenagers? · I would appreciate your full reference to Hopper and Kirschenbaum (1979), if you get a chance to drop me a card.
In summary, let me say that the MEPS seerr..s to be reliable and sensitive enough to pick up significant differences between groups.
It relates to a variety of measures of level of adjustment.