A total of 452 male (38.1%) and 733 female students (61.9%), aged between 18 and 24 years (M = 18.98, SD = 1.21), were recruited from a large public university in the Southwest of the US. Males (M age=19.12, SD = 1.27 years) and females (M age=18.89, SD = 1.17 years) differed significantly in age, t (1,185)=3.16, p = .002, Cohen’s d = 0.19. Informed consent were obtained, and participants completed a battery of questionnaires, including a short set of demographic information and the SBQ-R. Participation was voluntary and confidentiality was maintained by not including any identifiers in the survey form. Participants were awarded partial course credit for participation in the study. The university’s Institutional Review Board approved all the study procedures. Criteria for excluding participants in the study were the same as the Chinese sample (described below). No participant was excluded as none of the questionnaires were incomplete.

To lessen the potential impacts of unbalanced sample sizes between the Chinese and US samples, the SPSS random sampling extraction procedures extracted 2,000 students aged between 18 to 24 years (M = 20.70, SD = 1.35) from a dataset consisted of 11,806 participants from seven provinces in China (Ningxia, Shandong, Shanghai, Jilin, Qinghai, Shaanxi, and Xinjiang). The original dataset was part of a three-phase study on suicidality among college students in China, with a particular focus on examining the risk and protective factors associated with students’ suicidality. Participants were excluded if: (1) the responses were out-of-range, which may be due to data entry error; (2) over 30% of the items were incomplete on the SBQ-R; or (3) key demographic information, such as gender and age, was missing. Participation was voluntary, with written consent obtained from participants. Confidentiality was maintained by not including any identifiers in the survey form. Relevant university’s Institutional Research Board (Shandong University) approved all study procedures. Males (M age=20.75, SD = 1.36 years) and females (M age=20.67, SD = 1.35 years) did not differ significantly in terms of age, t (2000)=1.29, p = .195 (see Table 2).

Suicidal Behaviors Questionnaire-Revised (SBQ-R). The SBQ-R (Osman et al., 2001) is a scale designed to measure the suicide-related behavior construct with four different suicide-related behavior parameters: past suicide attempt, suicidal ideation, suicide threat, and likelihood of suicide attempt. It allows researchers and healthcare professionals to assess the level of severity of suicidality and specific risk factors. The scale is unidimensional and made up of four items: assessing suicidal ideation and attempts in a lifetime (Item 1); the frequency of suicidal ideation over the past 12 months (Item 2); the threat of suicide attempt (Item 3); and future likelihood of suicidal behavior (Item 4). A sample item is, “How likely is it that you will attempt suicide in the future?” Summing each item score gives the total score for the SBQ-R which ranges between 3 to 18. Higher scores indicate higher levels of suicidality. A score of 7 or above was classified as suicidal, whereas scores below 7 was classified as non-suicidal. The reliability estimate of the SBQ-R for the undergraduate sample was reasonable (Cronbach’s alpha=.76; Osman et al., 2001). In this study, the Cronbach alpha value was .73 (95% CI [.71, .75]) for the Chinese sample and .83 (95% CI [.81, .85]) for the US sample.

In the US sample, the original scale in English (Osman et al., 2001) was used. For the China sample, the Chinese version of the SBQ-R was translated by the Shandong University Centre for Suicide Prevention Research in 2016 by Chinese and US collaborative research teams. Two independent bilingual experts with extensive experience in the area of mental health forward-translated the individual items and the instructions into Chinese. Next, an expert, a native Chinese speaker with fluency in English, back-translated the Chinese version into English. The Chinese version of the SBQ-R was reviewed and discussed to enhance clarity and legibility prior to administration. It was first evaluated systematically in 2017 with approximately 2,074 college samples from two universities in Jinan, Shandong, with an acceptable Cronbach’s alpha coefficient of .67 (Lew et al., 2019). The translated version was accepted with no further changes made.

Statistical analysis was conducted using the SPSS V20 (IBM Corp., 2011). Data analytic strategy differed for continuous variables and categorical variables.

Continuous variables. Each question of the SBQ-R is examined in terms of its raw score. Specifically, Item 1 has a four-point response option, ranging from 1 (never) to 4b (attempted to kill, …hoped to die). Item 2 has a five-point response option with scores ranging from 1 (never) to 5 (very often). Item 3 has a three-point response option with scores ranging from 1 (no) to 3b (yes, …really wanted to die). Item 4 has a seven-point response option with scores ranging from 0 (never) to 6 (very likely). Each item represents a specific suicide-related behavior parameter: (1) Life-time ideation / plans / attempts; (2) Past year ideation frequency; (3) Life-time threat (inform someone); and (4) Future attempt likelihood.

