The moderating effects of coping and self-esteem on the relationship between defeat, entrapment and suicidality in a sample of prisoners at high risk of suicide

Abstract

Background

Research is sparse which examines pathways to suicide, and resilience to suicide, in people who are particularly vulnerable to suicide, for example, prison inmates. The purpose of this study was to examine the ways in which perceptions of self-esteem and coping ability interacted with defeat and entrapment to both amplify suicidal thoughts and feelings, and to act as a buffer against suicidal thoughts and feelings.

Methods

Participants were 65 male prisoners at high risk of suicide. A cross-sectional questionnaire design was used. Questionnaire measures of depression, defeat, entrapment, self-esteem, coping ability and suicidal probability were administered.

Results

For the hopelessness component of the suicide probability measure, high levels of coping ability together with low levels of defeat resulted in the lowest levels of suicidality indicative of a resilience factor. In contrast, low levels of coping skills together with high levels of entrapment were a high risk factor for this hopelessness component of suicide. This pattern of results pertained when controlling for depression levels.

Conclusions

This is the first study to examine interactions between defeat, entrapment and appraisals of self-esteem and coping ability. Therapeutic interventions would benefit from boosting perceptions and appraisals of coping ability, in particular, in people who are at high risk for suicide.

1. Introduction

Suicidal thoughts, behaviours and deaths are a serious health concern [6,14]. The prevalence of suicidal thoughts and behaviours increases dramatically in those with severe mental illnesses. For example, survey results reported that one half of individuals with symptoms of schizophrenia attempt suicide [33]; the risk for suicide deaths in those with bipolar disorder is between 20–30 times greater than that in the general population [38]; and the high prevalence of suicide in those with unipolar depression is well documented [21].

More mental illnesses are experienced by prisoners compared to the general population [15]. Furthermore, prisoners are often vulnerable to suicide in other ways, such as, being homeless, or in financial difficulties, both of which are associated with increased suicidality in general [20,40]. In addition, aspects of the prison environment may intensify suicidality. For example, being incarcerated can lead to fear, a lack of agency, a lack of trust [29], and social isolation [41]. Occupying a single cell has been associated with suicide deaths in prisoners [16]. Consequently, prisoners are at a very high risk of experiencing suicidal thoughts, plans, and behaviours. Indeed suicide deaths have been documented as being from three to eight times higher in prisoners compared to the general population [15,17,36]. Furthermore, this may be an underestimate of the problem of suicidality in prisoners because of limited suicide risk assessment procedures, high vigilance, and restrictions on what prisoners are allowed to bring into their cells [29]. Thus, suicidal ideation may be a more sensitive indicant of suicide risk in prison contexts. Office for National Statistics for England and Wales figures showed that lifetime suicidal ideation for male and female prisoners were 41.5% and 55.5% respectively, whereas lifetime suicide attempts were 23.5% and 40.5% respectively [31]. In a meta-analysis, the odds of death by suicide were found to be 15 times greater in prisoners reporting suicidal ideation [16].

There is increasing evidence that perceptions of being defeated and trapped are strongly linked to suicidality [47]. It is important to determine which psychological factors exacerbate defeat and entrapment. An area of research, which requires expansion, concerns perceptions of coping ability and self-esteem, and the effect of such perceptions on defeat and entrapment. This is important therapeutically because if low levels of self-esteem and coping heighten defeat and entrapment, then this indicates a primary therapeutic target. Conversely, if high levels of self-esteem and coping weaken defeat and entrapment, then this identifies a focus for building suicide resilience [23].

Self-esteem describes feelings of self-worth [34]. There is a robust literature documenting associations between low self-esteem and suicidality in adults [5,8,12,26,37,43,49]. For example, one study reported that in depressed inpatients, those expressing suicidal ideation with a history of suicide attempts had significantly lower self-esteem compared to inpatients who had no such thoughts or histories [37]. Two studies have examined the relationship between self-esteem and suicidality in prisoner populations. The first had a focus on male inmates in the UK which found that self-esteem was significantly lower in prisoners who had made near-lethal suicide attempts (n = 60), compared to those who had not made such attempts (n = 60) [41]. Another study conducted in China reported that depression, low self-esteem, and poor social support were the strongest predictors of suicidal ideation, in that order, but only in female prisoners (n = 201) [49]. Notably, the literature addressing self-esteem in relation to suicidality is sparse for prisoner populations at high risk of suicide. Furthermore, moderation/interaction effects between levels of self-esteem and other key psychological factors, such as, perceptions of defeat and entrapment, in the route to suicidality have received relatively scant attention in the general suicide literature [4,12,26,28,48].

