This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Journal of Medical Internet Research, is properly cited. The complete bibliographic information, a link to the original publication on http://www.jmir.org/, as well as this copyright and license information must be included.
A limited number of studies have examined the differences in suicide-related social media use behaviors between suicide ideators and suicide attempters or have sought to elucidate how these social media usage behaviors contributed to the transition from suicidal ideation to suicide attempt.
Suicide attempts can be acquired through suicide-related social media use behaviors. This study aimed to propose 3 suicide-related social media use behaviors (ie, attending to suicide information, commenting on or reposting suicide information, or talking about suicide) based on social cognitive theory, which proposes that successive processes governing behavior transition include attentional, retention, production, and motivational processes.
We aimed to examine the mediating role of suicide-related social media use behaviors in Chinese social media users with suicidal risks. A sample of 569 Chinese social media users with suicidal ideation completed measures on suicidal ideation, suicide attempt, and suicide-related social media use behaviors.
The results demonstrated that suicide attempters showed a significantly higher level of suicidal ideation (
Our findings thus support the social cognitive theory, and there are implications for population-based suicide prevention that can be achieved by identifying behavioral signals.
Suicide has been a critical public health problem, with approximately 1 million people committing suicide worldwide annually [
The ideation-to-action framework states that the development of suicidal ideation and the progression from ideation to suicide attempt are two distinct processes [
Today, the internet can provide new data sources because people record their lives in varying degrees on the website, and it has become an active interaction platform for young people, where they sometimes exchange thoughts portraying a susceptibility to suicidality [
People like to create and exchange user-generated content on social media [
Attentional process is the first step of observational learning. The attentional process revolves around what people selectively pay attention to about the observed models and what message they procure from a wealth of information. Preconceptions, value preferences, and other factors determine this process [
The processes for observational learning are retention and production. Retention involves the process of remembering the modeled activities. This process would involve not only simply copying but also proactively reconstructing the observed events. Production would be the deep processing of retention when learned behavior is generated through a “conception-matching process” (ie, through a production process, cognitive activities are developed into corresponding explicit behaviors) [
Finally, the motivational process is the final stage to make the decision on whether one would act out on the acquired modeled behaviors. Social cognitive theory differentiates internal acquisition from external behaviors because people would not show all the things that they have ever learned. They would have a higher possibility of performing the modeled behaviors that they felt were similar to their situation, and they value the outcomes [
In summary, the mechanism of the progression from suicide ideators to suicide attempters among social media users is unclear. On the basis of earlier research [
There are 2 hypotheses for suicidal social media users:
H1: Compared with suicide ideators, suicide attempters would report a higher level of suicide-related social media use behaviors, including
H2: Suicidal ideation would predict suicide attempt through the mediating chains of
As the most popular social media platform in China, Sina Weibo has nearly 300 million users, and most of them are younger than 30 years [
We sent a direct message that provided social support, referrals, and a link to questionnaires to these 4616 suicidal users. More details can be found in the study by Liu et al [
The demographic information of participants was collected, including their sex, age, education level, marital status, and living status (living alone, with family or partner, with friends, or with others).
Suicidal ideation was measured by the 4-item version of the Adult Suicidal Ideation Questionnaire (ASIQ) [
Suicide attempt was measured by one item “Have you ever tried to kill yourself?” Participants responded with binary choices (yes or no). As suicide attempt is a binary dependent variable, participants responded no (0) or yes (1) for suicide attempt. Participants who responded no to the suicide attempt item were grouped as suicide ideators, and those who responded yes to the suicide attempt item were categorized as suicide attempters [
Suicide-related social media use behaviors were developed by our team. This measure consists of
Mplus7 (Muthen and Muthen, Beijing, China) and SPSS21 (SPSS Inc, Beijing, China) were used for the statistical analysis. Descriptive statistics for demographic characteristics and study variables were also tabulated. Exploratory factor analysis (EFA) and confirmatory factor analysis (CFA) were conducted to examine the psychometric properties of the self-developed measure of suicide-related social media use behaviors.
The flowchart for data analyses. CFA: confirmatory factor analysis; EFA: exploratory factor analysis; SEM: structural equation modeling; H1: hypothesis 1; H2: hypothesis 2.
