Suicidal Ideation among Youths at Risk of School Dropout: Impact of Student Demographics, Stressors, and Academic Self-Concept

abstract

In the United States, suicidal ideation is an issue for high school–age youths. Research supports that youths who have learning difficulties and who are at risk of high school dropout are at greater risk for suicidal ideation. Although alternative high schools address both student academics and emotional health, they are underused, nonclinical settings for understanding and addressing suicidal ideation. This study aimed to examine the impact of student identity, external stressors, and academic self-concept on suicidal ideation among sexual and ethnic minority and underserved students enrolled in an alternative education public high school. The student sample (N = 103) completed a onetime survey comprised of the Suicidal Ideation Questionnaire-Junior, the Cultural Assessment of Risk of Suicide, the Coddington Life Events Scale for Adolescents, and the Piers Harris 2. Results of hierarchical linear regression indicated that identifying as nonheterosexual and experiencing discrimination were associated with greater student reporting of suicidal ideation. Identifying as Latino and other non-White was associated with lower reporting of suicidal ideation. Controlling for student demographics and external stressors, positive academic self-concept was associated with a lower reporting of suicidal ideation. Findings have future implications for health social work, suicide prevention and intervention, and education policy.

Suicidal ideation is an issue for high school–age youths across the U.S. education system (Kann et al., 2018). According to the Youth Risk Behavior Survey (YRBS), 17.2 percent of high school students sampled in the United States reported seriously considering suicide in the past 12 months (Kann et al., 2018). The prevalence of suicidal ideation varies by student demographics, and the risk of suicide may be elevated among gender and sexual and ethnic minority individuals (Chu, Goldblum, Floyd, & Bongar, 2010). For example, boys are more likely to die by suicide (Stone et al., 2018), whereas girls are more likely to report suicidal ideation and suicide attempts (Kann et al., 2018). Since 2007, the YRBS has documented significant increases in suicidal ideation among high school students who identified as White or Latino (Centers for Disease Control and Prevention, 2017). In addition, sexual minority individuals are at greater risk of suicidal ideation than heterosexual individuals (Kann et al., 2018; Russell & Joyner, 2001). Yet, there are few existing studies on the implications of ethnic and sexual minority identity and environmental stressors on youth suicidal ideation (Chu et al., 2010).

Students at risk of dropping out are also at risk of suicidal ideation (Daniel et al., 2006). Studies have documented that suicidal ideation is associated with truancy (Epstein et al., 2019) and failing coursework (Kosidou, Dalman, Fredlund, & Magnusson, 2014). Additional research suggests that students who struggle with reading and writing are at greater risk of experiencing suicidal ideation (Daniel et al., 2006).

Alternative education high schools are unique yet underrepresented settings to study suicidal ideation. Alternative education schools are public or private institutions that use nontraditional education techniques to address issues specific to students at risk of dropping out, such as truancy, poor grades, and mental health issues (Franklin, Kelly, & Szlyk, 2016; Grunbaum, Lowry, & Kann, 2001). Students often come from families of lower socioeconomic status, may identify as ethnic minorities, and are often disproportionately burdened with mental health and behavioral issues (Lagana-Riordan et al., 2011). Research suggests that alternative student populations may report higher rates of engagement in risky behaviors, such as substance use, physical fighting, and suicide attempts, than their traditional school peers (Grunbaum et al., 2001; Johnson, McMorris, & Kubik, 2013; Johnson & Taliaferro, 2012).

Students’ perceptions of the world and of themselves are shaped by experiences and feedback from their environment (Bong & Skaalvik, 2003). Students who have a more positive perception of themselves and their communities are often healthier (Patton et al., 2011) and perform well academically (Bong & Skaalvik, 2003). In contrast, students who display fatalistic and pessimistic thinking about themselves and the future are at risk for emotional issues, such as suicidality (Piña-Watson & Abraído-Lanza, 2017).

Two factors suggest that alternative high school youths have greater negative perceptions of themselves and of the world as compared with traditional high school youths. Ethnic and sexual minority students and students of lower socioeconomic backgrounds, who often make up alternative schools (Lagana-Riordan et al., 2011), may experience external stressors, including discrimination and adverse life events, that promote a bleaker outlook on their personal safety and future success (Dohrenwend, 1959; Jamieson & Romer, 2008; Piña-Watson & Abraído-Lanza, 2017). Students who feel fatalistic about their lives are more likely to drop out of school (Jamieson & Romer, 2008) or have poor school performance (Thompson, Mazza, Herting, Randell, & Eggert, 2005).