The same issue occurs with coping ability. Coping skills refers to the capacity to offset problems and overcome, or tolerate, stressors [27]. As might be expected, perceptions of having poor coping skills have been linked with suicidal thoughts and behaviours. Work with medical inpatients pointed to an association between high levels of dysfunctional coping skills and high levels of suicidal ideation [30]. Similarly, avoidant coping, i.e., avoiding dealing with stressors, together with the use of substances to cope, was also linked to greater levels of suicidality in homeless young people [25]. Overall, research investigating the relationship between coping skills and suicidality are relatively few. In addition, as with the self-esteem literature, this relationship has been minimally investigated in people who are at high risk of suicide, such as prisoners, nor have moderation/interaction effects between coping ability and defeat and entrapment been sufficiently examined.

Hence, the current study had three aims with a broad focus of understanding suicidal thoughts and behaviours in male prisoners who were most vulnerable to suicide. The first aim was to investigate the impact of low levels of self-esteem and dysfunctional coping abilities on suicidal thoughts and behaviours. The second aim was to examine the extent to which low levels of self-esteem and dysfunctional coping abilities amplified suicidality because they worsen defeat and entrapment. Conversely, the third aim was to determine whether high levels of self-esteem and coping skills buffer perceptions of defeat and entrapment in the pathway to suicidality. Our measure of suicide provided an overall assessment of suicidal thoughts and feelings, but it also allowed an examination of four different components of suicidality. These components reflected suicidal ideation, hopelessness, negative evaluations and hostility. An exploratory goal was to investigate our three aims for each of these components of our suicide measure.

2. Methods

2.1. Design

A cross-sectional questionnaire design was used. The outcome variables were total and sub-scale scores on the Suicide Probability Scale [9]. Predictor variables were defeat and entrapment. Moderator variables were coping and self-esteem. Depression was a control variable.

2.2. Participants

Male volunteers were recruited from a high security prison in England as part of a pilot RCT assessing a cognitive therapy for suicide [39]. Potential participants were identified by prison staff. Participants received no remuneration or privileges for taking part. At the time of recruitment, all participants had to have been identified as being at risk of self-harm or suicide in the past month by prison or prison health care staff according to the Assessment, Care in Custody and Teamwork system of the UK prison service (ACCT). The ACCT system is specific to the prison service in England and Wales. It is the care planning and management system used to identify, monitor, and manage prisoners at risk of suicide or self-harm. No later than 24 hours after the ACCT has been initiated, as part of an ACCT assessment, the at-risk prisoner must be interviewed by a trained ACCT assessor [32].

There were four additional inclusion criteria, namely:

• being over 18 years of age;

• having sufficient English language ability to understand the participant information sheet and consent form;

• having capacity to consent to participation;

• and not considered by prison staff to pose a risk to the researcher.

2.3. Measures

The measures were part of a larger battery of tests conducted as part of a pilot randomised control trial assessing cognitive behavioural therapy for suicidality in prisoners [39]. Cronbach’s alpha reliability scores for each measure are presented in Table 1.

Table 1.

Mean scores, standard deviations, sample size (N), Cronbach’s alpha reliability scores, and correlations between key variables.

	M	SD	N	Alpha	2	3	4	5	6	7	8	9	10
1. SPS Total	86.94	20.47	63	.87	.86**	.86**	.70**	.61**	.72**	.61**	.67**	−.32*	−.70**
2. SPS ideation	24.48	7.89	63	.81	-	.64**	.51**	.39**	.68**	.59**	.61**	−.33**	−.58**
3. SPS hopelessness	28.10	8.43	63	.76		-	.54**	.36**	.68**	.64**	.70**	−.16	−.71**
4. SPS negative evaluation	18.43	4.43	63	.65			-	.22	.51**	.40**	.40**	−.41**	−.49**
5. SPS hostility	15.74	5.47	61	.63				-	.22	.18	.20	−.16	−.30*
6. Depression	35.11	12.33	64	.89					-	.58**	.70**	−.22	−.67**
7. Defeat	54.89	14.41	63	.93						-	.68**	−.01	−.73**
8. Entrapment	56.22	15.23	64	.92							-	−.09	−.69**
9. Coping skills	141.41	31.49	64	.91								-	.25
10. Self-esteem	101.52	25.71	61	.80									-

^*P ≤ .05;

^**P ≤ .01. Only the defeat scores were not normally distributed meaning that correlations with this variable were analysed with Spearman’s correlation coefficient. All other correlations used Pearson’s Product Moment correlation coefficient. SPS: Suicide Probability Scale.

