|
|
A multisystems analysis of adolescent suicide attempters.
Samantha R. Levy; Gregory L. Jurkovic; Anthony Spirito.
Journal of Abnormal Child Psychology, April 1995 v23 n2 p221(14)
Abstract: A study was conducted to examine the factors which influence suicidal behavior in adolescents using an ecologically based model. Multiple-regression analyses of the behavior of adolescent males and females revealed hopelessness as the strongest and most reliable predictor of suicidal intent and ideation. However, hopelessness was observed to vary within the group of suicide attempters studied indicating time as an important factor that should be considered in future studies.
Full Text: COPYRIGHT Plenum Publishing Corporation 1995
As the rate of adolescent suicide and parasuicide has increased, so have investigative efforts to examine factors associated with these phenomena. No single predictor of suicide has been identified. Rather, suicidal behavior, ideation, and intent appear related to a variety of factors (Spirito, Brown, Overholser, & Fritz, 1989), spanning different levels of analysis (individual, family, and sociocultural).
On the individual level, research has pointed to the central role of hopelessness. Beck and his colleagues (1985, 1990) studies of the predictive validity of hopelessness in adults have indicated that cognitive aspects of depression are more closely related to suicide than affective aspects of depression. Studies with psychiatric inpatient children have also found that hopelessness predicts suicidal intent and ideation, independent of depression (Asanrow & Guthrie, 1989; Carlson & Cantwell, 1982). Although few studies have included exclusively adolescent samples, studies of adolescent psychiatric inpatients have found that hopelessness covaries with number of suicidal gestures, seriousness of intent, and medical lethality of attempts (Brent, Kolko, Goldstein, Allan, & Brown, 1989; Robbins & Alessi, 1985; Topol & Reznikoff, 1982). In these studies, suicidal psychiatric inpatients' hopelessness scores were significantly higher than both normal controls and psychiatric controls, indicating that hopelessness in adolescent suicide attempters is not merely a function of psychological disturbance but plays a specific role in suicide.
Family-level variables also have received considerable attention in the adolescent suicide literature. It appears that family interactional processes are more consistently related to suicidal thinking and behavior in adolescents than such demographic and descriptive variables as divorce, family history of suicide, and physical and sexual abuse (see Spirito, Stark, Fristad, Hart, & Owens-Stively, 1987). Problematic family patterns associated with suicidal behavior in adolescents include conflict (e.g., Hawton, O'Grady, Osborn, & Cole, 1982; Mattsson, Seese, & Hawkins, 1969; McIntire & Angle, 1980), poor communication (e.g., Sanborn, Sanborn, & Cimbolic, 1973; Williams & Lyons, 1976), lack of cohesion (Miller, King, Shain, & Naylor, 1992), and rigidity (Miller et al., 1992). Conflict with parents is reported as the precipitant to adolescent suicide in 40% (Mattsson et al. 1969) to 76% (Hawton et al., 1982) of the samples. Family conflict is also sometimes expressed by either subtle or blatant expression by the family of a "death wish" for the patient (Sabbath, 1969). Communication problems are especially relevant to the study of adolescent suicide attempters. Hawton et al. (1982), for example, found that significantly more adolescent suicide attempters reported communication problems with their parents than nonattempters in a normal control sample.
A small number of researchers has also examined the relation of sociocultural variables to adolescent suicide. For example, as with other social and psychological problems, rates of suicide and suicidal ideation are inversely related to socioeconomic status (SES) (Dubow, Kausch, Blum, Reed, & Bush, 1989; Garfinkel, Froese, & Hood, 1982; Stack, 1982).
A major limitation of research in this area is that variables associated with adolescent suicidal behavior and ideation have been examined largely in isolation. The scope, moreover, of studies exploring more than one factor (e.g., depression and hopelessness) has typically been restricted to a single level of analysis (e.g., Rotheram-Borus & Trautman, 1988). To the extent that more than one level is represented, the interrelation of variables is not considered within the context of a coherent explanatory framework (e.g., Brent, 1987; Topol & Reznikoff, 1982). The role of SES, for example, is often included merely as a background variable. Finally, none of the previous studies in this area have attempted to predict suicidal ideation and intent within a sample of attempters. Rather, suicide attempters have been merely compared to nonattempters on various measures. To account for and to explain the variance in adolescent suicidal behavior, research is needed that (a) tests causal models, (b) includes multiple levels of analysis (see Bronfenbrenner, 1977), and (c) explores within-group variability.
