STRESS-BUFFERING FACTORS RELATED TO ADOLESCENT COPING: A PATH ANALYSIS.

Printz, Brian L. ; Shermis, Mark D. ; Webb, Patrick M. 等

ABSTRACT

To uncover those factors that buffer the impact of stressful negative experiences on adolescent adjustment, a theoretical model of adolescent stress and coping, with social support and social problem solving proposed as moderators, was investigated using path analysis. The study was conducted with 122 ninth and tenth-grade nonreferred high school students. Using the LISREL statistical package (Joreskog & Sorbom, 1986), it was found that a recursive loop leading from stress outcomes back to negative stressors did not allow for a successful solution to the model. However, the effects of stressful events on adjustment were mediated by coping resources, which included a combination of problem-solving abilities and social support. Overall, the findings replicated previous investigations that have demonstrated direct relationships among stressful life events, social support, problem solving, and adolescent adjustment. While a successful fit to the theoretical model was not attained, it was concluded that a refined model may provide a more acceptable solution.

INTRODUCTION

Unhealthy adaptation to stress can take many forms, such as school maladjustment. For example, stressors at home and school may lead to reduced attention span and to diminished motivation to succeed academically (Pryor-Brown & Cowen, 1989). Some students develop socially maladaptive coping patterns, including verbal and physical aggression toward others, defiance of authority, acting out, and juvenile delinquency (Compas, Howell, Phares, Williams, & Giunta, 1989). Anxiety (Swearingen & Cohen, 1985a, 1985b), depression, and suicidal

ideation (Cohen-Sandier, Berman, & King, 1982) are other reactions to stress. Moreover, some youths experience psychophysiological symptoms in response to chronic or severe levels of stress (Walker & Greene, 1987).

The availability of coping resources varies considerably among students. During times of high stress, these resources, whether social or individual, may be insufficient. When coping resources, such as problem-solving skills, are inadequate, stressful situations may give rise to unhealthy outcomes (Spirito, Overholser, & Vinnick, 1995; Wills, Vaccaro, & Benson, 1995).

Research has sought to explicate factors associated with adolescents' stress-related coping. Models have been constructed to identify the key variables in the stress-adjustment relationship (see Lazarus & Folkman, 1984; Petersen & Spiga, 1982; Rice, Herman, & Petersen, 1993; Shermis & Coleman, 1990). However, such models have rarely been examined using rigorous statistical analyses that allow for modification of the particular model to provide for a goodness of fit of key variables based upon the findings. Moreover, there continues to be little understanding as to how the results of such investigations can be applied to interventions in the area of adolescent health. The intent of the present investigation was to specify and test a model of stress-related coping among high school students, examine the model's key components, and utilize this information to provide guidelines for school-based interventions to address adolescent health issues.

Shermis and Coleman Model

While adult models of stress and coping processes have been postulated (e.g., Lazarus & Folkman, 1984), there is a paucity of models for adolescents. Shermis and Coleman (1990) have offered a cognitive-behavioral model of adolescent stress and coping (see Figure 1). It has five major components: environmental stressors, environmental moderators, personal factors, stress outcomes, and behavioral outcomes.

Environmental stressors include daily hassles (e.g., getting involved in an argument, experiencing bad weather, having plans change unexpectedly) and major life events (e.g., parental divorce, death of a friend or relative, serious illness or injury), with differential effects (Compas, 1987b). Environmental moderators include support from family members, peers, and school personnel. These individuals may offer advice, teach skills, provide material aid, help the adolescent overcome emotional distress, and share responsibilities (Compas, 1987a). Shermis and Coleman (1990) suggest that it is the adolescent's perception of support that actually determines the extent to which the effects of stress are moderated.

Personal factors--cognition being prominent--also may impact affective and behavioral outcomes. Shermis and Coleman (1990) have identified self-talk as one form of cognitive coping. An earlier conceptualization (Chandler, 1985) included age, cognitive appraisal (e.g., perceptions of threat or loss, perceptions of control), self-esteem, and problem-solving skills as personal moderating factors.

Stress outcomes may include physical and psychological symptoms. Chandler (1985) noted the outcomes of stress in the areas of school functioning and social relations.

Finally, behavioral outcomes, which are linked to stress outcomes, can be thought of as secondary responses to stress. Shermis and Coleman (1990) have listed drug abuse, delinquency, pregnancy, and dropping out of school as maladaptive behaviorial responses to stress.

Stressful Life Events

The emerging consensus within the literature is that the experience of stressful events during childhood greatly increases an individual's vulnerability to behavioral and psychological maladjustment, in addition to physical illness. Pearson correlation coefficients have been found to range from .10 to .68, with the majority of studies reporting coefficients between .20 and .30 (see Compas, 1987b). In general, major life events were predictive of daily hassles, which in turn were associated with adjustment.

Typically, life stress models conceptualize mediating factors as being either personal or environmental. For example, investigations of personal factors have examined demographic variables (Dohrenwend, Krasnoff, Askenasy, & Dohrenwend, 1978), temperament (Kagan, 1983), attachment and separation during infancy (Ainsworth, 1979), social problem solving (Mullins, Siegal, & Hodges, 1985), and Type A and B behavior patterns (Dweck & Wortman, 1982). Investigations of environmental factors have focused on social support as a resource for coping (Barrera, 1981; Compas, Wagner, Slavin, & Vannatta, 1986; Pryor-Brown & Cowen, 1989; Walker & Greene, 1987). The present study targeted those factors that can be influenced via school-based interventions (i.e., social support and problem solving). Moderating variables that are relatively permanent or immune to treatment (i.e., temperament, personality type, and attachment during infancy) were not assessed.