2.3.1. Suicidal probability

The Suicide Probability Scale (SPS) [9] is a 36-item questionnaire, which has four sub-scales of hopelessness, suicidal ideation, hostility, and negative self-evaluation. Examples include “When I get mad I throw things”, “I feel so lonely I cannot stand it”. Participants respond to each item using four response-options. This scale provides a total score, and a three-category probability of suicide indicant, based on the profile of sub-scale scores and contextual factors. The total score was used in the current study with a possible range of 30–146.

2.3.2. Depression

The Beck Depression Inventory-second edition (BDI-II) [2] is a 21-item multiple-choice inventory assessing the severity of a range of depressive symptoms (e.g., sadness, pessimism). Items are rated for their severity over the past two weeks using a 4-point scale with a potential range of 0–63.

2.3.3. Self-esteem

The Robson Self Concept Questionnaire (RSCQ) [42] is a 30-item self-report measure, e.g., “I never feel down in the dumps for very long”. A 7-point scale is used giving a possible range of scores from 30–210.

2.3.4. Coping ability

The Coping Inventory for Stressful situations (CISS) [13] is a 48-item measure, e.g., “I can adjust my priorities.” Items are responded to using a 5-point scale ranging from “not at all” to “very much” with a range of 48–240.

2.3.5. Defeat

The defeat scale [19] is a 16-item questionnaire assessing appraisals of defeat including a sense of failed struggle and low social rank (e.g., “I feel that I have not made it in life”). Items are rated on a 5-point scale from ‘Never’ to ‘Always’ for their prevalence during the past week. Scores can range from 0–64.

2.3.6. Entrapment

The entrapment scale [19] is also a 16-item scale assessing feelings of being trapped by internal and external events as well as feelings of a desire to escape (e.g., “I want to get away from myself”). Items are rated on a 5-point scale from ‘Not at all like me’ to ‘Extremely like me’. Scores can range from 0–64.

2.4. Procedure

Participants were asked to read the participant information sheet at least 24 hours prior to signing the consent form. Questionnaires were completed in the order stated above in a private room in the prison, in testing sessions lasting around two hours. This study was approved by a UK NHS ethics committee.

2.5. Analysis strategy

Pearson’s Product Moment or Spearman’s rho correlation coefficients were calculated as appropriate for the data distribution.

Hierarchical multiple regression analysis was used to determine which variables predicted outcome measures as assessed by the suicide probability questionnaire, and its component sub-scales. In the first analysis, the control variable of depressed mood scores was entered into the first step of the regression model, predictor variables of scores on defeat, entrapment, and self-esteem measures were entered into the second step, and moderation effects, which were reflected by interaction/multiplication terms between i defeat and self-esteem, and ii entrapment and self-esteem, were entered into the third step. The second hierarchical regression analysis had the same structure as the first, but coping scores replaced self-esteem. The enter method was used which enters all variables in a step simultaneously [18]. All control and predictor variables were standardised to z scores prior to entry and prior to the calculation of the interaction terms [18]. Outcome variables were checked for distribution normality (skewedness divided by the standard error) [18]. Multi-colinearity was considered problematic if correlations between predictor variables exceeded .8 [18]. The alpha level was set at .05, two-tailed. Data were analysed with IBM SPSS statistics, version 20. Algorithms provided by Dawson et al. were used to plot interaction effects [1,10,11].

3. Results

3.1. Participant characteristics

Participants were 65 male prisoners, aged 22–60 (M = 35.8, SD = 10.95). Most were white (n = 58). Christianity was the most frequent religion (n = 36). The mean number of times that participants had been imprisoned was 3.48 (SD = 6.7). The mean age at which the sample had first been imprisoned was 25.55 (SD = 11.68). The most frequent index offences were sexual crimes (n = 24) and acquisitive offences (n = 22). Psychiatric diagnosis data was drawn from clinical records provided by one consultant psychiatrist at the host prison. Eight participants had no psychiatric diagnoses; 22 had depression; 10 had comorbid depression and anxiety, three were on the schizophrenia disorder spectrum, one had a personality disorder, and the remainder had numerous comorbid symptoms involving severe mental illnesses (e.g., schizophrenia and depression).

3.2. Descriptive data and correlations

The mean scores, standard deviations and correlation coefficients are shown for all key variables in Table 1.

