亲社会行为泛指一切符合社会期望，对他人、群体和社会有益的行为，是发生在人际之间的积极社会互动，表现为捐赠、志愿服务、合作等多种形式。影响亲社会行为的因素很多，随着贫富差距的扩大和社会阶层差异的日益凸显，个人所属社会阶层对亲社会行为的影响逐渐成为研究者关注的热点。社会阶层(也常被称为社会经济地位)是人们在社会层级结构中的等级和地位，它既取决于个体所拥有的物质与社会资源(收入、受教育水平和职业常被用作衡量指标，其中收入或物质资源是界定社会阶层的核心因素)，也包括个体通过社会比较而感知到的自己在社会中的相对位置。但关于社会阶层影响亲社会行为的研究主要基于实施者视角，重点探讨了实施者社会阶层、接受者社会阶层以及施-受双方社会阶层的交互作用对实施者亲社会行为的影响。亲社会行为是一种人际互动行为，它不仅涉及实施者、接受者与旁观者的一次行动，而且还会随着接受者或旁观者转换为实施者而变成持续发生的行为。也就是说，亲社会行为往往不是单一行动，而很可能是连锁行动。

我们认为，由实施者的亲社会行为引发的接受者对一般他人实施的亲社会行为(即接受者亲社会行为传递)和旁观者以实施者为榜样而向一般他人实施的亲社会行为(即旁观者亲社会行为依从或从众)都可被看作是亲社会行为的广义传递(不包括接受者回报实施者的行为和旁观者帮助实施者的行为)。鉴于对前人相关研究的分析，本研究主要探讨发生在陌生人之间的亲社会行为广义传递现象，具体来说，是基于亲社会互动的不同角色，探讨实施者社会阶层是否以及如何影响接受者和旁观者的指向一般他人的亲社会行为，并探讨经济不平等凸显对其的调节作用。

研究1探讨实施者社会阶层对接受者亲社会行为传递的影响，以及感恩的中介作用和经济不平等凸显的调节作用，共包含5项子研究(总样本量N = 1154)。研究1a使用问卷星平台在线招募成人被试(N = 167)，采用第三人称假设情境和投射测验，通过收入、受教育水平和职业，操纵实施者的社会阶层(低，高)，测量被试推测情境中的主人公在大病中受到高/低阶层者大额捐助后的感恩情绪及指向一般他人的亲社会行为，初步检验实施者社会阶层与接受者亲社会行为传递的关系和感恩的中介作用。研究1b使用Credamo平台在线招募大学生被试(N = 160)，采用第一人称假设情境，通过父母的收入、受教育水平和职业地位，以及实施者的日常消费水平操纵实施者的社会阶层(低，高)，测量被试在假设的日常生活情境中受到高/低阶层者小额帮助后的感恩情绪及指向一般他人的亲社会行为，并检验中介模型。研究1c使用Credamo平台在线招募大学生被试(N = 159)，采用在线互动的实验情境，使用独裁者博弈中的金钱分配测量被试实际的亲社会行为，通过父母的收入、受教育水平和职业地位，以及实施者的日常消费水平操纵实施者的社会阶层(低，高)，让被试首先作为接受者受到高/低阶层者的慷慨金钱分配，测量被试的感恩情绪，然后让被试作为分配者给另一轮实验中的他人分配金钱，从行为层面再次验证中介模型。研究1d使用问卷星平台在线招募成人被试(N = 350)，采用同研究1a的第三人称假设情境和投射测验，同时操纵实施者的社会阶层(低，高)与经济不平等凸显(凸显，控制)，检验经济不平等凸显对实施者社会阶层影响接受者亲社会行为传递的调节作用。研究1e使用Credamo平台在线招募大学生被试(N = 318)，采用和研究1b相同的第一人称假设情境，同时操纵实施者的社会阶层(低，高)与经济不平等凸显(凸显，控制)，进一步检验经济不平等凸显对实施者社会阶层影响接受者亲社会行为传递的调节作用。研究1a-1e的五项研究结果均表明，与高阶层者相比，低阶层实施者的亲社会行为使接受者产生更高的感恩情绪，进而愿意做出更多的指向他人的亲社会行为，即传递亲社会行为，而且感恩在实施者社会阶层影响接受者亲社会行为传递之间起中介作用。研究1d和1e的结果均表明，经济不平等凸显并不调节实施者社会阶层对接受者感恩情绪和亲社会行为传递的影响。