The present study explored how particular individual, family, and sociocultural factors influenced suicidal intent and ideation in a sample of adolescent suicide attempters. Of the various factors that have been explored in this area, family discord, socioeconomic status, and hopelessness have consistently emerged as significant. Empirically documented relationships among these and related variables suggest a causal model. Dukes and Lorch (1989), for example, found that family conflict covaried with feelings of hopelessness in adolescents. Low SES has been linked with feelings of low control (Paltiel, 1987), which relate to feelings of helplessness (Glass & Signer, 1972; Seligman, 1975). In light of the data linking family discord and SES to both hopelessness and suicidality, and unequivocal evidence of the association of hopelessness with suicidality, it seemed plausible that hopelessness mediates the effects of family discord and SES. Thus, of interest in this study was whether hopelessness is a "generative mechanism" (Baron & Kenny, 1986) through which family and sociocultural factors affect suicidal behavior.
METHOD
Subjects
The sample in the present study was drawn from a larger data base (e.g., Spirito, Lewander, Levy, Kurkjian, & Fritz, 1994), consisting of consecutive adolescent suicide attempters who presented at the emergency room or pediatrics ward of a large northeast metropolitan hospital. Adolescents immediately admitted psychiatrically following their attempts were not included. Written consent was obtained by each adolescent and at least one parent or guardian before they participated in the study.
The sample, which included 76 adolescents, ranged in age from 12 to 18 years, with a mean age of 15.48 (SD = 1.16). There were 61 (80%) females and 15 (20%) males. Seventy-five percent of the sample was Caucasian; The remainder were Hispanic (12%), African-American (10%), and Asian-American (3%). Patient socioeconomic status was classified by the Department of Health Census Track data based on street and town addresses (see Measures section for details). Distribution for SES was fairly even: 34% poverty, 24% low, 25% middle, and 17% high. Ninety-two percent of the subjects attempted suicide by drug overdose. Most of the subjects (72%) reported never having previously attempted suicide. The most commonly reported precipitating events for their attempts were problems with parents (36%) and with boyfriends/girlfriends (25%). Eighty-two percent of the sample was admitted to the pediatric ward for medical treatment immediately following the attempt, and 18% were sent home.
Procedure
Suicide attempters were defined as adolescents admitting to any degree of intentional self-inflicted harm. The majority of evaluations were conducted in the patient's room on the pediatric ward after he or she was admitted for medical treatment. The remainder were interviewed in the emergency department or returned within three days of the attempt. All measures were administered to the adolescents without family members present, which is important for validity (Robbins & Alessi, 1985). Patients admitted immediately to a psychiatric hospital were not included in the present study.
Measures
Hopelessness Scale for Children (HSC). The HSC (Kazdin, Rodgas, & Colbus, 1986) is a 17-item true/false questionnaire that measures pessimism about the future. The scale was adapted from the Hopelessness Scale (Beck, Weissman, Lester, & Trexler, 1974b), developed for adults. Kazdin and colleagues reported internal consistency (alpha = .97) and test-retest reliability (r = .52). Construct validity demonstrated a positive correlation with depression (r = .58) and a negative correlation with self-esteem (r = -0.61) and social skills (r = -.39) for child psychiatric inpatients. Spirito, Williams, Stark, and Hart (1988) evaluated the psychometric properties of the HSC with a sample of adolescent suicide attempters. The test-retest reliability after 10 weeks for controls was r = .53, and the split-half reliability for controls and attempters was .75 and .91, respectively. Tests for criterion validity comparing total HSC scores for controls (M = 3.7, SD = 3.2) and attempters (M = 6.7, SD = 4.5) showed a significant difference, t(923) = 6.56, p [less than] .001. Support for HSC validity was also provided by positive correlations with depression (Children's Depression Inventory) and depressive attributional style (Children's Attributional Style Questionnaire). The findings of this study suggest that the HSC is an appropriate measure to be used in research with adolescent suicide attempters.