Social Support

Perceived social support is the cognitive appraisal of the presence and quality of interpersonal ties (Barrera, 1986). For school-aged children, these sources of support primarily include family members, school personnel, and peers.

Overall, research suggests that perceived social support and psychological and physical symptoms are negatively correlated (Billings & Moos, 1981). For example, results of three path analytic studies (Dean & Ensel, 1982; Lin & Dean, 1984; Lin & Ensel, 1984) indicate that perceived social support is negatively correlated with depressive symptoms.

The results have been mixed concerning the stress-buffering role of social support. Several investigations with child and adolescent samples (Compas et al., 1986; Gad & Johnson, 1980; Rowlison & Felner, 1988) did not note any significant interactions among life events, social support, and symptom levels. In contrast, Walker and Greene (1987) found that as negative life events increased, adolescent males with low peer support reported significantly more symptoms than did males with high peer support. Additionally, they noted that at low levels of life stress, peer support was not significantly associated with males' symptomatology. However, peer support appeared more necessary for the well-being of adolescent females, regardless of the level of life stress. Family support was negatively associated with males' and females' reports of symptomatology.

The study of social support has produced findings consistent with several different models of stress, support, and adjustment. While critics (Barrera, 1986; Compas, 1987a; Heller & Swindle, 1983) do not rule out the possibility of the stress-buffering position, they indicate that conceptual and methodological deficiencies hinder definitive conclusions.

Social Problem Solving

Problem solving has been proposed as a moderating variable in models of stress and coping (e.g., Nezu, 1987). Studies by Spivack, Platt, Shure, and their associates have investigated deficits in the ability to generate multiple ways of solving a problem. Means-ends performance among impulsive adolescents at a residential treatment center (Spivack & Levine, 1963), adolescent psychiatric patients (Platt, Spivack, Altman, Altman, & Peizer, 1974), and emotionally disturbed children (Larcen, Spivack, & Shure, 1972; Shure & Spivack, 1972) was found to be markedly lower than that of control groups.

Investigations of adult populations (Heppner & Anderson, 1985; Heppner & Petersen, 1982; Heppner, Hibel, Neal, Weinstein, & Rabinowitz, 1982; Heppner, Reeder, & Larson, 1983; Nezu, 1985, 1986) have demonstrated the mediating role of self-perceived problem-solving ability. Individuals who rated themselves as ineffective at solving problems were more likely to exhibit depressive symptomatology (Nezu, 1986), reported more problems (Heppner et al., 1982) and general symptoms of maladjustment (Heppner & Anderson, 1985), and had more irrational beliefs and lower self-concepts (Heppner et al., 1983). Examinations of adolescent self-appraised problem-solving abilities are needed.

In short, the ultimate objective is to identify the relationships among key components of adolescent coping. Thus, the purpose of this study was to determine how well the empirical data from an adolescent sample fit the conceptual model presented by Shermis and Coleman (1990).

METHOD

Subjects

One hundred twenty-two high school students, 65 males (53%) and 57 females (47%), served as subjects. They were between 14 and 19 years old (M = 15.25, SD = 0.98). Thirty-nine were freshmen, 79 were sophomores, 3 were juniors, and 1 was a senior. These students were not receiving special education services.

Sixty-seven students lived with their biological mother and father, 23 lived with a biological parent and a stepparent, 28 lived with one parent, and 4 had adoptive parents or lived with other relatives. The racial distribution of the students was as follows: Anglo (n = 74, 61%), Asian (n = 21, 17%), Hispanic (n = 19, 16%), African-American (n = 6, 5%), and other racial groups (n = 2, 2%). The high school from which they were drawn was located in a lower-middle-class to middle-class suburb of Houston, Texas.

Procedure

Teachers from six health classes explained the purpose of the study to their students. Informed consent (parent and child) was obtained for 144 of the 160 eligible students (90%).

To reduce order effects, the five instruments, administered over two days, were counterbalanced. Twenty-two students did not complete the assessment phase for various reasons (e.g., absence from class). Those whose scores indicated that they were experiencing a. high degree of emotional distress were contacted, along with their parents, and referral information and counseling opportunities were provided.

Instrumentation

Environmental stressors. The high school version of the Adolescent Perceived Events Scale (APES; Compas, Davis, Forsythe, & Wagner, 1987) includes separate scales for daily (APES-D) and major (APESM) events, and for positive and negative events. Compas et al. (1987) reported high test-retest reliability over a two-week period. Several studies examining relationships between negative life events and psychological and behavioral symptoms have employed the APES. For example, using weighted negative events and the Hopkins Symptom Checklist (Derogatis, Lipman, Rickels, Uhlenhuth, & Covi, 1974), Wagner, Compas, and Howell (1988) reported correlations ranging from .23 to .32 for major events, and from .54 to .68 for daily events.

Social support. The Perceived Social Support self-report scales (PSS; Procidano & Heller, 1983) are designed to assess perceptions of the adequacy of support from family (PSS-Fa) and friends (PSS-Fr). Procidano and Heller reported high internal consistency (alpha coefficients of .88 for the PSS-Fa and .90 for the PSS-Fr).

