More Interactions,More Prosociality? An Investigation of the Prosocial Effect of Online Social Interactions Among Adolescents

Abstract

Adolescence is a key stage for prosocial advancement. A plethora of research suggested that the consumption of prosocial-themed media content can facilitate the development of adolescents' prosociality. However, little has examined whether various behaviors of media use may play a role here, such as online social interactions (OSIs). This study aimed at exploring whether and how OSIs, an omnipresent activity we engage in on a daily basis, could enhance prosocial tendencies among adolescents. Empathy and trust were proposed as two possible explanatory variables linking OSIs and prosocial tendencies. A survey was conducted among 533 secondary school students (46.5 percent females; age: 12–18 years old, M_age = 14.82, SD_age = 1.77) in China, assessing their OSIs on Weibo, prosocial tendencies, trust, and empathy. The results confirmed that OSIs are positively associated with adolescents' prosocial tendencies, and this association goes through trust, not empathy. Further, it was found that age serves as a moderator such that the indirect association via trust is significant only for middle and late adolescents, not the early group. The findings provide preliminary evidence and a theoretical basis for further investigation of prosociality development from OSIs. It also opens up a possible direction of adolescents' prosociality promotion in practice.

Introduction

Broadly speaking, prosocial behaviors refer to the voluntary acts performed to benefit others, including helping, guiding, and rescuing others.^1,2 There is no doubt that prosocial behaviors are not only essential for the functioning of our society, but also important for individuals' personal growth and success,^3,4 particularly for adolescents. Research shows that adolescents' prosocial behaviors predict more peer acceptance,⁵ better academic performance,⁶ and greater achievement at later life stages.⁷

Moreover, given that adolescence is the stage when one's values and worldviews are being formed, it is often recognized as a key period for prosocial development.⁸ Therefore, considerable research endeavor has been devoted to exploring possible ways to nurture and advance adolescents' prosocial behaviors. Consuming prosocial-themed media is proved to be such an activity, including video game playing⁹ and media content exposure.^10,11

Research in prosocial development has long recognized that environment-related factors such as media exposure act as another significant contributor to one's prosocial behaviors, besides the genetic influence.¹² Given the prevalence of social media use, this study seeks to explore the impact of such an environmental factor and proposes that prosocial tendencies can be brewed by online social interactions (OSIs), a ubiquitous social media activity that we frequently engage in.

Our premise rests on the socialization theory, which posits that one's personality traits and behaviors are largely shaped by the socialization process they get involved in.¹³ When it comes to prosocial behaviors, the literature suggests that children and adolescents' prosociality develops through their socialization process during which ongoing and dynamic social interactions take place with other agents in society.¹⁴

The observed association between social interactions and prosociality would also make sense when considering the neurobiological finding that human connection behaviors such as touching and social interactions can release oxytocin, the so-called love hormone contributing to human social bonding.¹⁵ It was further found that oxytocin may play a role in regulating people's moral behaviors.¹⁶ These findings, along with the socialization theories, suggest that social interactions may foster prosocial tendencies. Similarly, a series of psychological research found that social exclusion decreases prosocial behaviors, as being excluded impairs one's capacity for an empathic understanding of others.¹⁷ Therefore, in this study, we seek to investigate whether this pattern could apply to OSIs.

The OSIs are often characterized by anonymity, fleetingness, and lack of physical cues.¹⁸ Nevertheless, a group of computer-mediated communication theories (i.e., the Social Information Process Theory, Hyperpersonal Model) suggests that the effectiveness of online communication can approach or even exceed that of the face-to-face one if given sufficient interaction time.^19,20 Therefore, we hypothesize that frequent OSIs are associated with stronger prosocial tendencies (H1).

If the association stands, we propose empathy and trust as two possible pathways explaining the association between OSIs and prosocial tendencies since both of them are identified as important precursors of prosocial behaviors in psychology research.²¹ First, empathy refers to an emotional state triggered by another's emotional state or situation, in which one internally stimulates the feelings of others.²² One strand of psychology research such as Hoffman's theory of empathy-based moral development²³ and Batson's empathy-altruism hypothesis²⁴ believes that the altruistic motivation in prosocial behaviors is based on empathy.²⁵

For example, several longitudinal studies indicated that children with stronger empathy easily apprehend others' experiences, acknowledge others' needs, and then are more likely to show prosocial behaviors.^2,26 Further, it was found that OSIs would foster a sense of connectedness, which is found to covary with empathy.²⁷ Therefore, we propose empathy as one explanatory pathway linking OSIs and prosocial tendencies (H2).