研究2探讨实施者社会阶层对旁观者亲社会行为依从的影响，以及道德提升的中介作用和经济不平等凸显的调节作用，共包含5项子研究(总样本量N = 1073)。研究2a使用Credamo平台在线招募成人被试(N = 240)，让被试阅读新闻报道中的捐款助人材料，通过收入、受教育水平和职业操纵实施者的社会阶层(低，高)，并控制实施者亲社会行为的代价(捐赠数额占高低阶层者年可支配收入的比例相同或数额相同)，把被试随机分入高阶层-相同捐赠数额组(捐赠1250元的高阶层者)、高阶层-相同捐赠比例组(捐赠10万元的高阶层者)，及低阶层组(捐赠1250元的低阶层者)中，测量被试作为旁观者看到新闻中的高/低阶层者捐款后的道德提升情绪以及指向一般他人的亲社会行为。初步验证实施者的社会阶层与旁观者亲社会行为依从的关系和道德提升的中介作用。研究2b使用Credamo平台在线招募成人被试(N = 239)，同研究2a一样，采用新闻报道中的捐款助人材料，操纵实施者的社会阶层(低，高)，并控制实施者亲社会行为的代价，把被试随机分入高阶层-相同捐赠数额组、高阶层-相同捐赠比例组或低阶层组中，测量被试作为旁观者看到新闻中的高/低阶层者捐款后的道德提升情绪以及向慈善机构实际捐赠的金钱数额，在行为层面检验中介模型。研究2c使用Credamo平台在线招募成人被试(N = 116)，采用回忆书写任务，让被试回忆自己在生活中看到的高/低社会阶层者捐款助人的情境，测量被试作为旁观者的道德提升情绪以及向慈善机构实际捐赠的金钱数额，再次检验中介模型。研究2d使用Credamo平台在线招募成人被试(N = 318)，采用同研究2a中的捐款助人的新闻材料和亲社会行为的测量方式，同时操纵实施者的社会阶层(低，高)与经济不平等凸显(凸显，控制)，控制实施者亲社会行为的代价(高/低阶层者-相同捐赠比例)，检验经济不平等凸显对实施者社会阶层影响旁观者亲社会行为依从的调节作用。研究2e使用Credamo平台在线招募成人被试(N = 320)，采用同研究2b中的捐款助人的新闻材料，测量被试向慈善机构实际捐赠的金钱数额，同时操纵实施者的社会阶层(低，高)与经济不平等凸显(凸显，控制)，控制实施者的亲社会行为代价(高/低阶层者-相同捐赠比例)，进一步检验经济不平等凸显对实施者社会阶层影响旁观者亲社会行为依从的调节作用。研究2a-2e的五项研究结果均表明，无论捐赠数额相同，还是捐赠比例相同，与高阶层者相比，低阶层者的捐款行为使旁观者产生了更高的道德提升，进而表现出更多的指向他人的亲社会行为，而且道德提升在实施者社会阶层影响旁观者亲社会行为依从之间起中介作用。研究2d和2e的结果均表明，经济不平等凸显不调节实施者社会阶层对旁观者道德提升和亲社会行为依从的影响。

在贫富差距较大和社会阶层差异凸显的社会背景下，本研究基于亲社会互动的接受者视角和旁观者视角，探讨了实施者社会阶层对亲社会行为广义传递的影响。研究从宏观层面的经济不平等、中观层面的亲社会情境，以及微观层面的实施者社会阶层、接受者的感恩情绪、旁观者的道德提升情绪等多层面和多角度的交互的框架下，剖析了陌生人之间亲社会行为传递的机制。本研究也使用多样化的亲社会行为情境和亲社会行为测量指标对研究假设进行了反复验证，研究结果具有可重复性和较好的生态效度。本研究结果对于理解复杂生态环境中的人际互动现象，揭示亲社会互动的机制具有理论价值；对于如何激发接受者和旁观者更多的亲社会行为，提高社会环境中的公民社会参与，维持人际、群际和社会和谐都具有现实启发作用，也为提高社会治理和社会心理服务提供了有效建议。