Social problem solving. The self-report Social Problem-Solving Inventory (SPSI; D'Zurilla & Nezu, 1988) measures social problem-solving ability and consists of 10 subscales comprising 70 items. Two composite scales, the Problem Orientation Scale (POS) and the Problem Solving Skills Scale (PSSS), were of interest here. D'Zurilla and Nezu (1988) reported test-retest coefficients of .83 for the POS and .88 for the PSSS. Internal consistency was also high (alpha coefficients of .89 and .81, respectively). Concurrent validity was assessed by comparing the SPSI with the Problem Solving Inventory (PSI; Heppner & Petersen, 1982) and the Means-Ends Problem-Solving Procedure (MEPS; Platt & Spivack, 1975). Correlations between the PSI and the P05 and PSSS were - .62 and - .69, respectively. Correlations between the MEPS and the POS and PSSS were .61 and .65, respectively.

Psychological adjustment. The Stress Response Scale (SRS; Chandler, 1983) is a 40-item behavior rating instrument that was designed to assess the emotional adjustment reactions of children. Teachers or parents are required to rate children's behaviors on a six-point scale. The SRS is derived from Chandler's (1983) stress response model and is supported by factor analyses (Chandler, 1982). The five-factor solution that forms the basis for the subscales--Impulsive (acting out), Passive-Aggressive, Impulsive (overactive), Repressed, and Dependent--accounted for 62% of the variance in a clinically referred sample (Chandler, 1982), 64% of the variance in a nonreferred sample (Chandler & Lundahl, 1983), and 65% of the variance with a randomly selected group (n = 962) of children (Shermis & Chandler, 1985).

The Youth Self-Report (YSR; Achenbach & Edelbrock, 1987) consists of 112 items taken from the Child Behavior Checklist (CBCL; Achenback & Edelbrock, 1983, 1986) and reworded in the first person. Designed for youth aged 11-18, the YSR is a relatively stable measure of self-reported social competence and behavior problems. Achenbach and Edelbrock (1987), examining a sample of youths referred to a community mental health facility, reported good test-retest reliability (.69) for total behavior problem scores over a 6-month period. They also found a high degree of consistency between CBCL ratings (by mothers and clinicians) and YSR ratings at the time of intake and at 6-month follow-up.

Data Analysis

The hypotheses were tested using techniques appropriate for latent variable models, including covariance structure analysis (CSA) and factor analysis. CSA allows for testing of bidirectional relationships between variables, as well as relationships involving recursive loops. In addition, CSA and path analysis permit statements concerning the direction of causal flow between variables ("causal" refers to the degree of impact one variable, or a set of variables, has on another variable). Thus, they are a step beyond the predictive power of traditional correlational and regression analyses. Factor analysis is commonly used to confirm the relationships among the structural and measured variables, or as an exploratory tool to assess problems within the observed model.

CSA entails the construction of diagrams that specify causal flow among the variables within a model, and then solving for these relationships (Loehlin, 1992). A path diagram, initially constructed based on existing knowledge or assumptions, depicts relationships among latent (theoretical) and/or observed variables. Each of the relationships in a path model can be expressed with a linear structural equation. There will be numerous structural equations to account for a given model. The CSA process involves solving for the unknown coefficients (which are indicative of the strength of the causal relationships) in each of the equations simultaneously, thus reducing the amount of error (Aaronson, Frey, & Boyd, 1988). The goal is to seek the most parsimonious explanation for the data. Through the process of simultaneous iteration, computer programs such as LISREL6 (Joreskog & Sorbom, 1986) and SIMPLIS (Joreskog & Sorbom, 1987) use matrix algebra to solve for the unknown path coefficients and to test the degree of variance, or "fit," between the implied (theoretical) and the observed models.

LISREL6, the computer program used in the present investigation, is based on the LISREL model, which provides for the inclusion of both latent and observed variables. Observed, or measured, variables are thought to be indicators for latent variables. A latent variable is best estimated when it serves as a composite for several observed variables, provided that the observed variables

are valid and reliable indicators of the theoretical construct.

The LISREL model requires the construction of a structural model and a measurement model. The structural model consists of the relationships among the theoretical constructs, or latent variables. For this study, the structural model was composed of four latent variables, as illustrated in Figure 2: negative stressors, perceived social support, problem solving, and stress outcomes. In contrast, the measurement model, as shown in Figure 3, identified the observed variables (i.e., instruments or indices) used to measure the latent variables in the structural model. Five pencil-and-paper instruments, including several subscales, were used in the measurement model: APES (APES-D, APES-M), PSS (PSS-Fr, PSS-Fa), SPSI (POS, PSSS), SRS, and the YSR. With the exception of the SRS, which involved teacher reports, all were youth self-report measures.

RESULTS

An initial correlation matrix was generated that included each of the variables present in the model. Although a covariance matrix could have been generated for the input variables, the correlation matrix was substituted to facilitate interpretation of the results. Because correlations are standardized variables, they allow for more straightforward comparisons across variables.

The initial model proposed that all of the latent variables be placed on the Y-side (right) of the model. This allowed for a feedback loop from stress outcomes back to environmental stressors. Further, paths between each of the mediator variables (social problem solving, social support) and stress outcome variables were identified. Personal factors were excluded due to insufficient sample size.