Although fatalistic and pessimistic thinking are foundational concepts in suicidology (Beck, Brown, Berchick, Stewart, & Steer, 1990; Dohrenwend, 1959; Krajniak, Miranda, & Wheeler, 2013), no known study has examined the relationship between academic self-concept and suicidal ideation among students enrolled at an alternative high school. Academic self-concept encompasses a student’s perception of how others view their academic abilities and their mastery of certain skills (Bong & Skaalvik, 2003). This construct extends the suicidology literature beyond the importance of belonging (Joiner, 2006; Van Orden et al., 2010) and grade point average (Sörberg Wallin et al., 2018) and focuses on the potential protection of perceiving oneself as a capable student. As the concepts of empowerment and self-efficacy are deeply rooted in social work (Parsons & East, 2013), the study of how academic self-concept relates to suicidal ideation may also help to leverage the profession as a leader in youth suicide prevention and intervention.

Informed by the interdisciplinary literature, this study intended to examine the impact of identity, external stressors such as discrimination and life events, and academic self-concept on suicidal ideation among a sample of sexual and ethnic minority and underserved students at an alternative school. The study aimed to explore how student self-reported data informed the following three hypotheses:

1. A minority identity (determined by having a self-reported ethnicity or sexual orientation that is considered a minority population in the United States) would be associated with a higher level of suicidal ideation.

2. External stressors (perceived discrimination and adverse life events) would be associated with a higher level of suicidal ideation.

3. Positive academic self-perception would be associated with a lower level of suicidal ideation.

Method

Study Design

This study used quantitative data from a mixed-methods sequential design project that aimed to examine and understand the relationships among suicidal ideation and behavior, academic progress, adverse life events, and resiliency within an alternative high school population. Quantitative data were collected through a questionnaire at one time point, which included several standardized measurements (see “Study Instruments”). The author and research assistant administered the questionnaire to students individually on school campus. The questionnaire included the domains of suicidal ideation, student demographics, recent life events, perceived discrimination, and academic self-concept. Students completed the questionnaire with paper and pencil.

Participants

The students were from one alternative high school located in the southwestern United States that uses a solution-focused approach to teaching. The high school is the only alternative education, nondisciplinary program for a public school district of 85,000 students. The students at the alternative high school are ethnically diverse (46.9 percent identify as Latino/Hispanic, followed by 42.2 percent White, and 4.8 percent Black), and 34 percent of students qualify as economically disadvantaged. High school administrators and parents identified students as at risk of dropping out. In this article, I use the terms “White,” “Black,” and “Latino” to refer to socially, politically, and culturally constructed ethnic group identities and recognize that these groups, like all other ethnic groups in the United States, are heterogeneous (Graves, 2001; Sussman, 2014; Zuberi, 2001; Zuberi & Bonilla-Silva, 2008).

From a student body of approximately 200 students, 103 students agreed to participate in the study. Participants’ ages ranged from 15 to 20 years. The study was inclusive of all genders and ethnic groups. Only students who received written consent from their parents or guardians and assented to study participation were in the final sample. Students age 18 years or older consented for themselves. The University of Texas at Austin and the Washington University in St. Louis’s institutional review boards and the school district’s research review board granted approval for the project.

Study Instruments

Suicidal Ideation

The Suicidal Ideation Questionnaire-Junior (SIQ-JR) (Reynolds & Mazza, 1999) was used to measure student suicidal ideation. The SIQ-JR has been used with high school–age participants and young adults (Jacobson, Marrocco, Kleinman, & Gould, 2011; Labouliere, Kleinman, & Gould, 2015; Thompson & Eggert, 1999). The 15-item questionnaire asked students to respond to symptoms over the past month and to assess thoughts about suicide (for example, “I wished I were dead” and “I thought that no cared if I lived or died”). Items are scored on a seven-point scale ranging from 0 to 6, with greater scores reflecting greater frequencies of suicidal ideation severity. Total scores of 31 and greater suggest elevated risk of suicide (King, O’Mara, Hayward, & Cunningham, 2009). Reliability of the SIQ-JR is high, ranging from .91 to .96 (Reynolds & Mazza, 1999) for internal consistency (Reynolds & Mazza, 1999) and from .87 to .93 for test–retest reliability (Klomek et al., 2012). The SIQ-JR has demonstrated criterion validity (Reynolds & Mazza, 1999) and construct validity in community (Reynolds & Mazza, 1999) and clinical samples (King, Hill, Naylor, Evans, & Shain, 1993). For the present study, Cronbach’s alpha was .94.

