基于格式塔流派的观点，Campbel(1958)提出群体实体性的概念，并将其定义为一个群体有意义或能称之为群体的程度。实体性是群体的基本特征。它是一个连续体，不同的群体具有不同水平的实体性。以往关于如何知觉实体性群体的研究主要存在以下问题。首先，在不同群际关系背景下，实体性对群际态度的作用是矛盾的。在群际合作背景下，外群体的实体性越高，受到的喜爱越多，而在群际冲突背景下，外群体的实体性越高，遭受的歧视和怀疑越多。其次，在群体印象形成过程中，实体性对群体知觉的影响不一致。有些研究发现，目标群体的实体性越高，知觉者对其评价越极端。但也有研究发现，目标群体的实体性越高，知觉者对其评价越积极。第三，实体性的不同知觉线索(相似性和互动性)在群体知觉中的作用尚不明确。由此可见，实体性对群体知觉的影响是有待解决的科学问题。 这一科学问题可以进一步分解为三个子问题。第一，在群际知觉过程中，即考虑本群体与外群体之间群际关系时，外群体的实体性如何影响本群体对外群体的知觉？第二，在群体印象形成过程中，即排除了群际关系后，目标群体的实体性如何影响个体对目标群体的知觉？第三，实体性不同线索对群体知觉的效应是否相同？ 本研究基于一个理论支点三条路径来解决这些问题。就理论支点而言，本研究通过引入“大二”模型(Fiske et al., 2002)，区分群体知觉的两个基本维度，来探讨实体性对群体热情知觉和能力知觉的不同作用。根据“大二”模型，对群体的热情知觉取决于对群体意图的判断，而对群体的能力知觉取决于对群体实现其意图的实力的评价。已有研究表明，高实体性群体同时具有强烈的行为意图以及实现群体意图的能力。因此，实体性同时作用于对群体的热情知觉和能力知觉。基于这个理论支点，本研究采取了以下三条路径，每条路径中均蕴含着研究假设的推衍过程。 首先，本研究从群际知觉视角，探讨在本群体和外群体处于合作或冲突的群际关系背景下，外群体的实体性如何影响本群体对其的热情知觉和能力知觉。当本群体与外群体处于合作关系时，外群体的实体性越高，其积极意图越强烈且实现积极意图的实力越强，因此，本群体对其热情知觉越高，能力知觉越高；当本群体与外群体处于冲突关系时，外群体的实体性越高，其消极意图越强烈且实现消极意图的实力越强，因此，本群体对其热情知觉越低，能力知觉越高。基于上述推断，我们假设：在群际知觉过程中，实体性对外群体热情知觉具有极化效应(假设1)，而对其能力知觉则具有正向效应(假设2)。 其次，为考察在排除合作或冲突的群际关系背景后，上述群际情境中的效应是否具有普适性，本研究基于群体印象形成的视角，探讨在没有群际关系的情境中，目标群体的实体性如何影响个体对其进行热情知觉和能力知觉。当目标群体持有积极意图时，实体性水平越高，积极意图越强烈，个体对其热情评价越高，然而，当目标群体持有消极意图时，实体性水平越高，消极意图越强烈，个体对其热情评价越低。不论目标群体初始能力如何，其实体性越高，个体对其能力评价越高。基于上述推论，我们假设：个体对目标群体的印象形成过程中，实体性对目标群体热情知觉具有极化效应(假设3)，而对其能力知觉则具有正向效应(假设4)。 第三，本研究从群体特征本身入手，区分实体性的两种线索并比较相似性和互动性对群体热情知觉和能力知觉的影响。相似性是指群体成员具有共同的属性和特征，互动性指群体成员拥有共同的目标并相互依赖。相似性会使群体特征泛化到每个成员，知觉者会对高相似性群体进行刻板化和特质化知觉，这使得对群体的评价更加极端。因此，我们假设：基于相似性的实体性对热情知觉和能力知觉均具有极化效应(假设5)。高互动性群体具有更强的意图，并且群体内部的合作使得群体更有可能实现其意图。因此，我们假设：基于互动性的实体性对群体热情知觉具有极化效应，而对其能力知觉具有正向效应(假设6)。 为了解决上述科学问题，并验证6个假设，本文基于上述三条路径，共进行了4个研究，9个子研究。紧密围绕实体性对群体知觉的影响这一主题，本文采取群际知觉视角与群体印象形成视角相结合、整体实体性与实体性线索相结合、实证研究与元分析研究相结合的方式，环环相扣，层层递进。 研究1从群际知觉视角探讨了实体性如何影响本群体对外群体的热情知觉和能力知觉。研究包括3个子研究。研究1.1在实验室情境中设置具有不同实体性水平的外群体，并操纵其与被试所在群体之间的群际关系。研究1.2和1.3选取了汉族人作为被试，测量了真实外群体(维族人)的实体性，并测量了汉维关系。研究发现，在群际合作背景下，实体性同时促进对外群体的热情知觉和能力知觉；在群际冲突背景下，实体性降低对外群体的热情知觉，促进对外群体的能力知觉。3个子研究的结果一致地验证了假设1和2，即在群际知觉过程中，实体性对热情知觉具有极化效应，而对能力知觉具有正向效应。 研究2从群体印象形成视角探讨了实体性如何影响个体对目标群体的群体知觉。研究包括3个子研究。研究2.1和2.2创设了虚拟目标群体，并通过行为描述设置了群体的效价，考察了实体性对群体热情知觉(研究2.1)和能力知觉(研究2.2)的作用。研究2.3选用了四种不同类型的真实社会群体，并用图片呈现的方式操纵其实体性，考察了实体性对群体热情知觉和能力知觉的影响。3个子研究的结果一致地验证了假设3和4，即在群体印象形成过程中，实体性对热情知觉具有极化效应，而对能力知觉具有正向效应。 研究3分解了实体性的不同线索，进一步探讨了相似性和互动性对群体知觉的影响。研究包括2个子研究。研究3.1基于群际知觉视角，选用美国被试，以真实群体(“一带一路”联盟)作为外群体，并操纵其相似性和互动性；同时，测量美国与“一带一路”联盟之间的群际关系作为热情知觉的初始水平，并通过操纵“一带一路”联盟的经济地位设置能力知觉的初始水平。研究3.2基于群体印象形成视角，选用四个虚拟目标群体，并操纵其相似性或互动性；同时，通过呈现行为条目的方法设置热情知觉和能力知觉的初始水平。2个子研究的结果一致验证了假设5和6，即基于相似性的实体性对热情知觉和能力知觉均具有极化效应；基于互动性的实体性对热情知觉具有极化效应，而对能力知觉具有正向效应。 研究4采用元分析的方法，将上述3个实证研究的结果进行定量合并，估计实体性及其线索对群体知觉影响的效应大小。研究对实体性的操作性定义(实体性、相似性或互动性)、群体知觉的维度(热情或能力)以及群体在知觉维度上的效价（积极或消极）进行编码，将每个效应经过样本量的权重处理，基于随机模型计算总体和单独效应的平均估计值。结果不仅验证了6个研究假设，揭示了实体性及其线索对热情知觉和能力知觉具有中等大小的效应，还表明互动性比相似性更接近实体性的本质。 本研究首次将热情和能力这两个群体知觉的基本维度引入到对实体性的研究。通过4个研究验证了6个研究假设，解决了实体性对群体知觉的影响这一科学问题，回答了在群际知觉视角以及群体印象形成视角下，实体性及其线索如何影响群体热情知觉和能力知觉这些问题，具有重要的理论贡献和现实意义。 本研究的理论贡献表现在以下三个方面。首先，本研究揭示了实体性对于群体知觉的影响同时体现在热情和能力两个维度上，这充分解释了以往研究的不一致结果，同时加深了对实体性本质的理解。其次，本研究拓展了社会认知的“大二”模型，揭示了群体知觉不仅基于社会结构因素，还受到实体性这一群体自身特征的影响。第三，本研究通过区分相似性和互动性对群体知觉的影响，反映了不同群体形成方式对群体知觉的作用，为解决社会分类理论和群体互动理论关于群体本质的争议提供了新的思路。 本研究为改善群际关系和提升群体形象提供了科学依据。一方面，污名化的群体(例如，穆斯林、朝鲜)可以通过弱化自身实体性的方式提升人们对其的热情知觉，进而改善群际关系。另一方面，联盟性群体(例如，“一带一路”联盟、欧盟)可以通过彰显自身实体性的方式来增强人们对其的能力知觉, 进而提升群体形象。

Groups vary along an entitativity continuum, that is, the extent to which an aggregate has real existence and is meaningful (Campbell, 1958). Entitativity is seen as the fundamental group characteristic. Three unsolved questions arise from existing research about social perceptions of entitative groups. First, groups with high entitativity are treated with divergent responses. Some highly entitative groups are associated with more suspicion and discrimination, while the others are be-loved. Second, the effect of entitativity on group judgments is inconsistent. Some studies suggest that entitativity is associated