Categorical variables. Categorical variables were formed from the SBQ-R items based on the following criteria:

Item 1. Participants who selected option 2 were included in the Lifetime ideation category. Those who selected either 3a or 3b were included in the Lifetime plan category. Those who selected either 4a or 4b were assigned to the Lifetime attempt category.

Item 2. Participants who selected options of 2 to 5 were included in the Past-year ideation category.

Item 3. Participants who endorsed response options 2a, 2b, 3a or 3b were categorized into Life-time threat (inform someone).

Item 4. Participants who chose response options 4, 5, or 6 were categorized into the Future attempt likelihood category. In addition to assessing responses to the individual SBQ-R items, the overall severity of suicide risk score was computed by summing scores of the individual items. Following the cutoff score proposed by Osman and colleagues (2001), total scores at or above 7 in the current study indicated a risk for suicide-related behaviors.

For all continuous variables, the following analyses were undertaken:

• 1

  US participants were compared with Chinese participants by gender and also by total sample.

• 2

  Independent sample t-testes were used to evaluate the level of significance.

For all categorical variables, the following analyses were undertaken:

• 1

  US participants were compared with Chinese participants by gender and also by total sample.

• 2

  Pearson Chi-Square was calculated to show the level of significance.

• 3

  Odds Ratio (OR) was calculated for each comparison.

Results are presented from Tables 3 to 5. Item-level comparison results are summarized as follows.

Continuous variables-lifetime ideation, plans, and attempts. Both the US male (1.57±0.80) and female (1.79±0.91) students (overall- US: 1.71±0.88, Chinese: 1.44±0.69, t (3185) = ±9.61, p < .001; Cohen’s d = 0.34) had significantly higher total suicidal ideation than Chinese male students (1.42±0.72), t (1239) = ±3.38, p< .001; Cohen’s d=0.20, and Chinese female students (1.45±0.67), t (1946) = ±9.45, p < .001; Cohen’s d = 0.43.

Categorical variables-lifetime ideation. Both the US and Chinese male students (25.2% and 24.1%, respectively), χ2(1239)=0.18 (n.s.) and the US and Chinese female students (29.7% and 27.9% respectively, χ2(1946)=0.72 (n.s.) reported similar lifetime ideation (overall- the US: 28.0%, Chinese: 26.5%, χ2(3185)=0.84 (n.s.). The odds of having lifetime ideation were all significantly higher for the US than Chinese students (overall and breakdown by sex).

Categorical variables-lifetime plans of suicide. The US male students reported higher lifetime plans of suicide than the Chinese male students (11.7% vs 4.6%, respectively), χ2(1239)=21.67, p < .001, with a significant OR of 0.36 (95% CI=0.23-0.56). Similarly, more US female students reported lifetime plans than the Chinese students (16.4% vs. 6.8%, respectively), χ2(1946)=45.10, p < .001, and with a significant OR of 0.37 (95% CI=0.28-0.50). Overall, US students reported more lifetime plans compared to the Chinese students (14.6% vs. 5.9%, respectively), χ2(3185)=67.82, p < .001, with a significant OR of 0.37 (95% CI=0.29-0.47).

Categorical variables-lifetime attempts. US male students did not differ significantly than Chinese male students in lifetime attempts of suicide (2.7% vs 2.9%), χ2(1239)=0.04 (n.s.), and with a non-significant OR of 1.10 (95% CI=0.54-2.24). However, more US female students reported lifetime suicide attempts than Chinese female students (5.6% vs. 1.2%, respectively), χ2(1946)=31.85, p < .001, with a significant OR of 0.20 (95% CI=0.11-0.36). Overall, more US students reported lifetime suicide attempts than Chinese students (4.5% vs. 1.9%, respectively), χ2(3185)=18.05, p < .001, with a significant OR at 0.40 (95% CI=0.26-0.62).

Overall, American students were at higher risk of suicide-related behaviors compared to Chinese students, at 24.3% compared to 17.0%, respectively. American students scored higher than Chinese students on: (1) life-time ideation, plan, and attempt; (2) past-year suicidal ideation; (3) life-time threat; and (4) total SBQ-R score. Despite no substantive differences on life-time ideation, Chinese students scored higher on the likelihood of future attempt item, and were 2.5 times more likely to be at risk of future attempts.

The mean SBQ-R total score of 5.12 for the US sample and findings indicating that 24.3% of American students are at risk of suicide-related behaviors in this study were similar to findings by Becker, Holdaway, and Luebbe (2017), who reported a mean SBQ-R total score of 5.17, constituting 26.1% of high-risk individuals among US college students.

Chinese students reported significantly fewer suicide-related behaviors such as lifetime ideation/plan/attempt, past-year ideation frequency and lifetime threat, reflected by significantly lower total SBQ-R scores than their US counterparts. Despite this, the likelihood of future attempt was significantly higher among the Chinese students compared to the American students.