Prosocial behavior is broadly defined as the act that benefits others, groups or society. Prosocial behavior is a positive interpersonal interaction, manifested in various forms such as donation, volunteering and cooperation. However, most of the previous research on prosocial behavior focused on the perspective of actor but overlooked the perspectives of recipient and bystander, examing how actors’ characteristics influences their own prosocial behavior, how recipiets’ characteristics and the interaction of actors’ and recipients’ characterstics influence actors’ prosocial behavior. Actually, actors’ prosocial behavior can elicit recipients’ and bystanders’ prosocial behavior, including recipients’ and bystanders’ prosocial behavior towards general other third parties. We define recipients’ paying the prosociality forward, and bystanders’ prosocial conformity as generalized transmission of prosocial behavior, which excludes recipients’ and bystanders’ prosocial behavior towards the intial prosocial actor.

In recent years, the gap between the rich and the poor has gradually become prominent across the globe along with the development of social economy. Social class has become an important social identity characteristic of individual. Social class refers to an individual’s rank in the social hierarchy and is predominantly rooted in experiences of different levels of material resources (a composition of objective indicators in terms of income, education, and occupational prestige), and is rooted in the construal of one’s place in a class hierarchy vis-a`-vis others. An emerging body of research has showed that social class shapes a diverse array of domains that include psysical health, sujective well-being, cognitive preference, and prosocial behavior.

Previous research mainly explored how actors’ social class influences their own prosocial behavior, how recipients’ social class elicits actors’ prosocial behavior, and how actors’ and recipients’ social class interact to affect actors’ prosocial behavior. From the perspective of generalized transmission of prosocial behavior, the current research aims to examine the effects of actors’ social class on recipients’ paying the prosociality forward, and on bystanders’ prosocial conformity, as well as their respective mediating mechanism. In addition, we test the moderating role of economic inequality salience in these effects.

Study 1 (1a-1e, N = 1154) focused on the recipients’ perspective, examining the effect of actors’ social class on recipients’ prosocial behavior towards general other third parties (i.e., pay it forward), testing the mediating role of gratitude and the moderating role of economic inequality salience. In Study 1a, we aimed to provide initial evidence for the hypothesis that actors’ social class reduce recipients’ prosocial behavior towards others and that gratitude mediates this effect. We recruited adult participants (N = 167) and asked them to read one of two scenarios in which the protagonist received help from a high-class or low-class actor. Then, we measured participants’ perceived gratitude of the protagonist towards the actor and his prosocial behavior towards others. In study 1b, we aimed to further examine our hypothesized mediation model. We recruited college students as participants (N = 160) and asked them to read one of two scenarios that manipulated actors’ social class (high vs. low), and to imagine that they received help from a high-class or low-class actor in the scenario. Then, we measured participants’ gratitude towards the actor and prosocial intention towards general others. In study 1c, we aimed to further examine our hypothesized mediation model by measuring participants’ actual prosocial behavior. We recruited college students as participants (N = 159) and used allocation of money in the dictator game to indicate prosocial behavior. In round 1, participants received generous allocation of money from a high-class or low-class allocator. Afterwards, their gratitude towards the allocator was measured. In round 2, participants were asked to divide additional money with a new anonymous person. In study 1d, we recruited adult participants (N = 350) and used the same scenarios as those in Study 1a that manipulated actors’ social class (high vs. low). In addition, we manipulated economic inequality salience (control vs. salient) to test whether it moderates the effects of actors’ social class on recipients’ gratitude towards them and prosocial behavior towards others. In Study 1e, we recruited adult participants (N = 318) and used the same scenarios as those in Study 1b that manipulated actors’ social class (high vs. low). In addition, we manipulated economic inequality salience (control vs. salient) to further test whether it moderates the effects of actors’ social class on recipients’ gratitude towards them and prosocial behavior towards others. Across five studies, we find that recipients’ gratitude mediates the effect of actors’ social class on recipients’ prosocial behavior towards others. That is, compared to a high-class actor, the recipient feels more grateful towards a low-class actor, and thus engage in more prosocial behavior towards others. Across Studies 1d and 1e, we find that economic inequality salience does not moderate the effects of actors’ social class on recipients’ gratitude towards them and prosocial behavior towards others.