Specification of the initial model failed to produce a significant match between the data and the model. A [x.sup.2] value of 58.22 with 9 degrees of freedom (p [less than] .001) and a goodness of fit (GFI) index of .901 was obtained. Moreover, some of the residuals proved to be negative. Further, the SRS did not perform as expected; instead, it performed in the reverse direction of the YSR. In addition, when personal factors were included in the model, the problems were exacerbated. These findings indicated that adjustments to the model were necessary.

Respecification of the Model

In order to identify the sources of error in the original model, several methods suggested by Joreskog and Sorbom (1986) were employed: (1) consideration of alternative plausible models, (2) factor analysis, and (3) elimination of constructed variables. First, however, problems encountered in attempting to fit the original model necessitated the exploration of a more parsimonious relationship among the data. The initial step in reducing the complexity of the model involved removing the reciprocal path leading from stress outcomes back to negative stressors. In addition, the negative stressor variable and measures were placed on the X-side of the model, leaving the mediator and outcome variables on the Y-side. Due to its poor performance, the SRS was eliminated from the analysis, which left only the YSR as a measure of stress outcomes. Since it could not be assumed that the YSR was a perfect measure of the stress outcome variable, the YSR's reliability was used as a best estimate of the path from the latent v ariable to YSR. Furthermore, based on LISREL's analysis of where the structural model was most problematic, a direct path from negative stressors to stress outcomes was included. Once these changes were implemented, several variations of the model were explored.

Alternative plausible models. The strength of each of the mediator values was tested. In succession, single paths leading from the mediating variables to the stress outcome variable were removed to determine their individual contribution to the model. Each of the trials remained consistent with the underlying theoretical framework of the original model. The models produced identical [X.sup.2] values (39.54, df = 6, p[less than].001) and GFI values (.912), while adjusted GFIs ranged from .591 to .650, depending on df in the particular model. These findings suggested that each of the trials solved for the model equally well, though none of the exploratory models provided an acceptable fit for the data. To further assess problems with the model and to construct a better model, a confirmatory factor analysis was undertaken.

Factor analysis. To assess the possibility that a poorly constructed or implemented measurement model was responsible for the poor fit with the data, a confirmatory factor analysis was conducted. This involved allowing all of the variables between the latent variables in the original model (excluding SRS and personal characteristics) to be correlated. No change was detected, [X.sup.2](6) = 39.54, p = .001, suggesting that the poor fit may have been due to the measurement model.

Once the original model was rejected, the nature of the investigation became exploratory. An exploratory factor analysis was conducted in which the SRS was removed due to its poor performance, while the YSR was included since it was the only indicator for the stress outcomes variable. Using orthogonal rotation, two underlying factors were uncovered. As would be expected, the APES-M and APES-D loaded on one factor, identified as "negative stressors." Contrary to original projections, though in keeping with a more parsimonious conceptualization, the PSS-Fa, PSS-Fr, POS, and PSSS loaded on the second factor, identified as "coping resources."

Elimination of constructed variables. Indicators produced by the initial test, together with the results of the exploratory factor analysis, suggested that significant problems were inherent in the model. Due to these problems, the LISREL program was unable to generate t values and standard error estimates on the path coefficients. The significance and strength of the paths in the model could not be assessed without these indices. Further, the program located a major problem with the path from the PSS-Fr variable to the perceived support latent variable.

Due to the problematic nature of the PSS-Fr variable, it was important to determine whether eliminating perceived support's direct impact on stress outcomes would enhance the model. However, models that allowed both problem solving and perceived support to have a direct impact on stress outcomes demonstrated a significantly better fit of the data, [X.sup.2]diff(1) = 5.33, p [less than].05.

Based on the results of the factor analysis and trials with various models, a more parsimonious model was explored, involving three latent variables: negative stressors, coping resources, and stress outcomes. Only one mediating latent variable, consisting of the PSS-Fr, PSS-Fa, POS, and PSSS, was used in the final model (see Figure 4). Further, a path between the residuals of the PSS-Fr and the PSSS was freed, since a correlation between the two variables may have accounted for some of the variance in the model. The correlation matrix for the respecified model is shown in Table 1.

Assessment of the Goodness of Fit of the Respecified Model

Minimization programs typically produce several indicators of the goodness of fit of a model (Joreskog & Sorbom, 1987). The LISREL6 program generates the chi-square statistic, multiple correlation coefficients, GFI and adjusted GFI, root mean square residual (RMR), as well as a Q-plot of data, which assess the efficacy of a model. Perhaps most commonly reported is the chi-square statistic; a close approximation of the model to the data is indicated when the ratio of the chisquare to the degrees of freedom is small. Path analysis of the final model produced a relatively large chi-square (32.86, df = 11,p [less than] .001), suggesting that the respecifled model (three latent variables) did not fit the data, although the improvement over the original model was significant, [X.sup.2]diff(2) 25.36, p [less than] .00 1.

Squared multiple correlation coefficients indicate that a small amount of the variance was accounted for by the mediating variable ([R.sup.2] = .204), while stress outcomes accounted for a significant amount of the variance ([R.sup.2] = .873). Thus, the model does not account for all of the variance; factors outside of the model appear to be impacting the dependent variables.

The GFI, a measure of the amount of variance between the implied and the observed variables (Joreskog & Sorbom, 1987), for this model was .934 (.831 when adjusted for degrees of freedom). These two indices are relatively independent of the sample size and are affected minimally by departures from normality (Joreskog & Sorbom, 1987).