3.3. Regression analyses

The regression analyses tested the predictions that perceptions of low self-esteem and poor coping would be positively associated with suicidality. It was further predicted that low levels of these variables would amplify relationships between defeat and entrapment, and suicidality, and that high levels of these variables would buffer relationships between defeat and entrapment, and suicidality. All outcome variables were normally distributed.

3.4. Self-esteem, defeat, entrapment, and suicidality

The first set of analyses examined whether self-esteem moderated the relationships between defeat and suicidality, and entrapment and suicidality, measured by the suicide probability scale, and the four sub-scales of this measure (Table 2).

Table 2.

Results of the regression models examining predictive effects of self-esteem, defeat, and entrapment on suicide probability and the components of the suicide probability scale for steps which were significant. Coefficients are given only for steps of each model, which were significant.

Outcome variable	Step	Predictor variables	β	SE β	B	CI (95%)	t
SPS - total	1 F(1, 59) = 64.26, P < .0001, R2 = .52	Depression	15.11	1.86	.72	11.34–18.89	8.02, P < .0001
	2 F(3, 56) = 5.07, Δ R2 = .10	Depression	7.79	2.58	.37	2.63–12.95	3.02, P < .01
		Self-esteem	−5.81	2.74	−.28	−11.30 to −0.33	−2.12, P < .05
		Defeat	2.07	2.87	.10	−3.63–7.81	.72 ns
		Entrapment	2.96	2.78	0.14	−2.60–8.53	1.07, ns
SPS hopelessness	1 F(1, 59) = 57.49, P < .0001, R2 = .49	Depression	6.05	.80	.70	4.45–7.64	7.58, P < .0001
	2 F(3, 56) = 7.05, P < .0001, Δ R2 = .14	Depression	2.56	1.05	.30	.46–4.66	2.45, P < .05
		Self-esteem	−2.66	1.11	−.31	−4.89 to −0.43	−2.39, P < .05
		Defeat	.09	1.16	.01	−2.24–2.42	.08 ns
		Entrapment	2.36	1.13	0.28	0.1–4.62	2.09, P < .05
SPS ideation	1 F(1, 59) = 54.34, P < .0001, R2 = .48	Depression	5.59	.76	.69	4.07–7.10	7.37, P < .0001
SPS negative evaluation	1 F(1, 59) = 20.34, P < .0001, R2 = .26	Depression	2.29	.51	.51	1.27–3.31	4.51, P < .0001

SPS: Suicide Probability Scale.

When the suicide probability scale total scores were used as the outcome variable, the first two steps of the model were significant but not the third step containing the interaction/moderator effects (F(1, 54 = 0.34)). Depression and self-esteem were significant predictors. This pattern in the data was repeated to some extent for the hopelessness sub-scale scores where depression, entrapment, and self-esteem were significant predictors, but there were no significant moderation effects at step 3 (F(1, 54 = 0.30)). The suicidal ideation sub-scale outcome resulted in only the first step of the model, containing the depression control variable, being significant (F(3,56 = 2.61) for step 2; F(2,54 = .62) for step 3). Similarly, for the negative self-evaluation sub-scale outcome, only the first step of the model was significant (F(3,56 = 1.09) for step 2). No steps in the regression model were significant for the hostility sub-scale scores (F(1,59 = 2.90) for step 1; F(3,56 = .87) for step 2; F(2,54) = .86 for step 3).

3.5. Coping skills, defeat, entrapment, and suicidality

The second set of analyses followed the first but replaced self-esteem scores with perceived coping skills scores (Table 3).

Table 3.

Results of the regression models examining predictive effects of coping, defeat, and entrapment on suicide probability and the components of the suicide probability scale for steps which were significant. Coefficients are given only for steps of each model, which were significant.