Perceived Discrimination

The Cultural Assessment of Risk of Suicide (Chu et al., 2013) measure assesses for cultural risk factors of suicide among ethnic and sexual minority individuals. Items were selected from the domain of nonspecific minority stress (perceived discrimination and stigmatization). The subscale has four items (for example, “People treat me differently” and “I believe the world is a dangerous place”). Items are rated on a six-point Likert scale, with 1 = strongly disagree, 2 = moderately disagree, 3 = slightly disagree, 4 = slightly agree, 5 = moderately agree, and 6 = strongly agree). Cronbach’s alpha in the study was .76.

Recent Life Events

Recent life events were measured using the Coddington Life Events Scale for Adolescents (CLES-A) (Coddington, 1972). The full CLES-A measures adolescent exposure to life stressors within the past year. The CLES-A was developed for youths ages 13 to 19 years and asked respondents to rate the number of times a stressor occurred in the past year and how long ago. The full scale consists of 50 items (for example, “death of a parent” and “breaking up with a boyfriend/girlfriend”). The frequency of individual negative life event is multiplied by established weights, reflecting an event’s severity and impact on an individual’s life, to produce a total negative life change unit (LCU) score. The LCU is generated cumulatively after each time interval (zero to three months, four to six months, seven to nine months, 10 to 12 months). The CLES-A has been used in both clinical and community samples. The CLES-A has demonstrated test–retest reliability, interrater reliability, and validity among adolescents and high school students (Mental Health Systems, Inc., 1999). Four items were removed for this study because of internal review board requirements regarding survey content and illegal student behavior.

Academic Self-Concept

Academic self-concept was measured using the Intellectual and School Status (INT) subscale of the Piers Harris 2 (Piers & Herzberg, 2002). The INT has 16 items and requires a yes or no response to the statements (for example, “My friends like my ideas” and “I forget what I learn”). The Piers Harris 2 is a self-report measure that provides a holistic view of a child’s self-concept and is applicable for ages seven through 18 years. The full measure has six domains: (1) physical appearance and attributes, (2) freedom from anxiety, (3) intellectual and school status, (4) behavioral adjustment, (5) happiness and satisfaction, and (6) popularity. The total score and the scores of the six domain scales showed strong interscale correlations ranging from .84 to .73 (Piers & Herzberg, 2002). Cronbach’s alpha for this study was .70.

Demographics

Students reported preferred gendered identity (versus biological sex), ethnic identity (White, Black, Native American/Native Alaskan, and so on), Latino or non-Latino identity, sexual identity, age, and length of enrollment at the alternative high school.

Data Analysis

Due to a data collection error, an item from the SIQ-JR was missing from 49 of the surveys. The scale from those cases was prorated, as this strategy increases comparability with the rest of the sample and with other studies using the SIQ-JR. The potential score for the cases using the 14-item scale ranged from 0 = no suicidal ideation to 84 = severely high suicidal ideation. To ensure the quality of the responses using the 14-item scale, correlations between total scores from the 14-item prorated scale and those from the 15-item scale have been assessed using a subsample of the available data set and the correlation coefficient high (r = .99, p < .001). This prorating strategy is regarded as a best practice for addressing missing items from data collection (Haley, Lamonde, Han, Narramore, & Schonwetter, 2001; Jang, Mortimer, Haley, & Graves, 2004).