McMaster Family Assessment Device (FAD). The FAD (Epstein, Baldwin, & Bishop, 1983) assesses family communication patterns, as well as overall health/pathology of the family unit. The questionnaire was designed as a screening instrument of family interactions. It is based on the McMaster Model of Family Functioning (MMFF), which describes structural and organizational properties of families and patterns of transaction among family members. Because only family communication and overall family functioning were assessed in the brief screening battery, only two of the seven subscales, the Communication and the General Functioning scales, were used. The Communication scale focuses on "whether verbal messages are clear with respect to content and direct in the sense that the person spoken to is the person for whom the message is intended" (Epstein et al., 1983, p. 172). The General Functioning scale assesses overall health/pathology of the family. Subjects rate their agreement with how well each item describes their family on a 4-point scale. Psychometric properties of the FAD subscales include adequate internal consistency, with alphas ranging from .75 (Communication) to .92 (General Functioning). Test-retest reliabilities, with a 1-week interval, range from .66 to .76 (Miller, Epstein, & Bishop, 1985). Validity was evaluated by comparing scores from nonclinical and clinical families; nonclinical families had significantly lower (healthier) scores than clinical families, (p [less than] .001). The FAD subscales have low correlations with social desirability, ranging from .06 to .19 (Keitner et al., 1990).
Socioeconomic Status. SES was determined using values obtained through factor and cluster analysis of Census of Population and Housing data by the Office of Data Evaluation at the Rhode Island Department of Health. Very small geographical areas were analyzed on two factors, Wealth and Education. Then, small regions were combined into single clusters only when both level of Wealth and level of Education were equivalent across each region. The level for each cluster was then compared to the state average. Based on this, clusters were labeled on SES as either poverty, low, middle, or high. A one-way analysis of variance determined that the four SES groups differed significantly from one another on Wealth and Education factors. Finally, SES was determined for each subject by identifying which cluster the subject was in, based on street address and city.
Suicidal Ideation Questionnaire (SIQ; Reynolds, 1987). The SIQ is a self-report measure designed to assess the seriousness of suicidal thoughts in adolescents. The scale consists of 30 items rated on a 7-point scale, identifying the frequency with which the thought occurs. The statements include issues such as thoughts of killing oneself, thoughts of death in general, writing wills, and problems suicide would solve. Subjects were instructed to answer the questions in terms of their feelings during the month immediately preceding their attempts. The standardization sample for the SIQ (N = 2180) consisted of a racially and socioeconomically heterogeneous sample of junior and senior high school students. Coefficient alpha internal consistency has been reported as high as [Alpha] = .97. Adequate test-retest reliability, after 4 weeks, was found (r = .72). Reynolds and colleagues (1987) also demonstrated construct validity by high correlations with related constructs such as depression (Reynolds Adolescent Depression Inventory, Beck's Depression Inventory, Children's Depression Inventory), and hopelessness (Beck's Hopelessness Scale).
Suicide Intent Scale (SIS; Beck, Schyler, & Herman, 1974a). The SIS is a 15-item scale for assessing the seriousness of the intent to die during the suicide attempt. The scale includes two sections. The circumstantial or objective section addresses the factual aspects of the attempt, which provide information about the intent of lethality, such as degree of isolation at the time of the attempt, precautions taken against being found, and help sought during or after the attempt. The subjective section addresses retrospectively the attempter's thoughts and feelings at the time of the attempt, such as expectation about the lethality of the attempt and the degree to which the attempt was either impulsive or premeditated. Each item has three possible options, rated from 0 to 2. The overall SIS score is obtained by adding scores for each of the 15 items.
The SIS is administered in an interview format and scored via clinical ratings. Interrater reliability is high (r = .95) (Beck et al., 1974a). Internal consistency, using an odd-even computation, was reported at r = 0.82 (Beck et al., 1974a). To demonstrate construct validity of the circumstantial section of the scale, Beck et al. (1974a) found that fatal attempters (derived from a detailed analysis of case records at the medical examiner's office) differed from nonfatal attempters significantly (p [greater than] .0005) and that the correlation with intent was higher for hopelessness (r = .47) than for depression (r = .26). Reynolds and Eaton (1986) found, in a sample of adolescents and adults, that multiple attempters had significantly higher scores on the subjective section of the SIS than first-time attempters. Several studies have demonstrated the relation of the SIS to hopelessness and depression among adolescents (Brent, 1987; Reuben, Boeck, & Kurzon, 1987).
