Introduction

Suicide is a salient problem with similar rates around the world. For instance, the suicide prevalence rate for young people is 10.2 per 100,000 individuals in the US, 11.5 in Japan, and 8.7 in Singapore [1]. The suicide process is a progression starting from suicidal ideation, plans about taking one’s life, through suicide acts with increasing lethality and suicidal intent [2]. It is crucial to examine and measure risk factors for suicidal ideation.

A parsimonious theory of the extenuating contexts underlying suicidal ideation is the interpersonal-psychological theory of suicidal behavior [3], which posits that two key interpersonal factors are perceived burdensomeness (PB) and thwarted belongingness (TB). PB refers to the belief that one is a burden to significant others, and that they, therefore, would be better off if one were dead. TB refers to the belief that there is lack of care or concern from others, as well as lack of meaningful connections with others. When these two needs are unmet, there is psychological maladjustment, such as passive suicidal ideation [3]; and when there is hopelessness too, hopelessness predicts active suicidal ideation [4]. Other factors associated with suicidal ideation are depression, lack of meaning in life, lack of social support, and deficits in problem-solving appraisal [5–8].

The Interpersonal Needs Questionnaire (INQ) was created to measure the constructs of PB and TB [9], but only a few studies have employed it in Asian populations and, to the extent of our knowledge, no study has explored its psychometric properties in Asia. Zhang et al. [8] administrated an 18-item version of the INQ in China, and reported high internal consistency (Cronbach’s alpha = .87). In the US, Wong et al. [10] administrated the INQ-18 to a sample of Asian-American students. They reported internal consistencies of .74 for PB and .88 for TB, and found that PB was significantly and positively associated with suicidal ideation. The INQ-18 has also shown good internal consistency (.86 for PB and .84 for TB) in more ethnically heterogeneous populations in the US, but the two-factor structure of the INQ-18 could not be validated [11]. Zaroff et al. [12] employed the INQ-12 in Macau and found adequate internal consistency for PB (Cronbach’s alpha = .75) but low internal consistency for TB (Cronbach’s alpha = .43). In contrast, the internal consistency of the INQ-12 has been found to be adequate in the US [11,13], with Cronbach’s alphas between .87 and .93 for PB, and .80 to .92 for TB, but the INQ-12 showed poor factorial validity [11]. To summarize, the INQ-18 has demonstrated good internal consistency when used in China and the US. The INQ-12 has shown poor internal consistency when used with a Macau (China) sample but strong internal consistency in US samples. Moreover, the structural validity of the INQ-18 and INQ-12 does not fit the two-factor model hypothesized by the interpersonal-psychological theory of suicidal ideation [3], at least in North-American samples.

In contrast to the INQ-18 and INQ-12, the INQ-15 has shown good internal consistency and factorial validity, as well as concurrent and convergent validity in US populations [9, 11]. The current study tested the psychometric properties of the INQ-15 in Asia, in particular, Singapore. This is in line with the position held by Van Orden et al. [9], who suggested that there is a need for research on the INQ measurement model using samples of “increased diversity in terms of race/ethnicity” and that “before the INQ is used in other countries, the content validity of the INQ items for the specified culture should be examined.” (p. 212).

The lack of studies examining the construct validity of the INQ in Asia, and the finding regarding the low internal consistency of the TB scale in China call into question the construct validity of this subscale and invites exploration of its validity in Asia. Examining its factorial validity in an Asian context is a first step in testing the interpersonal psychological model [3]. It cannot be assumed that the latent concepts of PB and TB, which were conceptualized and tested on US samples, will be equivalent in Asian contexts. People in Western societies predominantly adopt an independent self-construal, characterized by the self as an autonomous entity and having attributes that drive the individual’s motives, behaviors and emotions. On the other hand, people in Asian societies are more likely to adopt an interdependent self-construal, whereby others are incorporated into one’s sense of self [14–17]. In Western societies, others are less centrally placed in one’s self-identity, whereas in Asian societies, there is a strong emphasis on family cohesion, self-restraint, and group harmony [18]. PB and TB might be an especially critical cultural risk factor for suicidal desire in Asians because of the importance of harmonious interpersonal relationships and group cohesion in their lives [10].

In the present study, it was hypothesized that the INQ-15 would fit a two-factor model (representing the PB and TB constructs) and have good internal consistency. It was also predicted a positive correlation between interpersonal needs, depression and problem-solving appraisal, and a negative correlation between interpersonal needs and hope. Additionally, good convergent validity was expected by finding high shared variance among the items forming each of the two factors of the model. Moreover, it was hypothesized that the scores in the INQ-15 would predict suicidal desire to a greater extent than suicidal plans. Discriminant validity was tested by demonstrating that the constructs of PB and TB are unique and different from the relationship between the two constructs.