Outcome variable	Step	Predictor variables	β	SE β	B	CI (95%)	t
SPS - total	1 F(1, 61) = 64.57, P < .0001, R2 = .51	Depression	14.79	1.84	.72	11.11–18.48	8.04, P < .0001
	2 F(3, 58) = 6.33, Δ R2 = .12	Depression	7.54	2.41	.37	2.72–12.36	3.13, P < .01
		Coping	−4.50	1.68	−.22	−7.86 to −1.15	−2.69, P < .01
		Defeat	4.84	2.38	.24	.08–9.59	2.03, P < .05
		Entrapment	4.76	2.57	.23	−.39–9.91	1.85, ns
SPS hopelessness	1 F(1, 61) = 52.72, P < .0001, R2 = .46	Depression	5.79	.80	.68	4.19–7.38	7.26, P < .0001
	2 F(3, 58) = 5.69, P < .01, Δ R2 = .12	Depression	2.64	1.06	.31	.53–4.76	2.51, P < .05
		Coping	−.53	.74	−.06	−2.00 to −0.95	−.72, ns
		Defeat	1.77	1.04	.21	−.31–3.86	1.70, ns
		Entrapment	2.81	1.13	.33	0.55–5.07	2.49, P < .05
	3 F(2, 56) = 3.16, P = .05, Δ R2 = .04	Depression	1.92	1.06	.23	−.21–4.05	1.81, ns
		Coping	−.98	.80	−.12	−2.59–.63	−1.22, ns
		Defeat	2.82	1.14	.34	.53–5.12	2.47, P < .05
		Entrapment	2.74	1.11	.32	.51–4.98	2.46, P < .05
		Defeat × coping	2.62	1.19	.27	.24–4.99	2.21, P < .05
		Entrapment × coping	−2.05	.90	−.24	−3.86 to −.25	2.21, P < .05
SPS ideation	1 F(1, 61) = 52.55, P < .0001, R2 = .46	Depression	5.41	.75	.68	3.92–6.90	7.25, P < .0001
	2 F(3, 58) = 5.51, P < .01, ΔR2 = .12	Depression	2.79	.99	.35	.81–4.78	2.82, P < .01
		Coping	−1.86	.69	−.24	−3.24 to −.47	−2.69, P < .01
		Defeat	2.34	.98	.30	.38–4.30	2.39, P < .05
		Entrapment	1.11	1.06	.14	−1.02–3.23	1.04, ns
SPS negative evaluation	1 F(1, 61) = 21.47, P < .0001, R2 = .26	Depression	2.28	.49	.51	1.3–3.26	4.63, P < .0001
	2 F(3, 58) = 3.50, P < .05, ΔR2 = .11	Depression	1.37	.68	.31	0–2.73	2.00, P = .05
		Coping	−1.49	.48	−.34	−2.44 to −.54	−3.14, P < .05
		Defeat	.55	.67	.12	−.8–1.9	.82, ns
		Entrapment	.36	.73	.08	−1.10–1.82	.49, ns

SPS: Suicide Probability Scale.

For the total suicide probability scores, the first and second steps of the model were significant, with depression, defeat and poor coping skills being significant predictors. The third step containing the moderation/interaction effects was not significant (F(2,56) = 2.06).

As with the pattern of results for self-esteem, none of the steps in the regression model were significant when the hostility sub-scale scores were entered as the outcome (F(1,59) = 2.90 for step 1; F(3,56) = .63 for step 2; F(2,54) = 51 for step 3). When suicidal ideation sub-scale scores were the outcome, the first two steps of the model were significant. Depression, defeat and coping skills were significant predictors. Step 3 was not significant (F(2,56 = 1.06). For the negative self-evaluation scores, the first two steps of the model were significant. Depression and coping skills, but not defeat or entrapment, were significant predictors. Again, the third step was not significant (F(2,56) = 0.60).

However, for the hopelessness sub-scale scores all three steps of the regression model were significant. There were significant effects for defeat, entrapment, and the interaction/moderation effects between defeat and coping skills, and between entrapment and coping skills. The interaction/moderation effects are depicted in Fig. 1a (defeat) and b (entrapment). Low levels of defeat together with high levels of coping skills appeared to be protective of hopelessness related to suicide probability. Considering perceptions of entrapment, high levels of entrapment coupled with low levels of coping skills seemed to be a risk factor for this hopelessness element of suicide probability.

Fig. 1.

a and b: the moderation effect between defeat and coping skills and the hopelessness sub-scale scores of the suicide probability scale (SPS) is shown in the first figure, a. The moderation effect between entrapment and coping skills and the hopelessness sub-scale scores of the suicide probability scale (SPS) is shown in the second figure, b.

4. Discussion

This is the first study to examine, in a group of individuals who were highly vulnerable to suicidal thoughts and behaviours:

• the relationship between self-esteem, coping, and suicidality;

• and the extent to which self-esteem and coping moderated defeat and entrapment to both amplify and ameliorate suicidality.

There were four key findings.