The SIS was chosen because it measures subjective suicidal intent rather than actual lethality of the attempt. Studies have shown that there is no significant correlation between intent to die and the lethality of the attempt (Plutchik, van Praag, Picard, Conte & Korn, 1989). Evaluating subjective intent of adolescents is also important because it is unclear whether they are cognitively able to assess the lethality of their suicide methods (Rotheram-Borus & Trautman, 1988). Further evidence for the need to measure subjective intent includes studies indicating that repeat attempts are positively correlated with the adolescents' belief that death was a likely result of their attempt, despite the actual medical lethality (Spirito et al., 1994).
RESULTS
Statistical Analyses
According to Baron and Kenny (1986), mediation is established if the following criteria are met: (1) The independent variables affect the mediating variable, (2) the independent variables affect the dependent variable, (3) the mediating variable affects the dependent variable, and (4) the independent variables affect the dependent variable less when the mediator is included in the regression than when it is not included. Thus, to determine whether hopelessness mediated the relation of SES and family dysfunction to suicidal ideation and intent, three regression equations were computed for each dependent variable.
Subjects who were sent home were compared with subjects who were medically admitted. The only measure on which they differed significantly was Family Discord (t = 3.09, p [less than] .01). Those admitted medically had more dysfunctional family scores.
Means, standard deviations, and ranges
for all variables are shown in Table I. Zero-order correlations
for all variables are presented in Table II. The two family scales,
Communication and General Functioning, were highly correlated
(r = .95, p [less than] .001). Given that the means and ranges
of scores on each scale were equivalent, the scale scores were
added to form a composite variable (family dysfunction).
Table I. Means, Standard Deviations (SDs), and Ranges for All
Variables, n = 76
Mean SD Range
SES(a) 2.25 1.11 1 to 4
Family discord 24.18 23.51 12 to 76
Hopelessness 4.92 4.07 0 to 16
Suicidal intent 9.55 5.99 0 to 25
Suicidal ideation 62.08 44.18 3 to 171
a SES = socioeconomic status.
Table II. Zero-Order Correlations Between All Variables, n = 76
Variable SIS(a) Hope SES(a) Family
Hopelessness .20(b)
SES(a) .19(b) -.08
Family dysfunction -.06 .11 -.03
Suicidal ideation -.39(c) .50(c) .03 .03
a SES = socioeconomic status; SIS = Suicide Intent Scale.
b p [less than] .05.
c p [less than] .01.
Results of the regression analyses are shown in Tables III and IV. Family dysfunction and SES failed to predict either hopelessness, suicidal ideation, or suicidal intent (Criteria 1 and 2). Hopelessness predicted both suicide intent and suicidal ideation (Criterion 3). Thus, hopelessness was found to be a significant predictor, but not a mediating variable.
Additional Analyses
During the course of this ongoing project,
some changes in the protocol were made. In particular, the Suicide
Intent Scale was not added until after 26 subjects had completed
the Suicidal Ideation Questionnaire. As an exploratory analysis,
these subjects were added to the sample, comprising a sample of
102 subjects. The model was retested with SIQ as the dependent
variable to determine whether an increase in the power to detect
small or medium-size effects would yield different results. The
76 original subjects were compared with the 26 additional subjects,
using chi-square analyses, on all demographic variables. No significant
differences between the groups were found.
Table III. Standard Multiple Regressions Used to Test the Mediation Model with Suicidal Ideation (SIQ)(a) as the Dependent Variable, n = 76 DV(a) IV(a) Beta [R.sup.2] [R.sup.2] adjusted
HSC(a)
Family .111
SES(a) -.081
Full model .019 .007
SIQ(a)
Family .030
SES(a) .034
Full model .002 .025
SIQ(a)
Family .026
SES(a) .075
HSC(a) .510(b)
Full model .257(b) .226(b)
a HSC = Hopelessness Scale for Children; SIQ = Suicide Ideation
Questionnaire; DV = dependent variable; IV = independent variable;
SES = socioeconomic status.
b p [less than] .001.