Considering the constant occurrence of life-changing events that is characteristic of the alternative high school population (that is, housing instability, familial separation, interpersonal issues), I decided to only include the LCU from zero to three months from the CLES-A. As mentioned in the Method section, I was instructed by the school district to remove four items from the CLES-A prior to data collection, as these items could reflect students’ illegal activity (that is, underage drinking, minor crimes). Despite these circumstances, the construct was measured in a meaningful way, as the CLES-A measures relative frequency and impact of events within a youth sample and there can be large variation in LCU scores across youth subgroups (Coddington, 1972). Recent studies with adolescent samples have documented adjustments to the measure to best suit their population and to align with the time frame of the other scales in their surveys (Gould et al., 2018; Rious & Cunningham, 2018). Literature also supports examining a more recent period of life events in relation to assessment for suicidal ideation and behaviors (Bagge, Glenn, & Lee, 2013; Stanley et al., 2013).

After data preparation, assumptions for homoscedasticity, normality, and multicollinearity for multivariate regression were tested (Pituch & Stevens, 2016). Missing data for coefficients and the dependent variable were under 10 percent (Bennett, 2001). Student demographic characteristics were summarized with descriptive statistics. Correlation analyses demonstrated the strength and direction of associations between the demographic characteristics, the external stressor variables, the academic self-efficacy variable, and the dependent variable of suicidal ideation. Predictors of suicidal ideation were determined using hierarchical linear regression. The first model included only student demographic characteristics. To test the impact of risk, the second model added the two external stressor variables (perceived discrimination and recent life-changing life events). The third and final model included academic self-efficacy to test the protective impact of this variable when controlling for student characteristics and external stressors. Analyses were performed using Stata, release 15 (StataCorp, 2017).

Results

Descriptive Characteristics

Slightly over half of the participants were female (51.5 percent, n = 53) and were age 18 or older (18 to 20 years old) (51.0 percent, n = 52) (see Table 1). Approximately 35 percent (n = 36) of participants identified ethnically as non-Latino and White, 27 percent (n = 27) of participants identified as Latinx and White, and 30 percent (n = 31) identified as Latino or non-White. Sixty percent (n = 61) of students identified as heterosexual. Fifty-two percent (n = 54) of students lived in a two-parent household, and 30 percent (n = 31) of students reported that they supplement their family’s income. The mean score for suicidal ideation was 16.11 (SD = 13.30), with a range of 0 to 60. The mean score for perceived discrimination was 7.82 (SD = 5.05), with a range of 0 to 24. The mean of the CLES-A (the past zero to three months) was 250.90 (SD = 178.99), with a range from 0 to 1,717. The mean academic self-concept score was 43.48 (SD = 6.96), with a range of 0 to 65.

Table 1:

Descriptive Characteristics of Alternative School Students (N  = 103)

Demographic Characteristic	M (SD)	n (%)
Gender
Male		47 (45.6)
Female		53 (51.5)
Other		3 (2.9)
Ethnicity
Non-Latino White		36 (34.3)
Latino White		27 (26.5)
Latino and other non-White		31 (30.4)
Sexuality
Heterosexual		61 (59.8)
Nonheterosexual		41 (40.2)
Age of students
18 years old+		52 (51.0)
Under 18 years old		50 (49.0)
Household characteristics
Student lives in two-parent household		54 (52.0)
Student supplements family income		31 (30.0)
External stressors
Perceived discrimination (CARS; 0–24)	7.82 (5.05)
Life event units, 0–3 months (CLES-A; 0–1,717)	250.90 (178.99)
Resources
Academic self-concept (INT; 0–65)	43.48 (6.96)
Suicide risk
Suicidal ideation (SIQ-JR; 0–60)	16.11 (13.30)

Notes: Nine participants did not report ethnicity, one did not report sexuality, one did not report age, and one did not report perceived discrimination. CARS = Cultural Assessment of Risk of Suicide; CLES-A = Coddington Life Events Scale for Adolescents; INT = Intellectual and School Status subscale; SIQ-JR = Suicidal Ideation Questionnaire-Junior.

Bivariate Correlation Analysis

Bivariate correlations demonstrate significant positive associations between suicidal ideation and student characteristics and the psychosocial domains. The variable that had the strongest significant association with suicidal ideation was student perceived discrimination (r = .415, p < .001). Identifying as nonheterosexual was strongly associated with greater reporting of suicidal ideation (r = .373, p < .001). Being 18 years or older at the time of study enrollment (r = –.024, p < .05) was associated with lower reporting of suicidal ideation. No other variables had significant associations with suicidal ideation. See the Appendix for the corresponding table.