Another possible pathway proposed in this study is trust, which entails the “expectation of the persistence and fulfillment of the natural and the moral orders.”²⁸ Trust has been demonstrated as one of the strongest predictors of prosocial behaviors from both psychological and neurobiological perspectives.^29,30 For instance, psychological experiments were performed, revealing that higher levels of trust are more likely to elicit prosocial behaviors compared with low levels of trust.³¹ Following this line, a neural conceptual model and associated neuroimaging studies explained that trust feelings could activate one's reward system, which further leads to prosocial behaviors.³⁰ Further, it was found that social interactions bring about trust.^32,33 We thereby hypothesize that trust serves as another explanatory variable linking OSIs and prosocial tendencies (H3).

Adolescence is a period of significant development in one's life, involving profound changes throughout the whole trajectory.³⁴ Research shows that adolescents at various stages display different cognitions³⁵ and behaviors.³⁶ When it comes to prosociality, some research suggests a cubic trend such that prosocial behaviors first experience a decline in early- and middle adolescence, but they rebound in late adolescence.^37,38

Moreover, in prior research, age is often found to be a factor moderating the prosocial impact of examined predictors.³⁹ Following previous literature, we seek to examine whether the proposed association between OSIs and prosocial tendencies via empathy (RQ1a) and trust (RQ1b) would differ from early to late adolescence.

Materials and Methods

Platform selection

The platform selected for this study is Weibo, a Chinese Twitter-like social networking service. The reasons for our choice are twofold. First, Weibo is one of the largest social media platforms in China, hosting 566 million active users.⁴⁰ Second, Weibo is an open social media platform where individuals can have OSIs with any other members on the platform, including both friends and strangers. This allows us to measure generic OSIs.

Procedures

Procedures used in this study were approved by the IRB at the authors' institution (no. H2021177I). A pilot study was first carried out in April 2021 to test the validity and reliability of the scales among 94 students enrolled in a Chinese secondary school. All scales showed satisfactory reliabilities (α ≥ 0.80). The main survey was conducted in June 2021 in four secondary schools in China. To estimate the required sample size for the survey, a power analysis was performed with the statistical software G*Power.⁴¹ Since no relatable study can be identified to estimate the effect size, a safe standard (small: ρ² = 0.02) was used.

Thus, assuming an effect size of 0.02, an α value of 0.05, a power of 0.80, and seven predictors, G*Power program indicated that a sample size of 474 participants would be necessary. On parental consent guaranteed, a total of 540 students were contacted by the third author and all voluntarily participated in the study. The survey invitation and data collection all took place in students' classrooms. After delivering the instructions, the researcher left the classrooms to protect students' privacy. The survey was in paper-and-pencil format, with all the data entered manually into the computer later by research assistants.

Participants

Out of the 540 participants, 7 participants had missing values in their questionnaire and thus were excluded from data analysis. The final sample consisted of 533 students (46.5 percent females; age: 12–18 years old, M_age = 14.82, SD_age = 1.77). Most students' parents hold a bachelor's degree (34.1 percent), followed by a master's or a doctoral degree (26.6 percent), high school (9.2 percent), middle school (15.6 percent), and elementary school and below (14.4 percent).^a

Measures

The constructs in our study were measured by the following scales. Unless indicated, all scales used in this study were on a 5-point Likert scale ranging from “Strongly Disagree” to “Strongly Agree.”

The construct of prosocial tendencies was measured by a 15-item scale adapted from previous research^42–44 (α = 0.934). Sample items were “I prefer to donate money without anyone knowing” and “I think that helping others without them knowing is the best type of situation.”

The variable of OSIs was evaluated by using a 3-item scale assessing the frequency of liking, commenting, and private messaging on Weibo platform (1 = never, 5 = always; α = 0.923).

Empathy was measured by an eight-item scale^21,45 (α = 0.919). Sample items included “I sometimes try to understand my friends better by imagining how things look from their perspective” and “I would describe myself as a pretty soft-hearted person.”