Study 2 (2a-2e, N = 1073) focused on the bystanders’ perspective, examining the effect of actors’ social class on bystanders’ prosocial behavior towards general other third parties (i.e., prosocial conformity), testing the mediating role of moral elevation and the moderating role of economic inequality salience. In Study 2a, we aimed to provide initial evidence for the hypothesis that actors’ social class reduce bystanders’ prosocial behavior towards others and that moral elevation mediates this effect. We recruited adult participants (N = 240) and asked them to read one of two scenarios in which a high-class or low-class actor donated money to victims in an earthquake. We set the donation amount to be the same in either the percentage of donors’ annual income or the absolute amount across three conditions (low-class donor vs. high-class donor with the same percentage of donation vs. high-class donor with the same amount of donation). Then, we measured participants’ moral elevation and prosocial intention towards others. In study 2b, we aimed to further examine our hypothesized mediation model. We recruited adult participants (N = 239) and used the same materials and procedures as those in Study 2a except that we measured participants actual donation of money to charities. In study 2c, we aimed to further examine our hypothesized mediation model by using the recalling and writing task to manipulate actors’ social class (high vs. low). We recruited adult participants (N = 116). After participants recalled and wrote a description of a high-class or low-class actor who donated money to help others, we measured their moral elevation and donation of money to charities. In study 1d, we recruited adult participants (N = 318) and used the same scenarios as those in Study 2a that manipulated actors’ social class (high-class donor with the same percentage of donation vs. low-class donor). In addition, we manipulated economic inequality salience (control vs. salient) to test whether it moderates the effects of actors’ social class on bystanders’ moral elevation and prosocial intention towards others. In Study 1e, we recruited adult participants (N = 320) and used the same scenarios as those in Study 2b that manipulated actors’ social class (high-class donor with the same percentage of donation vs. low-class donor). In addition, we manipulated economic inequality salience (control vs. salient) to further test whether it moderates the effects of actors’ social class on bystanders’ moral elevation and prosocial behavior towards others. Across five studies, we find that bystanders’ moral elevation mediates the effect of actors’ social class on bystanders’ prosocial behavior towards others. That is, compared to a high-class actor, the low-class actor’s donation makes the bystander feel more morally elevated, and thus engage in more prosocial behavior towards others. Across Studies 2d and 2e, we also find that economic inequality salience does not moderate the effects of actors’ social class on bystanders’ moral elevation and prosocial behavior towards others.

From the perspective of generalized transmission of prosocial behavior in social interaction among strangers, the current research examined the effects of actors’ social class on recipients’ paying the prosociality forward, and on bystanders’ prosocial conformity, as well as their respective mediating mechanism (i.e., gratitude and moral elevation) and the moderating role of economic inequality salience in these effects. We synthesized different factors from macro level (i.e., economic inequality), meso level (i.e., prosocial contexts) and micro level (actors’ socia class, recipients’ gratitude, and bystanders’ moral elevation) that affect prosocial behavior, and emphasized the importance to elicit recipients’ and bystanders’ prosocial behavior towards others. We tested our findings in multiple prosocial contexts by using multiple measures of prosocial behavior, including prosocial intention, money allocation, and donation behavior. Across these studies, the data repeatedly supported our hypothesized models, extending the reproducibility and generalizability of our findings. Our findings contribute to explaining and promoting the generalized transmission of prosocial behavior in the current social background of the large gap between the rich and the poor, and also provide important insights into how to engourage civil participation, how to cultivate citizens’ positive prosocial mentality as well as how to improve social psychological services.