Regression Analysis

In the first model (see Table 2), student demographic variables explained approximately 28 percent of the variance of suicidal ideation scores. Identifying as a nonheterosexual youth was associated with nearly a 10-point increase in suicidal ideation; identifying as female was associated with a six-point decrease in suicidal ideation. Identifying ethnically as Latino and other non-White was associated with nearly a nine-point decrease in suicidal ideation. In the second model, the inclusion of external stressors (perceived discrimination and recent life-changing events) added approximately 8 percent of the explained variance. A one-point increase in perceived discrimination was associated with an approximately 0.7-point increase in suicidal ideation, resulting in a total variance explained of 37 percent. In the third model, the addition of academic self-concept explained approximately 6 percent of the variance, for a total of 43 percent of variance explained. Reporting a one-point increase in academic self-concept was associated with approximately a 0.5 decrease in suicidal ideation. The statistically significant effect on suicidal ideation (b = –.54, p < .01) suggests that, while controlling for student characteristics and the presence of external stressors, positive academic self-concept was associated with lower rates of student suicidal ideation.

Table 2:

Direct Effects of Student Characteristics and Experiences on Suicidal Ideation

	Model 1		Model 2		Model 3
Demographic Characteristic	B	SE	B	SE	B	SE
Gender
Male (reference)
Female	–6.02*	2.78	–4.53	2.71	–6.08*	2.64
Ethnicity
Non-Latino White  (reference)
Latino White	–5.47	3.42	–3.96	3.39	–2.76	3.16
Latino and other  non-White	–7.44*	3.35	–7.13*	3.19	–8.63**	3.08
Sexuality
Heterosexual (reference)
Nonheterosexual	13.56***	2.86	10.05**	2.94	9.71**	2.81
Age
18 years+	–2.52	2.84	–2.85	2.84	–3.86	2.61
External stressors
Perceived discrimination			.799**	.284	.689*	.273
Life events			–.006	.008	.009	.008
Resources
Academic self-concept					–.537**	.181
ΔR2	.276***		.089**		.061*
Overall R2	.276***		.371***		.433***

^* p < .05. **p < .01. ***p < .001.

Discussion

Findings extend the growing literature on the risk of suicidal ideation among sexual minority youths (Russell & Joyner, 2001) and the role of perceived discrimination (Chu et al., 2010; Gomez, Miranda, & Polanco, 2011) to include students enrolled in alternative high schools. Findings also suggest that positive academic self-concept may decrease vulnerability to external stressors and suicidal thinking. Although this study underscores that schools are an important setting for youth health promotion, the implications highlight gaps relevant to health social work, suicide prevention and intervention, and education policy.

Current state and federal policies often do not reflect the relationship among students’ personal lives, academic hardships, and mental health. The Every Student Succeeds Act (ESSA) (U.S. Department of Education, n.d.) aims to increase academic opportunities and resources to lower-income and youth populations at risk of dropping out, yet the law does not directly delineate how a positive academic experience can prevent mental health issues, specifically suicidal ideation. Although funding from the ESSA can be used toward suicide prevention and intervention, it is unclear how states have accessed and used these resources for their own schools’ suicide-related trainings and programs. As more states mandate public schools to have protocols for suicide prevention and intervention (Education Commission of the States, 2020), these new policies also do not take into account the complex psychosocial issues that put students at risk for suicidal ideation.

Social–emotional learning (SEL) is a process that simultaneously integrates interpersonal, emotional intelligence, and goal-oriented techniques to holistically address students’ wellness and academic success (Collaborative for Academic, Social, and Emotional Learning [CASEL], 2019). Alternative schools use evidence-based approaches under the umbrella of SEL to inform teaching and school climate (Szlyk, 2018). CASEL (2019) is an organization that supports schools in implementing SEL curricula and guides schools in using ESSA funding to meet their SEL goals for their own campuses. Although most SEL programs emphasize academic outcomes and classroom behaviors (Durlak, Weissberg, Dymnicki, Taylor, & Schellinger, 2011), health social workers (HSWs) and researchers may partner with alternative high schools to additionally track the impact of SEL programs on suicidal ideation and related outcomes.