Trust was assessed by using the 6-item scale developed by Yamagishi and Yamagishi⁴⁶ (α = 0.874). Since we are interested in trust toward the society as a whole, as opposed to specific people, items tapping general trust were selected. Sample items were “Most people are honest,” “Most people are trustworthy,” and “Most people are good and kind.”

Data about participants' age, gender, parents' education level, and exposure to prosocial content on Weibo were also collected as potential covariates. For the exposure to prosocial content on Weibo, we asked the participants to rate their frequency of browsing prosocial content on Weibo (1 = never, 5 = always).

Data analyses

In this study, all the data analyses were conducted with SPSS⁴⁷ and PROCESS.⁴⁸ Correlation, regression, and mediation analysis (PROCESS model) were performed to test hypotheses and research questions.

Results

Preliminary findings

A preliminary analysis was first carried out by running correlation analysis. As shown in Table 1, the variable of prosocial tendencies was significantly associated with OSIs (r = 0.212, p < 0.01), empathy (r = 0.448, p < 0.01), trust (r = 0.490, p < 0.01), and prosocial content exposure (r = 0.296, p < 0.01).

Table 1.

Mean, Standard Deviation, and Correlations of Key Constructs

	M	SD	1	2	3	4
1. Prosocial tendencies	3.529	0.903	1
2. OSIs	2.263	1.375	0.212^**	1
3. Empathy	3.777	0.927	0.448^**	0.127^**	1
4. Trust	3.623	0.985	0.490^**	0.207^**	0.384^**	1
5. Prosocial content exposure	2.510	1.343	0.296^**	0.433^**	0.236^**	0.241^**

p < 0.01.

OSIs, online social interactions; M, mean; SD, standard deviation.

Main findings

To test H1, we performed a regression analysis with age, gender, parents' education level, and exposure to prosocial content on Weibo as four covariates. The results indicated that OSIs are positively associated with prosocial tendencies (Coeff. = 0.088, p = 0.007). To alleviate the concern of the possible bidirectional correlation between OSIs and prosocial tendencies, we resorted to the instrument approach.^49,50 Specifically, we carried out the two-stage least squares regression analysis with the measure of generic Internet use as an instrumental variable. It is selected for being positively associated with OSIs, but not directly with prosocial tendencies considering that it is the socializing component of OSIs rather than the general Internet use that plays an important role in predicting prosocial tendencies. Following the prior research practice, we first obtained the predicted value of OSIs with the instrument variable (i.e., generic Internet use), and we then estimated the impact of OSIs on prosocial tendencies with the predicated value in the second-stage regression. A significant positive estimator at the second stage would indicate an impact from OSIs on prosocial tendencies.

As shown in Table 2, the first-stage analysis with the Internet use regressed on OSIs yielded significant results (Coeff. = 0.199, p < 0.001), with F value greater than the conventional cutoff point of 10, F(5, 527) = 53.110, p < 0.001.⁵¹ Then, the predicted value at the second stage was also found to be significant (Coeff. = 0.429, p = 0.010), lending support to our H1.

Table 2.

Results of Two-Stage Least Squares Analysis

Antecedents	First stage			Second stage
	(DV = OSIs)			(DV = prosocial tendencies)
	Coeff.	SE	p	Coeff.	SE	p
Constant	−3.298	0.476	<0.001	4.325	0.552	<0.001
Internet use	0.199	0.044	<0.001
Predicated value				0.429	0.166	0.010
Control variables
Age	0.284	0.030	<0.001	−0.143	0.052	0.006
Gender	−0.144	0.099	0.146	−0.039	0.081	0.632
Educational level of parents	0.001	0.033	0.968	0.042	0.025	0.098
Prosocial content exposure	0.277	0.039	<0.001	0.073	0.059	0.215
	R² = 0.335			R² = 0.110
	F (5, 527) = 53.110, p < 0.001			F (5, 527) = 12.963, p < 0.001

DV, dependent variable.