The root mean square residual (RMR), a measure of the average size of the estimated residuals, for the final model was .096. The RMR statistic is most valuable for comparing the effectiveness of separate models (Joreskog & Sorbom, 1987). Among the models tested, the final model had the lowest RMR.

DISCUSSION

Similar to the earlier work of Chandler and Shermis (1990), the present study failed to support the notion that maladjustment exacerbates the amount of negative life stress experienced by high school students--the model's recursive loop was found to be ineffective. However, findings on the role of negative life events were consistent with other research indicating that an accumulation of stressful experiences increases vulnerability to maladjustment (Barrera, 1981; Compas et al., 1989; Compas & Phares, 1986; Gad & Johnson, 1980; Sandier, 1980). Most studies have reported correlations between stress and adjustment ranging from .20 to .30 (see Compas, 1987b). This study found a similar correlation for major events (r = .33, p [less than] .05), but a higher correlation for daily events (r = .46, p [less than] .001).

The latter finding suggests that adolescent adjustment is influenced less by discrete events than by chronic stressors (Compas, 1987b). While demonstrating some predictive power in the present study, major events proved to be a poor indicator of student distress in several previous investigations (e.g., Wagner, Compas, & Howell, 1988). Here, this variable was causally related to daily events, which, in turn, impacted adolescent adjustment. Moreover, other researchers (Compas et al., 1989; Compas, Wagner, Slavin, & Vannatta, 1986) have argued that the reciprocal relationship between stress and symptomatology excludes major life stressors.

While a successful fit between the respecified model and the data was not attained, a trend in the findings supported previous research on the stress-adjustment relationship and mediating factors (Peterson & Spiga, 1982; Sherniis & Coleman, 1990). High school students' perceptions of available coping resources, including social support and problem-solving abilities, significantly buffered the impact of stressors on symptomatology. When examined separately, students' problem-solving abilities and perceptions of social support from friends and family were not powerful enough to reduce the effects of negative stress in their lives.

In the attempts to find a solution for the structural model, a major source of the variance between the implied and observed models was attributed to the path between a measurement variable, PSS-Fr (peer support), and its predetermined latent variable, social support. PSS-Fr was found to correlate significantly (r = .39, p [less than] .005) with PSSS (problem-solving skills). In addition, the degree of association between PSS-Fr and POS (problem orientation), an indicator of cognitive appraisal rather than social skill, was nonsignificant (r = .13). These results suggest that support from friends and support from family, as coping resources, serve unequivocal functions.

Soliciting support from peers requires social competencies (the ability to build and maintain friendships). The difference in the correlations between peer support and the two problem-solving measurement variables supports the contention that PSS-Fr scores are indicative of social skills. Similarly, Procidano and Heller (1983) found that the PSS-Fr was more highly associated with measures of social competence than was the PSS-Fa. Taken together, these findings suggest that general orientation to solving problems and perceptions of family functioning better predict adolescent adjustment problems than do social skills.

Support from friends did not have a significant impact on distress. This may indicate that even during adolescence, the support received from family members is more vital for healthy functioning than is the quality of affective bonds with friends. Alternatively, it may suggest that peers serve as a secondary coping resource; their help is solicited when the family proves nonsupportive.

Like the two support variables, problem orientation and problem-solving skills may play very different roles. Students' problem orientation was more highly associated with measures of stress and adjustment, while problem-solving skills were more closely associated with perceived support from friends. Thus, cognitive appraisal of effectiveness at resolving problems appeared to have a greater impact than did actual skill (see Nezu, 1987).

Overall, these data for a nonreferred sample of ninth- and tenth-grade students provide some indication that coping resource variables are impacted by stressful events, and consequently impact self-rated adjustment. While scores suggest that students had adequate problem-solving skills, it appears that those with a negative orientation to problems were less successful at coping with stress. More specifically, cognitive and affective factors related to problem orientation may have impacted the ability of some of the students to employ effective coping strategies. Although the stress-buffering hypothesis was not supported with the combined social problem solving construct, it is highly probable that problem orientation, along with other cognitive factors, plays a moderating role. Further research into the nature of the cognitive factors that hinder effective coping is needed.

A limitation of this study was its reliance on self-report measures to assess the latent variables. In addition, methodological problems associated with the use of the SRS, a teacher-rated measure of student adjustment, severely limited the strength of the stress outcomes variable (large classrooms meant teachers lacked familiarity with students). Since the SRS did not provide discriminatory power, the YSR served as the lone index of adjustment. Future investigations should incorporate reports from others, observational data, and clinical interviews. Furthermore, assessments of other hypothesized moderators (e.g., school-based social support, cognitions, and temperament), as well as inclusion of several outcomes of stress (e.g., academic performance and additional forms of symptomatology), would be valuable.

CONCLUSIONS

The primary source of stress for adolescents appears to be chronic interpersonal and nonsocial problems. When faced with a seemingly overwhelming accumulation of micro-stressors (i.e., daily hassles), they may feel less capable of solving their problems, despite having the necessary skills. Instead, they often resort to avoidance, shift causal attributions to factors beyond their control, or adopt irrational beliefs. Therefore, interventions that focus on improving adolescents' orientation to problems--helping them effectively resolve daily stressors--are recommended for reducing symptomatology (see D'Zurilla, 1988).