First, as expected, low levels of self-esteem were associated with high scores on the suicide probability scale even when controlling for depression. This finding is in accord with previous literature sampling populations across countries, communities, and psychiatric settings [3,8,23,26,37]. Two further studies, found that low levels of self-esteem were associated with suicidality in male [41], and female prisoners [50]. In the current study, the relationship between low self-esteem and suicidality was driven by the hopelessness component of the suicidality measure, but not the hostility, suicidal ideation, or negative evaluation components. No published studies to date have attempted to further examine the components of the suicide probability scale that are affected by low self-esteem in this manner.

Second, contrary to predictions, self-esteem did not moderate defeat and entrapment to either amplify or buffer the effect that these perceptions had on suicidal thoughts and feelings. A review found that results were equivocal concerning the role of high levels of self-esteem as a protective factor, which buffers suicidality [23]. Some studies reporting positive moderation effects have sampled college students [12,26]. It can be questioned whether results from college students can be generalised beyond that population, especially to those who are highly vulnerable to suicidality.

The third finding was that low levels of perceived coping skills predicted the probability of suicidality, again, whilst controlling for depression. Other work has also found relationships between dysfunctional coping skills and suicidality in female prisoners [7,22], psychiatric inpatients [24], patients with physical illnesses [30], and homeless young people [25].

The fourth, and most impactful, finding of the current study concerned interaction/moderation effects of coping skills on the relationship between defeat and entrapment and suicide probability scores. These moderation effects were evident only when the hopelessness component of the suicide probability measure was examined. Perceptions of coping skills affected defeat and entrapment in different ways. High levels of coping ability coupled with low levels of defeat resulted in the lowest levels of suicidality related to hopelessness, which is indicative of a resilience factor [23]. In contrast, high levels of entrapment coupled with low levels of coping skills were a high risk factor for the hopelessness element of the suicide probability scale. An additional important point was that perceived coping skills did not correlate highly with the hopelessness element of the suicidality measure, but it did act as a moderator. This illustrates that it is essential that suicide research expands beyond the examination of main effects to investigate more complex interactive relationships [45].

It is important that the coupling between entrapment and coping was particularly harmful with respect to suicidality, given that the participants were prisoners. It is also important that these effects were observed with the hopelessness component of the suicidality measure because hopelessness, above and beyond depression, has been found to be a high risk factor for suicidal thoughts, behaviours, and death [35,47].

These findings have three clinical implications. First, in terms of assessing suicidality, it is prudent to examine suicidal amplifiers, specifically, a combination of perceptions of entrapment and low coping skills. Second, psychological interventions would benefit from boosting both negative perceptions [45,46] of self-esteem and coping skills, but more importantly the latter, given that negative perceptions of coping skills may amplify entrapment, and positive perceptions of coping skills may counter perceptions of defeat. This last point serves to highlight the clinical importance of considering factors, which confer resilience to suicidality [23]. Third, these patterns of findings pertained even when controlling for depression. This underscores the value of focusing on suicidality in any treatment plan as opposed to focusing on depression alone [44,45].

Four limitations of this study warrant discussion. First, it must be remembered that these findings are based on a sample of prison inmates. Hence, the differential effects of coping skills on defeat and entrapment may have been a function of the prison context. No previous work has examined defeat, entrapment, self-esteem, and coping skills in this interactive manner in a prison context. Consequently, there is no benchmark against which to compare the current findings. However, it should be highlighted that prison contexts overlap with some in-patient psychiatric settings and with those who have reduced mobility. The second limitation is that although we examined sub-components of suicidality our measures did not permit us to investigate different factors contributing to self-esteem. Third, this study was cross-sectional. Thus, causality cannot be inferred. Finally, the sample size was small, meaning that non-significant results may have become significant with a larger sample. However, small samples are not unusual when conducting research with people who are highly vulnerable to suicide and who are difficult to recruit, and for this reason underrepresented in the research literature. Indeed, this kind of data from prison samples is rare.

In conclusion, this study presents a novel contribution to the field of suicide research in four ways. First, the population sample was at high risk of suicidal thoughts and behaviours. Second, the findings indicated that low self-esteem and poor coping ability were associated with suicidality in individuals at high risk of suicide whilst controlling for depression levels. Third, dysfunctional coping skills worsened perceptions of entrapment and, thus, heightened the hopelessness component of the suicide probability measure, but functional coping skills ameliorated perceptions of defeat to lower the hopelessness component of the suicide probability scale. Furthermore, these differential findings were not due to depression. These results are exciting because they illustrate how nurturing positive perceptions of coping abilities have the potential to counter suicidality and improve resilience to suicidality. Moreover, many health professionals are well situated to deliver interventions, which improve coping abilities.