Next, we conducted the parallel mediator analysis using Model 4 in SPSS PROCESS to examine the indirect associations from OSIs to prosocial tendencies via empathy and trust. Demographic information and prosocial content exposure were entered as covariates. As Table 3 shows, although empathy was significantly related with prosocial tendencies (Coeff. = 0.274, p < 0.001), it was not significantly associated with OSIs (Coeff. = 0.036, p = 0.292). A 95 percent bias-corrected confidence interval (CI) based on 50,000 bootstrap samples indicated that the indirect effect through empathy (effect size = 0.010) straddled 0 (−0.009 to 0.032).

Table 3.

Results of the Parallel Multiple Mediator Model Analysis

Antecedents	Consequent
	Empathy			Trust			Prosocial tendencies
	Coeff.	SE	p	Coeff.	SE	p	Coeff.	SE	p
Constant	3.484	0.369	<0.001	4.060	0.385	<0.001	1.228	0.345	<0.001
OSIs	0.036	0.034	0.292	0.130	0.035	0.003	0.038	0.028	0.179
Empathy	—	—	—	—	—	—	0.274	0.038	<0.001
Trust	—	—	—	—	—	—	0.309	0.037	<0.001
Control variables
Age	−0.036	0.025	0.162	−0.091	0.026	0.006	−0.009	0.065	0.649
Gender	−0.059	0.079	0.451	−0.176	0.082	0.032	−0.030	0.065	0.649
Educational level of parents	0.084	0.027	0.002	0.075	0.027	0.007	−0.001	0.022	0.971
Prosocial content exposure	0.160	0.033	<0.001	0.154	0.034	<0.001	0.086	0.028	0.002
	R² = 0.280			R² = 0.333			R² = 0.584
	F (5, 527) = 8.971, p < 0.001			F (5, 527) = 13.162, p < 0.001			F (7, 525) = 38.857, p < 0.001

In contrast, the OSIs were significantly associated trust (Coeff. = 0.130, p = 0.003), which, in turn, had a significant impact on prosocial tendencies (Coeff. = 0.309, p < 0.001). The corresponding CI indicated that the indirect effect through trust (effect size = 0.040) was entirely above 0 (0.016–0.067). Thus, H2 was rejected, and H3 was supported.

To test the conditional indirect effects, Model 7 was then performed. Empathy and trust were entered as Mediator 1 and 2, respectively, with age as a moderator, and other variables as covariates. As displayed in Table 4, the interaction of OSIs and age is not significantly associated with empathy (Coeff. = 0.005, p = 0.799), but with trust (Coeff. = 0.062, p = 0.002). Next, the significance of indirect associations via empathy and trust was examined across different levels of age with estimated standard errors and the 95 percent Bootstrap CI.

Table 4.

Results of the Moderated Mediation Model Analysis

Antecedents	Consequent
	Empathy			Trust			Prosocial tendencies
	Coeff.	SE	p	Coeff.	SE	p	Coeff.	SE	p
Constant	3.635	0.699	<0.001	5.982	0.722	<0.001	1.1001	0.179	<0.001
OSIs	−0.039	0.295	0.895	−0.819	0.305	0.007	0.033	0.026	0.202
Empathy	—	—	—	—	—	—	0.274	0.038	<0.001
Trust	—	—	—	—	—	—	0.311	0.036	<0.001
Age	−0.045	0.047	0.329	−0.218	0.048	<0.001	—	—	—
OSIs × age	0.005	0.019	0.799	0.062	0.020	0.002	—	—	—
Control variables
Gender	−0.060	0.079	0.446	−0.187	0.081	0.022	−0.028	0.065	0.663
Educational level of parents	0.083	0.027	0.002	0.073	0.027	0.008	−0.001	0.022	0.972
Prosocial content exposure	0.159	0.033	<0.001	0.146	0.034	<0.001	0.083	0.027	0.002
	R² = 0.280			R² = 0.357			R² = 0.584
	F (6, 526) = 7.474, p < 0.001			F (6, 526) = 12.790, p < 0.001			F (6, 526) = 45.371, p < 0.001

As Table 5 shows, the index of moderated mediation showed that the 95 percent Bootstrap CI straddled 0 (−0.010 to 0.013) for the indirect association between OSIs and prosocial tendencies via empathy (Path 1), indicating that age did not moderate this indirect association. In contrast, the CI was entirely above 0 (0.005–0.349) for the indirect association via trust (Path 2), suggesting that age moderated this indirect association. Specifically, the indirect association through trust was only significant for middle (mean; 95 percent Bootstrap CI: 0.005–0.054) and late adolescents (mean +1 SD; 95 percent Bootstrap CI: 0.030–0.100), but not for early adolescents (Mean – 1 SD; 95 percent Bootstrap CI: −0.043 to 0.031).