Table IV. Standard Multiple Regressions Used to Test the
Mediation Model with Suicidal Intent (SIS)(a) as the Dependent
Variable, n = 76
DV(a) IV(a) Beta [R.sup.2] [R.sup.2] adjusted
HSC(a)
Family .111
SES(a) -.081
Full model .019 .007
SIS(a)
Family -.052
SES(a) .194(b)
Full model .041 .015
SIS(a)
Family -.077
SES(a) .213
HSC(a) .232(b)
Full model .094 .056
a HSC = Hopelessness Scale for Children; SIS = Suicide Intent
Scale; DV = dependent variable; IV = independent variable;
SES = socioeconomic status.
b p [less than] .05.
Results of the three regression analyses
are presented in Table V and Fig. 1. Family dysfunction predicted
hopelessness (Criterion 1); hopelessness predicted suicidal ideation
(Criterion 3); and the effect of family dysfunction decreased
when hopelessness was entered into the regression (Criterion 4).
Family dysfunction, however, did not significantly affect suicidal
ideation directly (Criterion 2). The SES variable did not predict
either hopelessness or suicidal ideation (Criteria 1 and 2).
Table V. Standard Multiple Regressions Used to Test the Mediation Model with Suicidal Ideation (SIQ)(a) as the Dependent Variable, n = 102 DV(a) IV(a) Beta [R.sup.2] [R.sup.2] adjusted
HSC(a)
Family .285(b)
SES(a) .011
Full model .082(b) .063(b)
SIQ(a)
Family .146
SES(a) .032
Full model .023 .003
SIQ(a)
Family .015
SES(a) .027
HSC(a) .457(c)
Full model .220(c) .191(b)
a HSC = Hopelessness Scale for Children; SIQ = Suicide Ideation
Questionnaire; DV = dependent variable; IV = independent variable;
SES = socioeconomic status.
b p [less than] .05.
c p [less than] .001.
DISCUSSION
Family dysfunction, SES, and hopelessness have previously been related either singularly or in combination to suicidal behavior in adolescents. The present study was the first to examine these variables within a sample of adolescent attempters and to test a causal model, explaining their inter-connections.
Results were consistent with past studies, relating hopelessness to suicidal intent and ideation (Robbins & Alessi, 1985; Topol & Reznikoff, 1982; Spirito et al., 1988). Of the three variables tested, hopelessness, the individual or internal variable, had the strongest effect. This result is important in further validating the use of hopelessness as a predictor for both suicidal thoughts and seriousness of attempts. Furthermore, while past studies typically focused on the ability of hopelessness to distinguish suicidal from nonsuicidal adolescents, the predictive ability of hopelessness within a sample of attempters indicates the heterogeneity of attempters. Determining within-group variability among suicide attempters increases the clinical utility of the findings.
Contrary to expectation, higher socioeconomic status related to more serious suicidal intent. The correlation between SIS and SES was statistically significant, and SES significantly predicted SIS, when not controlling for hopelessness. However, when hopelessness was included in the regression, the predictive validity of SES only approached significance. Although previous studies comparing attempters and controls have found that low-SES adolescents are more likely to attempt suicide (Dubow et al., 1989; Garfinkel et al., 1982; Stack, 1982), the present study examined differences within a sample of attempters. Perhaps low-SES youngsters think just as seriously about death and suicide but are not as serious in their intent to die during the actual attempt. Because these youngsters often play a more central role than their more advantaged peers in contributing to the family welfare, such as taking care of younger siblings and working part time, they may feel a greater responsibility to live.
A possible methodological problem is that perhaps the determination of SES was too centered on parental influences. As in most studies, SES was determined solely on the basis of parental factors, such as income and education. Perhaps economic status should be measured from the perspective of the youngster. For example, how does the youngster feel he or she compares economically to peers or to youth nationally? How hopeful is the youngster about his or her financial future? Does the youngster feel he or she has adequate clothing and other necessities? Does the youngster have to deal with violence, which often accompanies living in poor areas? Bringing the questions to a level that impacts the adolescents directly may give a more accurate picture of economic status.
