Emotional Intelligence is a Protective Factor for Suicidal Behavior

Objective: Little is known about what factors protect against the occurrence of suicide ideation and attempts. We tested whether emotional intelligence (EI)—the ability to perceive, integrate, understand, and manage one’s emotions—decreases the likelihood of suicide ideation and attempts among those at risk. Method: Adolescents ( N =54) aged 12-19 were recruited from local psychiatric clinics and the community to participate in this cross-sectional laboratory-based study. Analyses examined whether the relations between childhood sexual abuse (CSA) and suicide ideation and attempts were moderated by adolescents’ EI. These constructs were assessed using self-report, structured interviews, and performance-based tests, respectively. Results: Analyses revealed that EI is a protective factor for both suicide ideation and attempts. Specifically, CSA was strongly predictive of these outcomes among those with low EI, weakly predictive among those with medium EI, and completely unrelated among those with high EI. Follow-up analyses revealed that the protective effect of EI was driven primarily by differences in Strategic EI (i.e., ability to understand and manage emotions) but not Experiential EI (i.e., ability to perceive emotions and integrate emotions into thoughts). Conclusions: This study provides preliminary evidence that EI is a protective factor for suicide ideation and attempts. Important next steps include testing the moderating influence of EI on a wider range of stressful life events and self-injurious behaviors, as well as conducting experimental studies to determine whether enhancing EI decreases the subsequent occurrence of these behavior problems.

engage in suicidal behaviors. This raises the important question of what factors might protect against the occurrence of these outcomes. Protective factors are those that decrease the probability of a negative outcome among those at risk. 8 It is important to note that a protective factor is not merely defined by the absence of a risk factor. Rather, it is a third variable that modifies the strength or direction of the relation between a risk factor and outcome. For instance, several recent studies suggest that genetic and neurobiological factors can buffer the influence of stressful life events on the likelihood of suicide attempts. [9][10][11] Other research suggests that environmental factors such as reduced accessibility to firearms, 12, 13 religious affiliation, 14,15 and social support 16-18 may moderate the influence of stressful life events on suicide risk. The study of psychological protective factors for suicidal behaviors represents an important research direction given that such factors are likely to be more easily modified than other moderators identified by prior work (e.g., genetic polymorphisms, presence of social support).

Emotional Intelligence & Suicidal Behavior 4
Psychological theories of suicide suggest that people engage in suicidal behaviors due to an inability to tolerate or modulate the experience of negative affect. 19,20 It is therefore possible that those who are especially adept at perceiving, integrating, understanding, and managing their emotions would be at reduced risk for suicidal behaviors in response to stressful life events.
Accordingly, the current study was designed to test whether stronger abilities in each of these domains are protective against suicidal behaviors in the presence of significant stressful life events.
In the current study, stressful life events were operationalized as the experience of childhood sexual abuse (CSA). This decision was based on several factors. First, prior studies testing potential moderators of the relation between life stressors and suicide attempts and related forms of psychopathology (e.g., depression) have used childhood maltreatment 21 and CSA in particular 11 as a measure of stressful life events. The assessment of CSA in the current study therefore facilitates comparisons across studies. Second, the association between CSA and suicidal behaviors has been replicated across many studies. 22, 23 CSA is a stronger risk factor for suicidal behaviors than other forms of childhood maltreatment 24 or childhood adversities (e.g., loss of caregiver, family violence, parental psychopathology). 25, 26 Third, past work has reported that history of CSA affects the transmission of suicide risk from parents to offspring, 27, 28 further highlighting need to better understand this risk factor.
The way that emotions are experienced and their influence on thoughts and behaviors has become the focus of increasing interest and study in psychology and psychiatry. 29, 30 Most research in this area has focused on the construct of emotion regulation; although, this construct has been defined and measured in very different ways across studies. 31 In the current study, we sought to examine the moderating effect of a broad range of emotional processes or abilities, Emotional Intelligence & Suicidal Behavior 5 including (but not limited to) the way in which people manage their emotional experiences. To do so, we examined the moderating effect of Emotional Intelligence (EI)-a multi-dimensional construct that refers to the ability to reason about emotions, as well as the ability to use emotions in guiding one's reasoning and behavior. 32-35 EI, which has been the focus of a substantial body of research over the past two decades, can be decomposed into two main abilities (each composed of two sub-branches). Strategic EI is defined as the ability to understand and manage emotions (e.g., understanding what feelings are expected to emerge in response to different events, and knowing how to regulate emotions to help attain one's goals). In contrast, Experiential EI is defined as the ability to perceive emotions and use them to facilitate thought (e.g., accurately reading facial expressions and integrating emotions into thoughts). Strategic EI is considered to be a more developmentally complex ability that is involved in managing one's emotions in order to facilitate problem-solving and decision-making, whereas Experiential EI is considered to represent lower level abilities. [32][33][34][35] Some prior research has linked problems with EI and the experience of clinical behavior problems. For instance, several aspects of self-reported EI similar to Experiential (e.g., perceiving emotions) and Strategic (e.g., managing others' emotions) have been reported to moderate the relation between stress and the outcomes of hopelessness, depression, and suicidal ideation. 36 The current study extends this earlier work by using a more comprehensive and rigorous measure of EI and by examining the relation between EI and suicidal behaviors. The goal of the current study was to test the primary hypothesis that EI moderates the relation between CSA and the experience of suicidal behaviors such that higher EI will be associated with lower levels of suicide ideation and suicide attempts. Given that prior studies suggest that suicidal behavior may result from problems regulating one's negative emotions, 19, 20 a secondary Emotional Intelligence & Suicidal Behavior 6 hypothesis was that Strategic EI would be a particularly important moderator of the relation between CSA and suicidal behavior.