Table 5.

Indices of Moderated Mediation with 95 Percent Bootstrap Confidence Intervals

Path 1: OSIs>empathy>prosocial tendencies
Index of moderated mediation: Index (0.001); BootSE (0.006); BootCI (−0.010 to 0.013)
Age	Effect	BootSE	BootLLCI	BootULCI
Mean −1 SD (13.047)	0.007	0.016	−0.025	0.038
Mean (14.818)	0.009	0.011	−0.011	0.031
Mean +1 SD (16.589)	0.011	0.013	−0.012	0.039

Path 2: OSIs>trust>prosocial tendencies
Index of moderated mediation: Index (0.019); BootSE (0.008); BootCI (0.005 to 0.349)
Age	Effect	BootSE	BootLLCI	BootULCI
Mean −1 SD (13.047)	−0.005	0.019	−0.043	0.031
Mean (14.818)	0.029	0.013	0.005	0.054
Mean +1 SD (16.589)	0.063	0.018	0.030	0.100

Discussion

By taking a media use perspective, the current investigation represents one early attempt to confirm that the generic OSIs can foster prosocial tendencies among adolescents, as well as their variation depending on different phases from early to late adolescence. Potential pathways underlying the association were also studied. The main findings are threefold: (1) Adolescents who frequently engage in OSIs are more likely to display prosocial tendencies. (2) Trust serves as an explanatory pathway such that OSIs form a sense of generalized trust that is further associated with stronger prosocial tendencies. (3) The aforementioned association and explanatory pathway were only observed among middle and late adolescents.

The findings greatly advance our knowledge of prosocial development among adolescents by identifying this ubiquitous behavior as a potential facilitator of prosociality. Theoretically, it opens up a research line investigating how prosociality can be advanced via online interacting behaviors and may serve as a theoretical basis for future research in the same line. Practically, the findings point to a possible direction that can be used to impel adolescents' prosocial behaviors.

Prosociality has long been a vibrant topic in psychology,⁴ sociology,⁵² and neurobiology.^53,54 Various methods and techniques are created to stimulate people's prosocial tendencies and behaviors.^55,56 The current study found that engaging in OSIs is positively associated with one's prosocial tendencies. This result is in line with several neurobiological findings that connection behaviors can function as a stimulus releasing oxytocin, a hormone found to drive human moral behaviors. It is interesting to observe that in online settings, where social interactions often occur anonymously and without any real-time nonverbal connection such as touching or eye contact, this association still can be replicated.

This suggests that the primary incentive of oxytocin may be rooted in the fundamental nature of social interactions: the sense of connectedness formed through social dynamics regardless of whether it is mediated or face-to-face communication. This speculation needs to be further tested by rigorous social experiments.

In addition, the present study also looks into the mechanism explaining the associations between OSIs and prosociality. It has been well established that both empathy and trust are important precursors to prosocial behaviors.^31,57 In this study, nevertheless, only trust was confirmed as a pathway explaining the associations between OSIs and prosocial tendencies. This finding seems to resonate with the aforementioned speculation that generic social interactions may foster a sense of connectedness with one another, which further builds generalized trust, an important basis for promoting prosocial behaviors.

Interestingly, empathy, a commonly observed predictor of prosociality among adolescents,⁵⁸ was not found to be a pathway explaining the examined associations. One possible explanation is that unlike the formation of trust, empathy development is often rooted in an understanding of others' experiences and perspectives. This understanding may require meaningful information exchanging and even self-disclosure, which may not necessarily happen in generic OSIs.

Lastly, our study observed differences across various adolescence phases. Specifically, the prosocial impact of OSIs via trust was only found in middle and late adolescence, with a greater effect size among the late group. One possible way to interpret the result is that early adolescents are too young to develop a generalized trust toward other people rather than their caring relationships such as parents and close friends. Therefore, OSIs' function in stimulating prosociality only applies among middle and late adolescents who are better ready to be a member of larger social circles and develop generalized trust.