The assessment of stressors, problem solving, and social support offers valuable insight into adolescents' perceptions of their environment and ability to cope. Their problem orientation provides an indicator of coping efficacy, while appraisal of problem-solving skills suggests areas in need of improvement. Finally, indices of family support reveal the availability of important coping resources.

The assessment of life events, problem orientation, problem-solving skills, and social support should prove especially helpful to mental health professionals working at the high school level. Students exhibiting symptoms of maladjustment can be offered group counseling based on their specific needs and situation; for example, those requiring social skills training or cognitive restructuring, those who are experiencing family conflict versus peer relationship problems, and those who would benefit from general stress management techniques (e.g., relaxation training). Further examination of the respecified model of adolescent coping should be conducted, with appropriate linkages to school-based interventions.

This article is based on the first author's doctoral dissertation, completed at the University of Texas at Austin.

Mark D. Shermis, Ph.D., Director, Testing Center, and Associate Professor, Department of Psychology, Indiana University-Purdue University Indianapolis.

Patrick M. Webb, Ph.D., Testing Center and Department of Psychology, Indiana University-Purdue University Indianapolis.

Reprint requests to Brian Printz, Ph.D., Director of Special Education, Pikes Peak B.O.C.E.S., 4825 Lorna Place, Colorado Springs, Colorado 80915.

REFERENCES

Aaronson, L. S., Frey, M. A., & Boyd, C. J. (1988). Structural equation models and nursing research (part 2). Nursing Research, 37(5), 315-318.

Achenbach, T. M., & Edelbrock, C. (1983). Manual for the Child Behavior Checklist and Revised Child Behavior Profile. Burlington, VT: Department of Psychiatry, University of Vermont.

Achenbach, T. M., & Edelbrock, C. (1986). Manual for the teacher version of the Child Behavior Checklist and Profile. Burlington VT: Author.

Achenbach, T. M., & Edelbrock, C. (1987). Manual for the Youth Self-Report. Burlington, VT: Department of Psychiatry, University of Vermont.

Ainsworth, M. D. S. (1979). Infant-mother attachment. American Psychologist, 34, 932-937.

Barrera, M., Jr. (1981). Social support's role in the adjustment of pregnant adolescents: Assessment issues and findings. In B. H. Gottleib (Ed.), Social networks and social support in community mental health. Beverly Hills, CA: Sage.

Barrera, M., Jr. (1986). Distinctions between social support concepts, measures, and models. American Journal of Community Psychology, 14, 413-445.

Billings, A. G., & Moos, R. H. (1981). The role of coping responses and social resources in attenuating the stress of life events. Journal of Behavioral Medicine, 4, 139-157.

Chandler, L. A. (1982). The Stress Response Scale: A preliminary report. Paper presented at the convention of American Psychological Association, Washington, DC.

Chandler, L. A. (1983). The Stress Response Scale: An instrument for use in assessing emotional adjustment reactions. School Psychology Review, 12, 260-265.

Chandler, L. A. (1985). Assessing stress in children. New York: Praeger.

Chandler, L. A., & Lundahl, W. T. (1983). Empirical classification of emotional adjustment reactions. American Journal of Orthopsychiatry, 53, 460-467.

Chandler, L. A., & Shermis, M. D. (1990). A paradigm for the study of childhood stress. In J. H. Humphrey (Ed.), Human stress: Current selected research (Vol. 4). New York: AMS Press.

Cohen, L. H., Burt, C. E., & Bjork, J. P. (1987). Effects of life events experienced by young adolescents and their parents. Developmental Psychology, 23, 583-592.

Cohen-Sandler, R., Berman, A. L., & King, R. A. (1982). Life stress and symptomatology: Determinants of suicidal behavior in children. Journal of the American Academy of Child Psychiatry, 21, 178-186.

Compas, B. E. (1987a). Coping with stress during childhood and adolescence. Psychological Bulletin, 101, 393-403.

Compas, B. E. (1987b). Stress and life events during childhood and adolescence. Clinical Psychology Review, 7, 275-302.

Compas, B. E., Davis, G. E., Forsythe, C. J., & Wagner, B. M. (1987). Assessment of major and daily life events during adolescence: The Adolescent Perceived Events Scale. Journal of Consulting and Clinical Psychology, 55, 534-541.

Compas, B. E., Howell, D. C., Phares, V., Williams, R. A., & Giunta, C. T. (1989). Risk factors for emotional/behavioral problems in young adolescents: A prospective analysis of adolescent and parental stress and symptoms. Journal of Consulting and Clinical Psychology, 57, 732-740.

Compas, B. E., & Phares, V. (1986, August). Child and parental stress and symptomatology: An integrative analysis. Paper presented at the annual meeting of the American Psychological Association, Washington, DC.

Compas, B. E., Wagner, B. M., Slavin, L. A., & Vannatta, K.(1986). A prospective study of life events, social support, and psychological symptomatology during the transition from high school to college. American Journal of Community Psychology, 14, 241-257.

Dean, A., & Ensel, W. M. (1982). Modelling social support, life events, competence, and depression in the context of age and sex. Journal of Community Psychology, 10, 392-408.

Derogatis, L. R., Lipman, R. S., Rickels, K., Uhlenhuth, E. H., & Covi, L. (1974). The Hopkins Symptom Checklist. Pharmacopsychiatry, 7,79-110.