Method Participants
Fifty-four (46 female) adolescents (age in years: M=17.30, SD=1.92, range 12-19), 31 with a recent (i.e., past year) history of suicide ideation (57.4%) and/or attempts (14.8%) and 23 with no recent history, participated in the current study. Suicidal and non-suicidal groups were matched on age, sex, and race/ethnicity to rule out the possibility that any observed effects are due to these factors. They were recruited using study advertisements placed in local psychiatric clinics, newspapers, community bulletin boards, and the internet. We focused on adolescence because of the increased risk of self-injurious behavior during this developmental period. 2,3 These 54 participants were drawn from a larger sample of 94 adolescents who participated in a laboratory-based study of nonsuicidal self-injury. 37-39 This study is reported separately due to the novel hypothesis and constructs (i.e., EI, suicide ideation and attempts) that were the focus of current analyses, and because the EI measure was added to the study after it had begun and so was only administered to these 54 adolescents. Adolescents in the current sample did not differ from the 40 excluded on any of the study variables measured. The current sample size provides adequate statistical power to detect large effects (power=.95) and some medium effects (power=.62) using two-tailed tests with alpha set at .05. Ethnicity of the participants was selfidentified as European American (79.6%), Hispanic (5.6%), Asian American (3.7%), African American (1.9%), and Other (9.2%).

Measures
Childhood Trauma Questionnaire [CTQ]. The CTQ 40 is a 28-item (< 5 min) self-report measure of childhood and adolescent maltreatment, including: physical abuse, sexual abuse, Emotional Intelligence & Suicidal Behavior 7 emotional abuse, physical neglect, and emotional neglect. For the purposes of this paper, we focus on CSA alone-calculated by adding up responses to its five corresponding items. Items are rated on a five-point scale (0 = never true, 5 = very often true). The CTQ has demonstrated high internal consistency, convergent and discriminant validity, as well as good sensitivity and specificity for all forms of maltreatment among a clinical sample of adolescents. 40 The CSA subscale demonstrated excellent internal consistency (α = .89) in the current sample. Overall, 20.4% of the current sample endorsed at least one item from the CSA subscale. Self-Injurious Thoughts and Behaviors Interview [SITBI]. The SITBI 45 is a structured interview (3-15 minutes) that assesses presence, frequency, and other characteristics of a broad range of self-injurious thoughts and behaviors, including suicide ideation ("Did you ever have thoughts of killing yourself?") and attempt ("Did you ever make an actual attempt to kill yourself?"). In the current study, we focused on items measuring the frequency of suicide ideation and attempts in the past year. The SITBI has demonstrated reported strong inter-rater reliability, test-retest reliability, and concurrent validity. 45 The SITBI was administered by trained master's-and doctoral-level researchers, as well as by closely supervised bachelor's-level research assistants. At the conclusion of each interview, each participant underwent a risk assessment and received a referral for clinical services when indicated.