with more extreme judgments, while some researchers have found a positivity effect of entitativity on group judgments. Third, whether the two cues of entitativity exert the same effect on group judgments remains unknown. To answer these questions, we must examine the effect of entitativity on group judgments systematically. Present research proposes a theoretical resolution and three empirical strategies to examine how entitativity influences group judgments. Theoretically, inspired by Fiske et al.'s seminal work suggesting that groups are mainly judged along the two fundamental dimensions of warmth and competence (Fiske, Cuddy, Glick, & Xu, 2002), the present research aimed to simultaneously examine the effects of entitativity on warmth and competence judgments. High entitativity magnifies the strength of group-intention attribution (or intentionality) and the capability to enact the intent. From intergroup perception approach, we first explored how people perceive high and low entitativity outgroups in terms of warmth and competence under cooperative or conflictive relations. Based on findings from previous research, we hypothesized that, compared with low entitativity outgroups, high-entitativity outgroups would be viewed as warmer under cooperative relations and colder under conflictive relations (Hypothesis 1), whereas they would be perceived as more competent regardless of the intergroup relation (Hypothesis 2). To exclude the influence of intergroup relation, from group impression formation approach, we examined how entitativity influences warmth and competence judgments about target groups. We hypothesized that, when a target group had positive intent, the more entitative it was, the warmer it would be rated as, while when it had negative intents, the more entitative it was, the colder it would be rated as. Stated differently, when making impressions of groups, entitativity would have a polarization effect on warmth judgments (Hypothesis 3). Meanwhile, high entitativity target groups would be viewed as more competent regardless of the initial levels of capability of the groups, that is to say, entitativity would exert a positivity effect on competence judgments (Hypothesis 4). Furthermore, the present research differentiates the effects of similarity- and interaction-based entitativity on warmth and competence judgments, as similarity and interaction has been identified as two main perceptual cues of entitativity. For groups with high similarity-based entitativity, attribution of traits to members are made based on mere group membership (i.e., stereotyping) and people tend to make dispositional attributions. Therefore, we hypothesize that similarity-based entitativity polarizes warmth and competence judgments (Hypothesis 5). Groups with high interaction-based entitativity are seen as more deliberate and intentional and they are more capable to enact agenda and achieve goals. Therefore, we hypothesized that interaction-based entitativity facilitates polarized warmth judgments and enhanced competence judgments (Hypothesis 6). Four studies were conducted to test these hypotheses. Study 1 examined the effects of entitativity on outgroup judgments from intergroup perception approach. In Study 1.1, we experimentally created the outgroup (i.e., Group X) and manipulated entitativity and functional relations. In Studies 1.2 and 1.3, we chose a real outgroup (i.e., Uyghurs) and measured indicators of entitativity (Study 1.2) or entitativity itself (Study 1.3) and interethnic relations. In all studies, participants rated the outgroup on warmth and competence dimensions. The results suggested that, under cooperative functional relation, the outgroup with higher entitativity was perceived as more competent and warmer, thereby more beneficial. Conversely, when the functional relation was conflictive, the outgroup with higher