Practically, the findings shed light on the positive role of OSIs in stimulating and reinforcing prosocial tendencies, which are particularly effective for middle and late adolescents. The social media platform designers can develop relevant features facilitating adolescents' OSIs to boost their trust and prosociality, such as creating heated topics or a gamification system to boost discussions.

The current study is limited in three ways. First, it is a cross-sectional study, implying that the causal relationship should be further validated in experimental or longitudinal designs. Second, the sample is limited in its monoethnic background. Generalization to other cultural and ethnic groups is needed in future research. Third, only parents' education level was collected as an indicator of the families' socioeconomic status (SES). A more accurate assessment of one's SES should be warranted in future research due to its close relationship with prosociality.⁵⁹ Lastly, we measured generic OSIs without further distinguishing specific types of OSIs, such as public broadcasting or private chatting. Future research could examine the roles played by different types of OSIs in stimulating prosocial behaviors.

In conclusion, prosocial development has been viewed as a significant research area in various fields. The current study represents one important attempt to explore the possibility that OSIs may augment adolescents' prosocial tendencies. Specifically, we demonstrated that adolescents' OSIs can act to foster trust, which is further associated with greater prosocial tendencies, particularly in middle and late age groups.

Note

a. For the item of parents' education level, we asked the participants to report the highest degree earned by either parent.

Footnotes

Author Disclosure Statement

No competing financial interests exist.

Funding Information

This work is funded by the Institute of College Student Development, Shanghai Jiao Tong University [DFY-LL-2020081] and Qixing Charity Communication Research Project, Civic Media Organization (CM Chuanbo) [2021-2022-A06].

References

Dovidio

, Piliavin

, Schroeder

, et al. (2006) The social psychology of prosocial behavior. Mahwah, NJ: Erlbaum.

Eisenberg

, Guthrie

, Cumberland

, et al. Prosocial development in early adulthood: a longitudinal study. Journal of Personality and Social Psychology, 2002; 82:993–1006.

Aknin

, Van de Vondervoort

, Hamlin

. Positive feelings reward and promote prosocial behavior. Current Opinion in Psychology, 2018; 20:55–59.

Penner

, Dovidio

, Piliavin

, et al. Prosocial behavior: multilevel perspectives. Annual Review of Psychology, 2005; 56:365–392.

Gerbino

, Zuffianò

, Eisenberg

, et al. Adolescents' prosocial behavior predicts good grades beyond intelligence and personality traits. Journal of Personality, 2018; 86:247–260.

Closson

LM.

Aggressive and prosocial behaviors within early adolescent friendship cliques: what's status got to do with it?. Merrill-Palmer Quarterly, 2009; 55:517–520.

Toumbourou

JW.

Beneficial action within altruistic and prosocial behavior. Review of General Psychology, 2016; 20:245–258.

Blakemore

, Mills

. Is adolescence a sensitive period for sociocultural processing?. Annual Review of Psychology, 2014; 65:187–207.

Qiu

, Zhen

, Zhou

, et al. Short-term prosocial video game exposure influences attentional bias toward prosocial stimuli. Cyberpsychology, Behavior and Social Network, 2020. DOI: 10.1089/cyber.2019.0745.

10.

Prot

, Gentile

, Anderson

, et al. Long-term relations among prosocial-media use, empathy, and prosocial behavior. Psychological Science, 2014; 25:358–368.

11.

Coyne

, Padilla-Walker

, Holmgren

, et al. A meta-analysis of prosocial media on prosocial behavior, aggression, and empathic concern: a multidimensional approach. Developmental Psychology, 2018; 54:331–347.

12.

Deater-Deckard

, Dunn

, O'Connor

, et al. Using the stepfamily genetic design to examine gene-environment processes in child and family functioning. Marriage and Family Review, 2003. DOI: 10.1300/J002v33n02_02.

13.

Grusec

, Hastings

, eds. (2014). Handbook of socialization: theory and research. New York, NY: Guilford Publications.

14.

Eisenberg

, Eggum-Wilkens

, Spinrad

. (2015). The development of prosocial behavior. In Schroeder DA, Graziano WG, eds. The Oxford handbook of prosocial behavior. Oxford: Oxford University Press, pp. 114–136.