Procedures
Participants who responded to the study advertisements were invited to the laboratory and provided with a complete description of the study. Written informed consent was obtained, with parental consent obtained for participants <18 years. Assessments were completed during a Emotional Intelligence & Suicidal Behavior 9 baseline laboratory assessment, for which participants received $100. Because the MSCEIT:YV-R was incorporated into the larger study shortly after it had begun, MSCEIT:YV-R data were obtained from the first 11 participants (paid an extra $10) by sending them this measure via postal mail with a return-postage envelope. MSCEIT:YV-R scores did not significantly differ between these 11 participants and the 43 who completed this measure in the laboratory. All procedures were approved by the university's institutional review board.

Data Analysis
Before conducting analyses, suicide ideation and attempt variables were adjusted to reduce the influence of outliers. Following the recommendations of Tabachnick and Fidell, 46 we re-assigned outliers less extreme values (i.e., within 2 standard deviations of the mean) that retained their relative standing in the distribution (i.e., still the highest scores on each variable) and variables that were not normally distributed were transformed to more closely approximate normality. 46 Next, we tested the magnitude of the relations between CSA, EI, and suicide ideation and attempts using correlations. We then tested whether the relations between CSA and suicidal behaviors were moderated by EI following the recommendations for testing moderation and conducting post-hoc probing. 47, 48 Specifically, the CSA and EI variables were centered prior to computing the interaction variable. The moderation models were then examined through hierarchical linear regression analyses, entering the centered CSA and EI variables in the first step and the interaction variable in the second step. Significant interactions were plotted and probed through simple slope analyses 47, 48 using conditional variables computed based on high (+1SD), medium, and low (-1SD) moderator (i.e., EI) values.

Relations between EI, CSA, and Suicidal Behaviors
Descriptive statistics and correlations between EI, CSA, and suicidal behaviors are presented in Table 1. As expected, CSA was significantly correlated with both suicide ideation and attempts. EI was not significantly associated with CSA or with suicidal behaviors, with the correlations among these constructs representing mostly small effects. This means that EI is not merely a correlate of, or risk factor for, CSA and suicidal behaviors. Although the absence of significant associations between EI and these other variables was not hypothesized, this pattern of findings strengthens the potential argument for moderation. 49

Is EI a Protective Factor for Suicidal Behaviors?
Consistent with our primary hypothesis, overall EI significantly moderated the relation between CSA and both suicide ideation and attempts ( Table 2). As presented in Figure 1a, CSA was strongly associated with suicide ideation for participants with low EI, whereas this relation was weaker for those with medium scores on EI, and completely absent for those with high EI.
The same pattern was revealed for suicide attempts.
In order to better understand which aspects of EI were protective for suicidal behavior, we separately tested the moderating effects of Strategic and Experiential EI. In support of our secondary hypothesis, analyses revealed that Strategic EI significantly moderated the relation between CSA and both suicide ideation and attempts ( Table 2). As presented in Figure 1b, CSA was strongly associated with suicide ideation for participants with low Strategic EI, whereas this relation was weaker for those with medium scores on Strategic EI, and completely absent for those with high Strategic EI. The same pattern was again revealed for suicide attempts.

Emotional Intelligence & Suicidal Behavior 11
Experiential EI did not significantly moderate the relations between CSA and suicide ideation or attempts (Table 2 and Figure 1c).