entitativity was perceived as more competent but colder, and thus more harmful. Study 2 examined the effects of entitativity on group judgments from group impression formation approach. In Studies 2.1 and 2.2, the target group was a novel group; its entitativity was manipulated by an instruction set. We described the target group as positive/negative on a warmth (Study 2.1) or competence dimension (Study 2.2) with segments of behavior statements. In Study 2.3, we selected four types of real social groups as target groups based on the stereotypes that people had held toward them. The entitativity of these four groups was manipulated by pictures. The results revealed that entitativity exerted a polarization effect on warmth and a positivity effect on competence judgments. Study 3 differentiated the effects of two cues of entitativity (i.e., similarity and interaction) on warmth and competence judgments about groups. In Study 3.1, from intergroup perception approach, participants were Americans and the outgroup was the One Belt One Road union. The similarity or interaction (high or low) of the participating countries of OBOR union was manipulated by descriptions. In Study 3.2, from group impression formation approach, either similarity or interaction (high or low) of novel groups was manipulated by instrument sets. In both Studies 3.1 and 3.2, participants were asked to rate targets groups on traits measuring warmth and competence. Consistent with our theoretical perspective, similarity-based entitativity facilitated stereotyped judgments, whereas, interaction-based entitativity polarized intent-related judgments and enhanced capability-related judgments. To quantitatively synthesize the effect of entitativity on group judgments, Study 4 was a meta-analysis of the results obtained in Studies 1, 2 and 3. In this study, the operationalization of entitativity (entitativity, similarity, or interaction), the dimension of group judgments (warmth or competence) and the valence of group were coded. Results of random-effect models support all the hypotheses we proposed. Specifically, entitativity exerted a polarization effect on warmth and a positivity effect on competence judgments. Similarity polarized warmth and competence judgments. Interaction exerted a polarization effect on warmth judgments and a positivity effect on competence judgments. The present research integrates the two fundamental dimensions of social perception into entitativity research and differentiates the effects of two cues of entitativity on group judgments. It offers significant theoretical contributions. First, our findings can explain the inconsistent results of previous research and reveal the ambivalent nature of entitativity. Second, our findings also advance the Big-Two model by confirming that, besides social structural factors, characteristics of group per se (i.e., entitativity) can also influence the content of group judgments. Third, with the combination of two perspectives about group formation, that is social categorization theory and group dynamic theory, by differentiating effects of two cues of entitativity on group judgments, we support the proposal that group judgments depend on the way of group formation. The present research suggests some practical strategies to improve intergroup relationships and enhance group image. On one hand, for nations or ethnicities that have been seen as with ill will (e.g., the Jews, North Korea, the Muslims), an effective measure aimed at reducing entitativity, through facilitating intergroup contact for example, can be taken to improve people's attitudes toward such groups. On the other hand, measures (e.g., synchronized behavior) can also be taken to elevate group entitativity if groups in need of growth (e.g., the OBOR union or European Union) are motivated to portray themselves as more prosperous and stronger.