15.

Uvnäs-Moberg

Oxytocin may mediate the benefits of positive social interaction and emotions. Psychoneuroendocrinology, 1998. DOI: 10.1016/S0306-4530(98)00056-0.

16.

Kosfeld

, Heinrichs

, Zak

, et al. Oxytocin increases trust in humans. Nature, 2005. DOI: 10.1038/nature03701.

17.

Twenge

, Ciarocco

, Baumeister

, et al. Social exclusion decreases prosocial behavior. Journal of Personality and Social Psychology, 2007. DOI: 10.1037/0022-3514.92.1.56.

18.

Green

, Clark

. (2015) Real or ersatz? Determinants of benefits and costs of online social interactions. In Sundar

, ed. The handbook of the psychology of communication technology. Malden, MA: Wiley Blackwell, pp. 247–269.

19.

Walther

, Burgoon

. Relational communication in computer-mediated interaction. Human Communication Research, 1992; 19:50–88.

20.

Walther

JB.

Computer-mediated communication: impersonal, interpersonal, and hyperpersonal interaction. Communication Research, 1996; 23:1–43.

21.

Eisenberg

Meta-analytic contributions to the literature on prosocial behavior. Personality and Social Psychology Bulletin, 1991; 17:273–282.

22.

Davis

MH.

Measuring individual differences in empathy: evidence for a multidimensional approach. Journal of Personality and Social Psychology, 1983; 44:113–126.

23.

Hoffman

MT.

Is altruism part of human nature?. Journal of Personality and Social Psychology, 1981; 40:121–137.

24.

Batson

, Lishner

, Stocks

. (2015). The empathy—Altruism hypothesis. In Schroeder DA, Graziano WG, eds. The Oxford handbook of prosocial behavior. Oxford: Oxford University Press, pp. 259–281.

25.

Eisenberg

, Miller

. The relation of empathy to prosocial and related behaviors. Psychology Bulletin, 1987; 101:91–119.

26.

Barr

, Higgins-D'Alessandro

. Adolescent empathy and prosocial behavior in the multidimensional context of school culture. Journal of Genetic Psychology, 2007; 168:231–250.

27.

James

, Davis

, Charmaraman

, et al. Digital life and youth well-being, social connectedness, empathy, and narcissism. Pediatrics, 2017; 140(Suppl 2):S71–S75.

28.

Barber

(1983). The logic and limit of trust. New Brunswick, NJ: Rutgers University Press.

29.

Parks

, Hulbert

. High and low trusters' responses to fear in a payoff matrix. Journal of Conflict Resolution, 1995. DOI: 10.1177/0022002795039004006.

30.

Declerck

, Boone

, Emonds

. When do people cooperate? The neuroeconomics of prosocial decision making. Brain and Cognition, 2013; 81:95–117.

31.

Irwin

Prosocial behavior across cultures: the effects of institutional versus generalized trust. Advances in Group Processes, 2009; 26:165–198.

32.

Singh

TB.

A social interactions perspective on trust and its determinants. Journal of Trust Research, 2012; 2:107–135.

33.

Duan

, Ye

, Poggi

, et al. Gaze towards my choice: noneconomic social interaction changes interpersonal trust only with positive feedback. Psychonomic Bulletin and Review, 2020; 27:1362–1373.

34.

Shefska

, Backes

. Fulfilling the promise of adolescence: applying developmental knowledge to create systems change. Journal of Youth Development, 2020. DOI: 10.5195/JYD.2020.946.

35.

Steinberg

Cognitive and affective development in adolescence. Trends in Cognitive Science, 2005; 9:69–74.

36.

Verrastro

, Albanese

, Ritella

, et al. Empathy, social self-efficacy, problematic internet use, and problematic online gaming between early and late adolescence. Cyberpsychology, Behavior and Social Network, 2021; 24:806–814.

37.

Eisenberg

, Cumberland

, Guthrie

, et al. Age changes in prosocial responding and moral reasoning in adolescence and early adulthood. Journal of Research Adolescence, 2005; 15:235–260.

38.

Eisenberg

, Eggum

. (2008) Empathy-related and prosocial responding: Conceptions and correlates during development. In Sullivan BA, Snyder M, Sullivan JL, eds. Cooperation: The political psychology of effective human interaction. Malden, MA: Blackwell Publishing.