After excluding respondents that reported conflicting or missing data (146/569, 20.1%) and no suicidal ideation (the ASIQ score=0; 10/569, 1.4%), there were 569 valid participants who reported a certain degree of suicidal ideation. This number exceeds the required sample size to test the study question, as the minimum sample size is 295 (power=0.80; alpha=.05; H0: RMSEA=0, H1: RMSEA=0.05; and df=36) [
Demographic characteristics of participants by suicidal status (N=569).
Characteristic | Total (N=569), n (%) | Suicide ideators (n=277), n (%) | Suicide attempters (n=292), n (%) | Chi-square value ( |
||||
|
|
|
||||||
|
Male | 78 (13.7) | 40 (14.4) | 38 (13.0) |
|
|
||
|
Female | 491 (86.3) | 237 (85.6) | 254 (87.0) |
|
|
||
|
|
|
||||||
|
Primary | 28 (4.9) | 11 (4.0) | 17 (5.8) |
|
|
||
|
Secondary | 97 (17.1) | 53 (19.1) | 44 (15.1) |
|
|
||
|
Tertiary | 444 (78.0) | 213 (76.9) | 231 (79.1) |
|
|
||
|
|
|
||||||
|
Single | 531 (93.3) | 256 (92.4) | 275 (94.2) |
|
|
||
|
Married | 24 (4.2) | 12 (4.3) | 12 (4.1) |
|
|
||
|
Divorced or others | 14 (2.5) | 9 (3.3) | 5 (1.7) |
|
|
||
|
|
|
||||||
|
With families | 201 (35.3) | 90 (32.5) | 111 (38.0) |
|
|
||
|
With partner | 33 (5.8) | 16 (5.8) | 17 (5.8) |
|
|
||
|
With friends | 231 (40.6) | 115 (41.5) | 116 (39.7) |
|
|
||
|
Alone | 75 (13.2) | 44 (15.9) | 31 (10.6) |
|
|
||
|
Others | 29 (5.1) | 12 (4.3) | 17 (5.8) |
|
|
To examine its psychometric properties, we split the sample (n=569) into two random and equal halves. We randomly used half of the sample to conduct EFA and the other half to conduct CFA. See
The CFA model showed a good fit to the data: χ26=2.1 (
On the basis of whether the participants attempted suicide or not, they were grouped as suicide ideators or suicide attempters as mentioned above. The group differences on study variables by suicide status are shown in
Descriptive statistics and correlations of study variables are shown in
Exploratory factor analysis of 6-item suicide-related social media use behaviors.
Items | Attending toa | Commenting-repostingb | Talking aboutc |
Attending to suicide news | 0.88 | N/Ad | N/A |
Attending to friends who said they wanted to commit suicide | 0.74 | N/A | N/A |
Commenting on or reposting suicide news | —e | 0.71 | N/A |
Commenting on or reposting other people’s posts about killing themselves | N/A | 0.90 | N/A |
Talking about suicide in online suicide communities | N/A | N/A | 0.85 |
Talking on the website about one’s own concrete plan to commit suicide | N/A | N/A | 0.82 |
aAttending to suicide information.
bCommenting on or reposting suicide information.
cTalking about suicide.
dN/A: not applicable.
eLoading <0.04 was not shown in the table.
Group differences on study variables by suicidal status (N=569).
Variables | Suicide ideators (n=277), mean (SD) | Suicide attempters (n=292), mean (SD) | Effect size | ||
Suicidal ideation | 2.34 (1.69) | 3.10 (1.92) | −5.04 (563.64) | <.001 | 0.42 |
Attending toa | 2.86 (0.77) | 3.00 (0.91) | −1.94 (567) | .05 | 0.17 |
Commenting-repostingb | 2.41 (0.95) | 2.59 (1.04) | −2.12e (567) | .03 | 0.18 |
Talking aboutc | 1.70 (0.88) | 2.14 (1.15) | −5.12a (542.22) | <.001 | 0.43 |
aAttending to suicide information.
bCommenting on or reposting suicide information.
cTalking about suicide.
Descriptive statistics and bivariate correlations of the study variables.