Dobrenwend, B. S., Krasnoff, L., Askenasy, A. R., & Dohrenwend, B. P. (1978). Exemplification of a method for scaling life events: The PERI Life Events Scale. Journal of Health and Social Behavior, 19, 225-239.

Dweck, C. S., & Wortman, C. B. (1982). Learned helplessness, anxiety, and achievement motivation: Neglected parallels in cognitive, affective, and coping responses. In H. W. Krohne & L. Laux (Eds.), Achievement, stress, and anxiety (pp. 93-125). Washington, DC: Hemisphere.

D'Zurilla, T. J. (1988). Problem-solving therapies. In K. S. Dobson (Ed.), Handbook of cognitive-behavioral therapies. New York: Guilford Press.

D'Zurilla, T. J., & Nezu, A. (1982). Social problem solving in adults. In P. C. Kendall (Ed.), Advances in cognitive-behavioral research and therapy (Vol. 1, pp. 201-274). New York: Academic Press.

D'Zurilla, T. J., & Nezu, A. M. (1988). Development and preliminary evaluation of the Social Problem-Solving Inventory (SPSI). Manuscript submitted for publication.

Gad, M. T., & Johnson, J. H. (1980). Correlates of adolescent life stress as related to race, SES, and levels of perceived social support. Journal of Clinical Child Psychology, 9, 13-16.

Gersten, J. C., Languer, T. S., Eisenberg, J. G., & Orzek, L. (1974). Child behaviors and life events: Undesirable change or change per se? In B. S. Dobrenwend & B. P. Dohrenwend (Eds.), Stressful life events: Their nature and effects. New York: Wiley.

Heller, K, & Swindle, R. W. (1983). The effects of social support: Prevention and treatment implications. In R. D. Felner, L. A. Jason, J. N. Moritsugn, & S. S. Farber (Eds.), Preventive psychology: Theory, research and practice (pp. 87-103). New York: Pergamon Press.

Heppner, P. P., & Anderson, W. P. (1985). The relationship between problem solving self-appraisal and psychological adjustment. Cognitive Therapy and Research, 9, 415-427.

Heppner, P. P., Baumgardner, A., & Jackson, J. (1985). Problem-solving self-appraisal, depression, and attributional style: Are they related? Cognitive Therapy and Research, 9, 105-113.

Heppner, P. P., Hibel, J. H., Neal, G. W., Weinstein, C. L., & Rabinowitz, F. E. (1982). Personal problem solving: A descriptive study of individual differences. Journal of Counseling Psychology, 29, 580-590.

Heppner, P. P., Kampa, M., & Brunning, L. (1984, August). Problem-solving appraisal as a mediator of stress. Paper presented at the annual meeting of the American Psychological Association, Toronto.

Heppner, P. P., & Petersen, C. H. (1982). The development and implications of a personal problem solving inventory. Journal of Consulting and Clinical Psychology, 29, 66-75.

Heppner, P. P., Reeder, B. L., & Larson, L. M. (1983). Cognitive variables associated with personal problem-solving appraisal: Implications for counseling. Journal of Counseling Psychology, 30, 537-545.

Joreskog, K. G., & Sorbom, D. (1986). LISREL6: Computer program. Moorseville, IN: Scientific Software, Inc.

Joreskog, K. G., & Sorbom, D. (1987). SIMPLIS: Users guide. Moorseville, IN: Scientific Software, Inc.

Kagan, J. (1983). Stress and coping in early development. In N. Garmezy & M. Rutter (Eds.), Stress, coping, and development in children. New York: McGraw-Hill.

Larcen, S., Spivack, G., & Shure, M. B. (1972, April). Problem-solving thinking and adjustment among dependent-neglected preadolescents. Paper presented at the meeting of the Eastern Psychological Association, Boston.

Lazarus, R. S., & Folkman, S. (1984). Stress, appraisal, and coping. New York: Springer.

Lin, N., & Dean, A. (1984). Social support and depression: A panel study. Social Psychiatry, 19, 83-91.

Lin, N., & Ensel, W. M. (1984). Depression-mobility and its social etiology: The role of life events and social support. Journal of Health and Social Behavior, 25, 176-188.

Loehlin, J. C. (1992). Latent variable models: An introduction to factor, path, and structural analysis (2nd ed.). New Jersey: Lawrence Erlbaum Associates.

Mullins, L. L., Siegal, L. J., & Hodges, K. (1985). Cognitive problem-solving and life event correlates of depressive symptoms in children. Journal of Abnormal Child Psychology, 13, 305-314.

Nezu, A. M. (1985). Differences in psychological distress between effective and ineffective problem solvers. Journal of Counseling Psychology, 32, 135-138.

Nezu, A. M. (1986). Cognitive appraisal of problem-solving effectiveness: Relation to depression and depressive symptoms. Journal of Clinical Psychology, 42, 42-48.

Nezu, A. M. (1987). A problem-solving formulation of depression: A literature review and proposal of a pluralistic model. Clinical Psychology Review, 7, 121-144.

Nezu, A. M., & Ronan, G. F. (1985). Life stress, current problems, problem solving, and depressive symptoms: An integrative model. Journal of Consulting and Clinical Psychology, 53, 693-697.

Pearlin, L. I., Menaghan, E. G., Lieberman, M. A., & Mullan, J. T. (1981). The stress process. Journal of Health and Social Behavior, 22, 337-356.