In the larger sample tested using only suicidal ideation as the dependent variable, family discord predicted hopelessness and hopelessness predicted suicidal ideation. The mediating role of hopelessness was not fully supported, according to Baron and Kenny's (1986) criteria, because the family variable did not significantly predict suicidal ideation (Criterion 2). Others would argue, however, that the present findings are consistent with mediation in the traditional sense, inasmuch as all of the effects of family dysfunction on suicidal ideation are indirect (Cohen & Cohen, 1983). Yet, in light of the nonsignificant relationship between family dysfunction and suicidal ideation, hopelessness may be mediating a facet or correlate of the family variable that was not directly examined in this study (R. M. Baron & D. A. Kenny, personal communication, 1993). The mediation found in the larger sample was weak and needs further validation.
It is possible that methodological factors obscured the relation of family dysfunction to suicidal ideation and intent. Most previous studies compared attempters with nonattempters. The current data were drawn from a group of attempters who were not admitted to psychiatric hospitals following their attempts, partly because of their families' ability to monitor the children. Therefore, the family scores were probably skewed toward the less pathological end of the scale. In addition, only two of the seven subscales from the FAD were used for this study. Using all subscales may be necessary to obtain an accurate evaluation. The FAD also does not assess some important family factors hypothesized to relate to adolescent suicidal behavior, such as enmeshment and rigidity.
![[Expanded Picture]](file:///C:%5CWINDOWS%5CDESKTOP%5Cuni%5CMarked%20Articles_files%5Cimg(17).jpg)
![[Expanded Picture]](file:///C:%5CWINDOWS%5CDESKTOP%5Cuni%5CMarked%20Articles_files%5Cimg(18).jpg)
Another methodological constraint was the inability
to use more than one measure to test each construct, which is
ideal when model-testing. For example, including additional measures
that were not self-report would have been preferable. However,
given that the data were obtained from a clinical sample of adolescents
presenting at an emergency room following a suicide attempt, we
did not have the luxury of multiple sources. Each measure used,
however, was considered the most important for its level; all
measures were chosen carefully, based on clinical experience,
and are well-grounded empirically.
A difficulty of suicide research in general involves the timing of the assessment. Conducting an evaluation after the suicide attempt may not accurately reflect feelings prior to or at the time of the attempt (Strang & Orlofsky, 1990). For example, adolescents may perceive their families more negatively before their attempts than after, especially if their suicidal behavior occasions desirable reactions from their families. Direct observations of family interactions at the time of crisis may be more valid than self-report measures and a better index of family-level functioning (Keeney, 1983).
In addition, causal relations must always be interpreted cautiously when evaluating data from a cross-sectional study. For example, questions could be raised regarding the use of suicidal ideation as a dependent variable because it measures feelings prior to the attempt, and it is difficult to know the time frame used by the adolescents while answering the hopelessness and family dysfunction measures. Issues of timing underscore the need for longitudinal studies.
![[Expanded Picture]](file:///C:%5CWINDOWS%5CDESKTOP%5Cuni%5CMarked%20Articles_files%5Cimg(19).jpg)
This study was the first attempt at integrating
variables at multiple levels to form a causal model. The results
confirm the importance of hopelessness as a factor in adolescent
suicide and indicate that hopelessness varies within a sample
of attempters. The simultaneous analysis of multiple variables
at different levels of analysis indicates that hopelessness (the
internal variable) is the strongest predictor among them.
Additional analyses suggested that hopelessness may mediate the effects of family dysfunction on suicidal ideation. Further research that considers the limitations discussed earlier may help elucidate possible mediational models in this area. Explaining how various factors at different levels of analysis relate to one another is necessary to develop more effective prevention and intervention strategies.
REFERENCES
Asarnow, J. R., & Guthrie, D. (1989). Suicidal behavior, depression, and hopelessness in child psychiatric inpatients: A replication and extension. Journal of Clinical Child Psychology, 18(2), 129-136.
![[Expanded Picture]](file:///C:%5CWINDOWS%5CDESKTOP%5Cuni%5CMarked%20Articles_files%5Cimg(20).jpg)
Baron, R. M., & Kenny, D. A. (1986).
The moderator-mediator variable distinction in social psychological
research: Conceptual, strategic, and statistical considerations.
Journal of Personality and Social Psychology, 31, 1173-1182.
Beck, A. T., Brown, G., Berchick, R. J., Stewart, B. L., & Steer, R. A. (1990). Relationship between hopelessness and ultimate suicide: A replication with psychiatric outpatients. American Journal of Psychiatry, 147, 190-195.