39.

Putra

GNE

, Astell-Burt

, Cliff

, et al. The relationship between green space and prosocial behaviour among children and adolescents: a systematic review. Frontiers in Psychology, 2020; 11:859.

40.

Statista. (2021). China: MAU of Sina Weibo 2021. https://www.statista.com/statistics/1058070/china-sina-weibo-dau/ (accessed Dec. 5, 2021).

41.

Faul

, Erdfelder

, Lang

, et al. G* Power 3: a flexible statistical power analysis program for the social, behavioral, and biomedical sciences. Behavior Research Methods, 2007; 39:175–191.

42.

Carlo

, Randall

. The development of a measure of prosocial behaviors for late adolescents. Journal of Youth Adolescent. 2002. DOI: 10.1023/A:1014033032440.

43.

Carlo

, Hausmann

, Christiansen

, et al. Sociocognitive and behavioral correlates of a measure of prosocial tendencies for adolescents. Journal of Early Adolescence, 2003. DOI: 10.1177/0272431602239132.

44.

Brookhart

, Rassen

, Schneeweiss

Instrumental variable methods in comparative safety and effectiveness research. Pharmacoepidemiology and Drug Safety, 2010. DOI: 10.1002/pds.1908.

45.

Siu

AMH

, Shek

DTL

. Validation of the interpersonal reactivity index in a Chinese context. Research on Social Work Practice, 2005. DOI: 10.1177/1049731504270384.

46.

Yamagishi

, Yamagishi

Trust and commitment in the United States and Japan. Motivation and Emotion 1994. DOI: 10.1007/BF02249397.

47.

IBM Corp. IBM Corp. (2016). IBM SPSS statistics for Windows, Version 24.0. Armonk, NY: IBM Corp. (Released 2016). 2016 2016.

48.

Hayes

AF.

(2018) Introduction to mediation, moderation, and conditional process analysis: a regression-based perspective. 2nd ed. New York, NY: The Guilford Press.

49.

Heckman

, Vytlacil

, Urzua

. Understanding instrumental variables in models with essential heterogeneity. SSRN Electronic Journal, 2021; 19:537–554.

50.

Angrist

, Krueger

. Instrumental variables and the search for identification: from supply and demand to natural experiments. Journal of Economic Perspectives 2001. DOI: 10.1257/jep.15.4.69.

51.

Stock

, Wright

, Yogo

A survey of weak instruments and weak identification in generalized method of moments. Journal of Business and Economic Statistics 2002. DOI: 10.1198/073500102288618658.

52.

Simpson

, Willer

Beyond altruism: sociological foundations of cooperation and prosocial behavior. Annual Review of Sociology 2015. DOI: 10.1146/annurev-soc-073014-112242.

53.

Marsh

, Marsh

, Lee

, et al. Oxytocin and the neurobiology of prosocial behavior. Neuroscientist, 2020; 27:604–619.

54.

Chen

, Ye

, Liu

, et al. Editorial: the neural basis of human prosocial behavior. Frontiers in Psychology, 2019; 10:2058.

55.

Laguna

, Mazur

, Kędra

, et al. Interventions stimulating prosocial helping behavior: a systematic review. Journal of Applied Social Psychology 2020. DOI: 10.1111/jasp.12704.

56.

Saleme

, Pang

, Dietrich

, et al. Prosocial digital games for youth: a systematic review of interventions. Computers in Human Behavior Reports 2020. DOI: 10.1016/j.chbr.2020.100039.

57.

Eisenberg

, Fabes

. Empathy: conceptualization, measurement, and relation to prosocial behavior. Motivation and Emotion 1990. DOI: 10.1007/BF00991640.

58.

McMahon

, Wernsman

, Parnes

. Understanding prosocial behavior: the impact of empathy and gender among African American adolescents. Journal of Adolescent Health 2006. DOI: 10.1016/j.jadohealth.2005.10.008.

59.

Andreoni

, Nikiforakis

, Stoop

Higher socioeconomic status does not predict decreased prosocial behavior in a field experiment. Nature Communications 2021. DOI: 10.1038/s41467-021-24519-5.