Perkins, D. V. (1982). The assessment of stress using life events scales. In L. Goldberger & S. Breznitz (Eds.), Handbook of stress: Theoretical and clinical aspects. New York: MacMillan.

Petersen, A. C., & Spiga, R. (1982). Adolescence and stress. In L. Goldberger & S. Breznitz (Eds.), Handbook of stress: Theoretical and clinical aspects. New York: MacMillan.

Platt, J.J., & Spivack, G. (1972). Problem solving thinking of psychiatric patients. Journal of Consulting and Clinical Psychology, 39, 148-151.

Platt, J.J., & Spivack, G. (1973). Studies in problem-solving thinking of psychiatric patients: Patient-control differences and factorial structure of problem-solving thinking. Proceedings of the 81st Annual Convention of the American Psychological Association, 8, 461-462.

Platt, J.J., & Spivack, G. (1975). Manual for the Means-Ends Problem-Solving Procedure (MEPS). Philadelphia: Hahnemann Community Mental Health/Mental Retardation Center.

Platt, J. J., Spivack, G., Altman, N., Altman, D., & Peizer, S. B. (1974). Adolescent problem solving thinking. Journal of Consulting and Clinical Psychology, 42, 787-793.

Procidano, M. E., & Heller, K. (1983). Measures of perceived support from friends and family: Three validation studies. American Journal of Community Psychology, 11, 1-24.

Pryor-Brown, L., & Cowen, E. L. (1989). Stressful life events, support, and children's school adjustment. Journal of Clinical Child Psychology, 18, 214-220.

Rice, K. G., Herman, M. A., & Petersen, A. C. (1993). Coping with challenge in adolescence: A conceptual model and psycho-educational intervention. Journal of Adolescence, 16, 235-251.

Rowlison, R. T., & Felner, R. D. (1988). Major life events, hassles, and adaptation in adolescence: Confounding in the conceptualization and measurement of life stress and adjustment revisited. Journal of Personality and Social Psychology, 55, 432-444.

Sandler, R. N. (1980). Social support resources, stress and maladjustment of poor children. American Journal of Community Psychology, 8, 41-52.

Sandler, I. N., & Block, M. (1979). Life stress and maladaptation of children. American Journal of Community Psychology, 7, 425-439.

Shermis, M. D., & Chandler, L. A. (1985). A confirmatory factor analysis of the Stress Response Scale. Unpublished manuscript, University of Pittsburgh.

Shermis, M. D., & Coleman, M. (1990). Stress and coping: A cognitive-behavioral model. Unpublished manuscript.

Shermis, M. D., Rudin, D., & Chandler, L.A. (1990). An extension of the norms for the Stress Response Scale for Children. Paper presented at the annual meeting of the American Educational Research Association, Boston.

Shure, M. B., & Spivack, G. (1972). Means-ends thinking, adjustment and social class among elementary school aged children. Journal of Consulting and Clinical Psychology, 38, 348-353.

Spirito, A., Overholser, J., & Vinnick, L. (1995). Adolescent suicide attempters in general hospitals: Psychological evaluation and disposition planning. In J. L. Wallander, L. J. Siegel et al. (Eds.), Adolescent health problems: Behavioral perspectives, advances in pediatric psychology (pp. 97-116). New York: Guilford Press.

Spivack, G., & Levine, M. (1963). Self-regulation in acting-out and normal adolescents (Report M-4531). Washington, DC: National Institute of Mental Health.

Swearingen, E. M., & Cohen, L. H. (1985a). Measurement of adolescent's life events: The Junior High Life Experiences Survey. American Journal of Community Psychology, 13, 69-85.

Swearingen, E. M., & Cohen, L. H. (1985b). Life events and psychological distress: A propsective study of young adolescents. Developmental Psychology, 21, 1045-1054.

Wagner, B. M., Compas, B. E., & Howell, D. C. (1988). Daily and major life events: A test of an integrative model of psychosocial stress. American Journal of Community Psychology, 16, 189-205.

Walker, L. S., & Greene, J. W. (1987). Negative life events, psychosocial resources, and psychophysiological symptoms in adolescents. Journal of Clinical Child Psychology, 16, 29-36.

Wills, T. A., Vaccaro, D., & Benson, G. (1995). Coping and competence in adolescent alcohol and drug use. In J. L. Wallander, L. J. Siegel et al. (Eds.), Adolescent health problems: Behavioral perspectives, advances in pediatric psychology (pp. 160-178). New York: Guilford Press.

 Correlations of the Variables Involved in Model Verification
Variables SRS YSR APES-D APES-M PSS-Fr PSS-Fa
SRS 1.000
YSR .139 1.000
APES-D -.078 .457 [****] 1.000
APES-M -.050 .332 [*] .389 [****] 1.000
PSS-Fr -.130 -.131 .188 .200 1.000
PSS-Fa .032 -.496 [****] -.279 -.134 .245 1.000
POS -.063 -.450 [****] -.272 -.192 .133 .370 [***]
PSSS -.229 -.119 .009 .127 .387 [***] .262
Variables POS PSSS
SRS
YSR
APES-D
APES-M
PSS-Fr
PSS-Fa
POS 1.000
PSSS .351 [*] 1.00
(*.)p [less than] .05
(**.)p [less than] .01
(***.)p [less than] .005
(****.)p [less than] .001