Future funding opportunities and mandated school policies should integrate academic success and mental health initiatives. Too often, school policies and public health protocols are piecemeal; streamlined and comprehensive initiatives that address various domains of student well-being, such as SEL programs, may be more sustainable, affordable, and proactive for schools in preventing a spectrum of youth issues (Lyon & Bruns, 2019). HSWs are primed to understand the intersections of identity, life stressors, academic difficulties, and suicidal ideation, and the importance of student empowerment. Therefore, HSWs should partner with state and local-level departments of education to develop school suicide prevention and intervention policies and protocols that holistically evaluate and support the psychosocial domains of students’ identity, and both academic and home lives.

Last, it is imperative that HSWs collaborate with high-risk schools, such as alternative programs, to identify and implement practical and feasible strategies for suicide prevention and intervention. This study’s findings suggest that metrics of student perceptions of mastery and academic self-concept may be one strategy to measure resiliency against external stressors and growth in student self-perception (Bong & Skaalvik, 2003). Social workers may also encourage school administrators and staff to foster student resiliency through exploration of special interests and after-school clubs. HSWs should advocate for suicide prevention and intervention policies that are inclusive of diverse student demographics and acknowledge youths who are most vulnerable, such as lesbian, gay, bisexual, transgender, and queer students (Kann et al., 2018; Russell & Joyner, 2001).

Limitations

The present study was conducted at one alternative high school. Students were self-selected; data were also student reported. Institutional review board and school district limitations restricted the study to only examine suicidal ideation. As mentioned, one item on the SIQ-JR was prorated and four items on the CLES-A were removed to comply with school district policies for data collection. The study is also limited by its small sample size and cross-sectional design; thus, causality cannot be drawn from the findings. Future researchers may reproduce this study with a larger alternative student population and with a comparison sample at a traditional high school that measures outcomes over time of school enrollment. Due to the aforementioned limitations, measure-specific scores and overall outcomes may not be generalizable to other alternative high school populations.

Conclusion

This study’s results aim to support greater understanding of suicidal ideation among youths who attend an alternative high school and are at risk of dropping out. This study can demonstrate to HSWs, public health interventionists, and policymakers the need to address the experience of youths especially vulnerable to suicide and how positive academic self-concept may mitigate the risks of marginalized students. Future public health funding opportunities and mandated school policies should streamline academic and mental health initiatives rather than take a piecemeal approach to suicide prevention. HSWs may also wish to partner with schools that serve students at high risk of suicide, such as alternative schools, to additionally track the impact of holistic prevention initiatives, such as SEL, on student suicidal ideation and related outcomes.

 

 

Appendix: Bivariate Correlations among Study Variables

Measure	1	2	3	4	5	6	7	8	9	10
1.Female	—
2. Non-Latino	.115	—
White
3. Latino White	.160	–.500***	—
4. Other non-White	–.273**	–.553***	–.445***	—
5. Nonheterosexual	.251*	–.139	.005	.139	—
6. Adult student	.070	–.372**	.259*	.137	–.149	—
7. Perceived
discrimination	–.123	.061	–.197	.127	.351**	–.176	—
8. Life events	–.144	.022	–.079	–.053	.011	.026	.289	—
9. Academic self-concept	–.098	.012	.229*	–.232*	–.137	–.070	–.119	–.013	—
10. Suicidal ideation	–.032	.154	–.098	–.064	.373***	–.024*	.415***	.083	–.175	—

^* p < .05. **p < .01. ***p < .001.

 

Hannah S. Szlyk, PhD, LCSW, is assistant professor, Rutgers University School of Social Work, New Brunswick, NJ 08901; e-mail: hannah.szlyk@rutgers.edu. The study was funded by the Hogg Foundation’s Frances Fowler Memorial Award for Mental Health Dissertation Research and the Harry E. and Bernice M. Moore Fellowship for doctoral research, and the P.E.O. Scholar Award. Research reported in this article was supported by the National Institute of Mental Health of the National Institutes of Health (NIH) under Award Number T32MH019960. The content is solely the responsibility of the author and does not necessarily represent the official views of NIH. The author would like to thank the school staff, social workers, and students who made this project possible. Carolyn Widen, BSW, MSSW, assisted with data collection and review of preliminary analyses.