![[Expanded Picture]](file:///C:%5CWINDOWS%5CDESKTOP%5Cuni%5CMarked%20Articles_files%5Cimg(21).jpg)
Beck, A. T., Schyler, D., & Herman, L.
(1974a). Development of Suicidal Intent Scales. In A. T. Beck,
H. L. P. Resnick, & D. T. Lettieri (Eds.), The prediction
of suicide (pp. 45-46). Bowie, MD: Charles Press.
Beck, A. T., Steer, R. A., Kovacs, M., & Garrison, B. (1985). Hopelessness and eventual suicide: A 10-year prospective study of patients hospitalized with suicidal ideation. American Journal of Psychiatry, 142, 559-563.
![[Expanded Picture]](file:///C:%5CWINDOWS%5CDESKTOP%5Cuni%5CMarked%20Articles_files%5Cimg(22).jpg){kind=small}
![[Expanded Picture]](file:///C:%5CWINDOWS%5CDESKTOP%5Cuni%5CMarked%20Articles_files%5Cimg(23).jpg)
Beck, A. T., Weissman, A., Lester, D., & Trexler,
L. (1974b). Measurement of pessimism: The Hopelessness Scale.
Journal of Consulting and Clinical Psychology, 42, 861-865.
Brent, D. A. (1987). Correlates of the medical lethality of suicide attempts in children and adolescents. Journal of the American Academy of Child and Adolescent Psychiatry, 26, 87-89.
Brent, D. A., Kolko, D. J., Goldstein, C., Allan, M., & Brown, R. (1989, October). Cognitive distortion, familial stress, and suicidality in adolescent inpatients. Poster presented at the Annual Meeting of the American Academy of Child and Adolescent Psychiatry, New York.
Bronfenbrenner, U. (1977). Toward an experimental ecology of human development. American Psychologist, 32, 513-529.
Carlson, G. A., & Cantwell, D. P. (1982). Suicidal behavior and depression in children and adolescents. Journal of the American Academy of Child Psychiatry, 21, 361-368.
Cohen, J., & Cohen, P. (1983). Applied multiple regression/correlation analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ: Erlbaum.
Dubow, E. F., Kausch, D. F., Blum, M. C., Reed, J., & Bush, E. (1989). Correlates of suicidal ideation and attempts in a community sample of junior high and high school students. Journal of Clinical Child Psychology, 18, 158-166.
Dukes, R. L., & Lorch, B. (1989). The effects of school, family, self-concept, and deviant behavior on adolescent suicide ideation. Journal of Adolescence, 12, 239-251.
Epstein, N. B., Baldwin, L. M., & Bishop, D. S. (1983). The McMaster family assessment scale. Journal of Marital and Family Therapy, 9, 171-180.
Garfinkel, B. D., Froese, A., & Hood, J. (1982). Suicide attempts in children and adolescents. American Journal of Psychiatry, 139, 1257-1261.
Glass, D. C., & Singer, J. E. (1972). Urban Stress: Experiments on Noise and Social Stressors. New York: Academic Press.
Hawton, K., O'Grady, J., Osborn, M., & Cole, D. (1982). Adolescents who take overdoses: Their characteristics, problems and contacts with helping agencies. British Journal of Psychiatry, 140, 118-123.
Kazdin, A. E., Rodgas, A., & Colbus, D. (1986). The Hopelessness Scale for Children: Psychometric characteristics and concurrent validity. Journal of Consulting and Clinical Psychology, 54, 241-245.
Keeney, B. P. (1983). Ecological assessment. In J. C. Hansen & B. P. Keeney (Eds.), Diagnosis and assessment in family therapy (pp. 156-170). Rockville, MD: Aspen.
Keitner, G. I., Ryan, C. E., Miller, I. W., Epstein, N. B., Bishop, D. S., & Norman, W. H. (1990). Family functioning, social adjustment, and recurrence of suicidality. Psychiatry, 53, 17-30.
Mattsson, A., Seese, L. R., & Hawkins, J. W. (1969). Suicidal behavior as a child psychiatric emergency. Archives of General Psychiatry, 20, 100-109.
McIntire, M., & Angle, C. (1980). Suicide Attempts in children and youth. Hagertown: Harper and Row.