Many factors could account for this discrepancy, such as cultural attitudes toward suicide. Much like Japan, there is substantial cultural tolerance in China for and permissive attitudes toward suicide and suicidal behavior/intent (Otsuka et al., 2020), with suicide historically even condoned as a morally responsible action (Kawashima, Kawamoto, Shiraga, & Kawano, 2019). The Chinese are much less predisposed toward help-seeking for emotional and mental distress compared to Americans. Mental issues remain a taboo issue in China, with less than 5% of Chinese respondents reporting lifetime mental health help-seeking (Liu et al., 2018). Hence, when young Chinese individuals suffer from mental health issues, they may have much less recourse to the traditional sources of support that are prevalent in Western society. For example, in spite of the cultural value emphasizing strong family cohesion, having a frank conversation with family and friends on mental health struggles remains far from the norm.

Another cultural reason may be due to Confucian ethics. Especially among Chinese females, they have been carefully groomed and moulded to subordinate to certain societal norms and are obliged to fulfil the expectations of their elders, family, and society (Zhang & Liu, 2012). This added pressure may exacerbate the typical stress-diathesis of already overwhelming pressure to cope with young adulthood and college life. Radical economic development may explain decreased suicide-related behaviors in China (Hu et al., 2015). Follett (2018) noted that the drop in suicide rate coincided with the large-scale migration of rural workers to cities since 2001, and Wang, Chan, and Yip (2014) suggested that the changes in individuals’ circumstances such as improved income, standards of living and education might have lowered the overall suicide rates. Filial piety (Chinese: Xiao), a central Confucian concept (Chao, 1994), which socialized Chinese people to consider how one’s action would affect their parents, may discourage suicide. Limited access to lethal means, such as firearms in China, may further discourage suicide. The lower rate of suicide-related behaviors in the Chinese college sample may also be mediated by higher socio-economic status, unlike the more disadvantaged general population (Vijayakumar, 2015).

Lastly, the prevalence of firearms in the US may elevate suicide risk and increase the rate of suicide. Firearms are a highly lethal method of suicide (Goldstein, Prater, & Wickizer, 2019), and experiences with firing a gun are associated with lifetime suicide attempts (Anestis & Capron, 2017). Over 60% of completed suicides in the US are by firearms (Ajdacic-Gross et al., 2008), and a meta-analysis study suggested that access to firearms increases the risk for completed suicide (Anglemyer, Horvath, & Rutherford, 2014).

Interventions to reduce suicide should not only focus on addressing items on the SBQ-R. Studies of adults and adolescents in Germany highlighted the importance of positive mental health in protecting against suicidal ideation (Teismann et al., 2018; Teismann, Brailovskaia, & Margraf, 2019). Depressed patients with and without pre-treatment for suicidal ideation were found to have similar psychotherapy outcomes, indicating that suicidal ideation may not be a detriment to depression recovery (Von Brachel, Teismann, Feider, & Margraf, 2019). Factoring in cultural differences, more research is needed to unravel the complex issues in suicide.

Several limitations should be considered when interpreting the results of the current findings. First, findings may not be generalizable to clinical groups. Second, the study did not include parameters such as suicide intent and correlates of suicide. Third, the study did not address the issue of measurement invariance, but this is fully addressed in another study in preparation by some of the authors of this paper. Indeed, it was the results of the presence of the differential item functioning of the SBQ-R items which had led to the focus of this paper to undertake the analyses at the item level. Suicide-related behavior construct could hold different meanings to the US and China groups such that it cannot be meaningfully compared (Putnick & Bornstein, 2016). Fourth, there were more Chinese students and they were slightly older than their US counterparts. Fifth, students in the US sample were awarded a partial course credit for their participation while students in China were not awarded, which could lead to biases in participant responses. Sixth, since data from the US involved one large public university while data from China was from seven provinces, generalizability of the findings may be affected by sampling bias. Despite these limitations, results point to the importance of traditional classical test theory for undertaking analyses at the item-level in order to evaluate the performances of individual items empirically (Bichi, 2016).

In conclusion, the study provided empirical evidence on the cross-cultural differences in suicide-related behavior parameters by making careful item-level comparisons on the well-validated SBQ-R across the two countries. Further investigations on why Chinese students reported a higher likelihood of future attempts compared to their US counterparts are needed. Factors such as psychological distress, suicide stigma, help-seeking behavior, and hope and hopelessness about the future (Han, Batterham, Calear, & Ma, 2018; Huen, Ip, Ho, & Yip, 2015; Tang, Byrne, & Qin, 2018) can be explored in future studies. Reasons behind higher suicide-related behaviors among female US students also warrants further investigation. Specific interventions designed to alleviate college student suicide-related behavior in the US and China and studies on particular protective and risk factors in different sub-groups across the countries should also be conducted.