Variables | Correlation coefficient ( |
Value, mean (SD) | |||||
|
Suicidal ideation | Suicide attempt | Attending toa | Commenting-repostingb |
|
||
|
|
||||||
|
|
N/Ac | N/A | N/A | N/A |
|
|
|
N/A | N/A | N/A | N/A |
|
||
|
|
||||||
|
|
0.21 | N/A | N/A | N/A |
|
|
|
<.001 |
|
|
|
|
||
|
|
||||||
|
|
0.31 | 0.08 | N/A | N/A |
|
|
|
<.001 | .05 |
|
|
|
||
|
|
||||||
|
|
0.30 | 0.09 | 0.52 | N/A |
|
|
|
<.001 | .03 | <.001 |
|
|
||
|
|
||||||
|
|
0.46 | 0.21 | 0.46 | 0.56 |
|
|
|
<.001 | <.001 | <.001 | <.001 |
|
aAttending to suicide information.
bCommenting on or reposting suicide information.
cN/A: not applicable.
dTalking about suicide.
Results of SEM are depicted in
The mediating effect of social media use behaviors in the association between suicidal ideation and suicide attempt. a) Attended to suicide information; b) Commented on or reposted suicide information; c) Talked about suicide; d)
Further mediation analysis with the Monte Carlo method found significant indirect effects of (1)
Estimated mediation effects of attending to, commenting-reposting, and talking about behaviors in the association between suicidal ideation and suicide attempt.
Mediation effect | Indirect effect | 95% CI | |
Suicidal ideation→ |
0.19 | 0.11 to 0.27 | <.001 |
Suicidal ideation→ |
−0.10 | −0.50 to 0.31 | .64 |
Suicidal ideation→ |
−0.16 | −0.48 to 0.16 | .32 |
Suicidal ideation→ |
0.22 | 0.09 to 0.36 | .001 |
aAttending to suicide information.
bCommenting on or reposting suicide information.
cTalking about suicide.
On the basis of the social cognitive theory, this study aimed to identify web-based behavioral markers in social media that could distinguish suicide ideators from suicide attempters and illuminate the transition mechanisms that led from suicidal ideation to suicide attempt. Our findings demonstrated that when compared with Chinese social media users who were suicide ideators, suicide attempters reported more suicide-related social media use behaviors. More importantly, the effect of suicidal ideation on suicide attempt was mediated by suicide-related social media use behaviors. Our findings, therefore, make a theoretical contribution to the field by providing behavioral markers in the progression from suicidal ideation to suicide attempt with the Chinese social media user population that has suicidal ideation. To our knowledge, this is the first study to focus on the identification of suicide-related social media use behaviors and illustration of the progression from suicidal ideation to suicide attempt.
Consistent with previous studies [
Interestingly, our findings demonstrated that suicidal ideation predicted suicide attempt through the mediating chains of social media use behaviors, which included the
Our mediation findings also provide insight into explaining how suicide capacity is acquired in social interaction online, potentially complementing the ideation-to-action framework [
Although the study has made several contributions to the existing research, there are several limitations to this work. First, participants in this study were mainly suicidal, unmarried young females with college degrees. Our results are consistent with previous studies showing that single young females with a higher education were more inclined to talk about their suicidal ideation and seek help [
This study aimed to investigate the behavioral markers available for distinguishing suicide ideators from suicide attempters and elucidate the behavioral process that transits from suicidal ideation to suicide attempt. Our findings demonstrated that suicide attempters showed a significantly higher level of suicidal ideation and more suicide-related social media use behaviors, including
The measurement items for the whole study.
Adult Suicidal Ideation Questionnaire
confirmatory factor analysis
comparative fit index
exploratory factor analysis
root mean square error of approximation
structural equation modeling
Tucker Lewis Index
weighted root mean square
The authors would like to gratefully acknowledge the generous support of the National Basic Research Program of China (2014CB744600), China Social Science Fund (17AZD041), National Social Science Fund of China (16AZD058), National Natural Science Foundation of China (31700984), and the Research Grants Council of the Hong Kong Special Administrative Region, China (Collaborative Research Fund, Project No. C1031-18G). The sponsors had no role in the study design; in the collection, analysis, and interpretation of data; in the writing of the report; or in the decision to submit the paper for publication.
None declared.