Miller, I. W., Epstein, N. B., & Bishop, D. S. (1985), The McMaster family assessment device: Reliability and validity. Journal of Marital and Family Therapy, 11, 345-356.
Miller, K. E., King, C. A., Shain, B. N., & Naylor, M. W. (1992). Suicidal adolescents' perceptions of their family environment. Suicide and Life-Threatening Behavior, 22, 226-239.
Paltiel, F. L. (1987). Is being poor a mental health hazard? Special issue: Women, health, and poverty. Women and Health, 12(3-4), 189-211.
Plutchik, R., van Pragg, H. M., Picard, S., Conte, H. R., & Korn, M. (1989). Is there a relation between the seriousness of suicidal intent and the lethality of the suicide attempt? Psychiatry Research, 27, 71-79.
Reuben, N., Boeck, M., & Kurzon, M. (1987, March). Relation between suicide intent, depression and disposition in adolescent suicide attempters. Paper presented at the Society of Adolescent Medical Conference, Seattle.
Reynolds, P., & Eaton, P. (1986). Multiple attempters of suicide presenting at an emergency department. Canadian Journal of Psychiatry, 31, 328-330.
Reynolds, W. R. (1987). Suicide Ideation Questionnaire. Odessa, FL: Psychological Assessment Resources.
Robbins, D. R., & Alessi, N. E. (1985). Depressive symptoms and suicidal behavior in adolescents. American Journal of Psychiatry, 142, 588-592.
Rotheram-Borus, M. J., & Trautman, P. D. (1988). Hopelessness, depression, and suicidal intent among adolescent suicide attempters. Journal of the American Academy of Child and Adolescent Psychiatry, 27, 700-704.
Sabbath, J. C. (1969). The suicidal adolescent-the expendable child. Journal of the American Academy of Child Psychiatry, 8, 272-289.
Sanborn, D. E. Sanborn, C. J., & Cimbolic, P. (1973). Two years of suicide: A study of adolescent suicide in New Hampshire. Child Psychiatry and Human Development, 3, 234-242.
Seligman, M. E. P. (1975). Helplessness: On Depression, Development, and Death. San Francisco: Freeman.
Spirito, A., Brown, L., Overholser, J., & Fritz, G. (1989). Attempted suicide in adolescence: A review and critique of the literature. Clinical Psychology Review, 9, 335-363.
Spirito, A., Lewander, W., Levy, S., Kurkjian, J., & Fritz, G. (1994). Factors related to treatment compliance among adolescent suicide attempters evaluated in an emergency department. Pediatric Emergency Care, 10, 6-12.
Spirito, A., Stark, L. J., Fristad, M., Hart, K., & Owens-Stively, J. (1987). Adolescent suicide attempters hospitalized on a general pediatrics floor. Journal of Pediatric Psychology, 12, 171-189.
Spirito, A., Williams, C. A., Stark, L. J., & Hart, K. J. (1988). The Hopelessness Scale for children: Psychometric properties with normal and emotionally disturbed adolescents. Journal of Abnormal Child Psychology, 16, 445-458.
Stack, S. (1982). Suicide: A decade review of the sociological literature. Deviant Behavior: An Inter-Disciplinary Journal, 4, 41-66.
Strang, S. P., & Orlofsky, J. L. (1990). Factors underlying suicidal ideation among college students: A test of Teicher and Jacobs' model. Journal of Adolescence, 13, 39-52.
Topol, P., & Reznikoff, M. (1982). Perceived peer and family relationships, hopelessness and locus of control as factors in adolescent suicide attempts. Suicide and Life-Threatening Behavior, 12, 141-150.
Wenz, F. V. (1979). Sociological correlates of alienation among adolescent suicide attempts. Adolescence, 14, 19-29.
Williams, C., & Lyons, C. M. (1976). Family interaction and adolescent suicidal behavior: A preliminary investigation. Australian and New Zealand Journal of Psychiatry, 10, 243-252.
![[Expanded Picture]](file:///C:%5CWINDOWS%5CDESKTOP%5Cuni%5CMarked%20Articles_files%5Cimg(24).jpg)
![[Expanded Picture]](file:///C:%5CWINDOWS%5CDESKTOP%5Cuni%5CMarked%20Articles_files%5Cimg(25).jpg)