Method

Results

CFA

AMOS version 21 was used to perform two CFAs with maximum likelihood estimation. The first CFA tested a one-factor model of general distress, in which the 15 items loaded on a single dimension, as in Freedenthal et al.’s study [13]. The second CFA tested the hypothesis that the INQ represented the PB and TB factors and that they were related to each other, as illustrated in the interpersonal-psychological theory of suicidal behavior [3]. Multiple measures were used to assess goodness-of-fit, such as χ², the standardized squared root mean squared residual (SRMR), root mean squared error of approximation (RMSEA), comparative fit index (CFI), and Tucker-Lewis index (TLI). SRMR and RMSEA values less than .08, and CFI and TLI values equal to or more than .90 were considered adequate.

The data did not fit a one-factor model of general distress, χ²(90) = 1219.23, p < .001, SRMR = .13, RMSEA = .19, CFI = .72, TLI = .67. The results, however, indicated that the data fit a two-factor model, χ²(88) = 373.24, p < .001, SRMR = .07, RMSEA = .10, CFI = .93, TLI = .92, after an error covariance between two items corresponding to the TB factor was estimated, as in Van Orden et al.’s study [9]. Table 1 presents the CFA two-factor model and standardized factor loadings, all p < .001. Factor loadings were large for PB and moderate for TB. As theorized, there was a positive correlation between the two latent constructs (r = .66, p < .001). The results of the present study were similar to those of the bootstrap sample, suggesting that the results may be replicable in a normal sample.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. INQ -15 Model estimated unstandardized factor loadings (+SE), standardized factor loadings, R² values, and item-total correlations.

https://doi.org/10.1371/journal.pone.0198839.t001

Internal consistency

Medium and large item-total correlations indicated that items in the two factors represented the constructs of PB and TB adequately (Table 1). The Cronbach’s alpha values for the INQ and the other scales used in this study are presented in Table 2. All the scales had good internal consistency.

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Descriptives, internal reliability and bivarate correlations between DASS-D, hope, PSI, SIS, and INQ scores.

https://doi.org/10.1371/journal.pone.0198839.t002

Discussion

The INQ-15 showed good construct validity, internal consistency, concurrent validity, convergent validity, predictive validity, and discriminant validity in a Singapore male young adult sample.

CFA confirmed adequate fit for a two-factor model, which was comprised of PB and TB. The CFA values were similar to the ones obtained in previous research [9,11], although the RMSEA values obtained in this study were slightly above the cut-off criterion. Standardized factor loadings obtained in the current study mirrored the ones obtained in the US by Van Orden et al. [9], although the average of standardized factor loadings for PB were higher in this study (.88) than in Van Orden et al.’s study (.70). The two research teams obtained similar averages of standardized factor loadings for TB (current study = .69; Van order et al. = .65). The findings suggest that PB is a more robust construct than TB in Singapore than in the US, and that PB is more important than TB in suicide ideation, as shown by the correlations in Table 2. Wong et al. [10] also found that PB was more important than TB in predicting suicide ideation in Asians. The correlation between PB and TB obtained in this sample was .66, similar to the value of .67 obtained by Van Orden et al. [9]. It is worth noting that Van Orden et al.’s sample was composed of only 33% males, while the sample used in the present study was made up solely of males. Considering that previous research [9,11] generalized the two-factor model to males and females, older, and clinical populations in the US, and that the present researchers found that the model fitted an Asian male-only sample, the results suggest that the model could generalize to a wide range of different populations.

In line with the interpersonal-psychological theory of suicidal behavior [3], the results of the current study suggest that the INQ more strongly predicts desire for suicide than for suicidal planning. Furthermore, the findings demonstrated that the INQ was more associated to the constructs of depression (r = .74) than hope (r = -.58) and problem-solving appraisal (r = .50). Also, the findings are consistent with research showing that suicide risk is present in depression, and that problem-solving issues and hopelessness increase the risk [5,33]. In this study, it was found that PB and depression shared a large quantity of variance (r² = .49), suggesting that depression symptoms are closely linked to PB. PB (r = .75) was more strongly associated with suicidal ideation than TB (r = .55), as in previous research conducted in Asia [8,10]. These findings have important clinical implications. To reduce the risk of suicide, therapy will need to focus on feelings of PB and depression, which might be exacerbated by life transitions and low controllability such as serving the army. Van Order et al. [9] found convergent validity between PB and autonomy, responsibility to family and self-competence. From a therapeutic perspective, dealing with issues related to autonomy, responsibility, and self-competence might decrease PB and the risk of suicidal ideation. Hopelessness and problem-solving appraisal are also important aspects to be considered during therapy. Problem-solving appraisal is associated with suicide ideation, and it is mediated by hopelessness [5]. When people feel they cannot cope with their problems, they can become hopeless about the future, and suicidal ideation might arise [5]. Therefore, therapeutic strategies designed to deal with problem-solving appraisal are critical to prevent hopelessness, which is a predictor of active suicidal ideation [4].

Limitations of this study are that it focuses on young adult males only, and the findings cannot be generalized across age groups, the female gender, or to clinical populations in Singapore. Future studies should aim to replicate the current findings in more heterogeneous samples across Asia.

In conclusion, the INQ has good psychometric properties and can be used to assess how young adult males living in the Singapore community perceive meeting interpersonal psychological needs. This is also the first step necessary to test the interpersonal-psychological theory of suicide in Singapore.