Discussion
The goal of this study was to determine whether EI is a protective factor for suicidal behaviors among adolescents who have experienced significant life stressors. Consistent with our hypotheses, results revealed that EI moderates the relation between retrospectively reported CSA and past year suicide ideation and attempt, such that there is a strong relation between CSA and these suicidal behaviors among adolescents with low EI; a weaker relation among those with medium EI; and no significant relation among those with high EI. Further analyses revealed that Strategic EI drove the overall protective effect, and Experiential EI did not emerge as a significant protective factor for either suicide ideation or attempts. Several aspects of these findings warrant additional comment.
The specificity of the observed protective effect of Strategic EI (but not Experiential EI) for suicidal behaviors is an important strength of this study. The fact that effects were replicated across both suicide ideation and attempts increases confidence in the reliability of these findings.
These results are consistent with those from randomized controlled trials of cognitive therapy and dialectical behavior therapy (both of which include components aimed at enhanced emotion understanding and management) for the treatment of self-injurious behaviors, which cause decreases in suicide attempts. 50, 51 Taken together, these findings point toward the importance of abilities for understanding and managing emotions in the treatment and prevention of suicidal behaviors. Although studies on the treatment of suicide have not typically made explicit mention of EI, the constructs targeted in these treatments (e.g., emotion regulation skills, problem-solving skills) are similar to those studied by EI researchers. On balance, it is notable that we did not Emotional Intelligence & Suicidal Behavior 12 find a statistically significant association between Strategic EI-whose managing emotions branch has been equated to emotion regulation 52 -and either CSA or suicidal behaviors, despite prior findings indicating that emotion regulation is correlated with these variables. 19,20 Two aspects of this finding are notable. First, we did observe consistent small-to-medium associations among these variables (rs = -.15 to -.19) and so lack of statistical significance is partly the result of our lack of power for detecting such effects. Second, differences in associations with emotion regulation-related constructs may be due in large part to variations in the way such constructs are defined and measured. In the current study, we used the construct of Strategic EI and measurement strategy provided by the MSCEIT, which assesses knowledge about effective emotion regulation strategies, but not their actual use. Prior studies have found that although self-injurers and non-injurers are equally capable of proposing effective solutions to potential interpersonal problems, self-injurers are significantly worse at the selection and performance of solutions. 38 It is important that future work in this area attend carefully to differences in the terms, definitions, and measurement strategies used.
Nevertheless, some research suggests that EI is indeed a malleable target of change. One example of this has been demonstrated by Eack and colleagues, 53 reporting that Cognitive Enhancement Therapy 54 improves Strategic EI among patients with schizophrenia. Despite these findings, the majority of therapeutic approaches for suicide attempters emphasize improving Experiential EI (e.g., clarification of emotions experienced). 55 Studies are sorely needed to test the potential benefit of targeting Strategic EI in intervention and prevention programs. 35,56 Identifying EI as a protective factor represents a unique and important contribution to research on suicidal behaviors. The identification of risk factors has allowed research, treatment, and prevention efforts to isolate at-risk (e.g., sexually abused) populations; little has been done There are several limitations to this study. First, our sample was relatively small and so the results are based on a limited sample of episodes of suicide ideation, attempts and CSA. Our sample also consisted mostly of European American adolescent girls who were willing to participate in a research study, which may limit the generalizability of our findings. Moreover, Emotional Intelligence & Suicidal Behavior 14 because participants were recruited for a larger study of nonsuicidal self-injury, many of those with suicide ideation and all of those with suicide attempts in the current sample also had a history of nonsuicidal self-injury. These issues may limit the generalizability of the findings and it will be important to replicate these findings in a larger and more diverse sample. Second, the current study did not control for other psychological factors (e.g., personality traits, IQ) that might contribute to the effect of EI. Controlling for such factors may have strengthened our findings; however, it is worth noting that personality and IQ have been shown to relate specifically to self-reported EI and not with performance-based EI, 63 and so it is unlikely that these factors would have changed the observed effects. Third, the reports of childhood CSA are based on retrospective self-report. Recent frequency of suicide ideation and attempts was also retrospectively reported. As a result, it is possible that participants inaccurately recalled or reported CSA or suicidal behaviors. Prior work suggests that although there is error in long-term retrospective reporting of past events, risk estimates for suicide attempts (e.g., odds ratios) based on self-report data compared to official records are fairly similar. 64 Moreover, such risk estimates for suicide attempts based on retrospective recall have been found to be fairly stable over time. 65 Nevertheless, future studies should measure CSA and suicidal behaviors prospectively and using methods that do not rely exclusively on self-report. 37 Given these limitations, the results of this study should be considered preliminary evidence for the protective effect of EI that enhance our understanding of suicidal behaviors but also raise key issues to address in future research. First, it is important to replicate these findings among a larger, more diverse sample. Doing so would support the generalizability of these findings. Second, studies are needed to test prospectively whether EI protects against the effects of co-occurring stressors (e.g., poor-functioning family). 66 In a related vein, future research Emotional Intelligence & Suicidal Behavior 15 should assess whether EI buffers the effect of any mechanisms of CSA traumatization 67, 68 or any subsequent risk factors for suicide (e.g., psychopathology). 26, 69 Testing the degree of generalizability to other stressors and clinical risk factors is a necessary step to clarifying the nature of relations identified here. Third, it would be fruitful to test how EI works with other potential protective factors (e.g., social support) to reduce suicide risk. Studies also are needed to test the relations between EI and potential genetic 9,11 and neurobiological predispositions to suicidal behaviors. 10 The identification of additional protective factors and the subsequent development of more sophisticated theoretical and empirical models represent much needed directions in the effort to prevent these prevalent and harmful behavior problems.