Research on intergroup bias usually focuses on a single dimension of social categorization. In real life, however, people are aware of others' multiple group memberships and use this information to form attitudes about them. The present research tests the predictive power of identification, perceived conflict, and perceived symbolic threat in explaining the strength of intergroup bias on various dimensions of social categorization in multiple categorization settings. We conduct a factorial survey experiment, manipulating 9 dimensions of social categorization in diverse samples from 4 countries (n = 12,810 observations, 1,281 participants representing 103 social groups). The dimensions studied are age, gender, ethnicity, religion, place of residence, education, occupation, income, and 1 country-specific dimension. This approach allows exploring the generalizability of established determinants of bias across dimensions of categorization, contexts, and target groups. Identification and symbolic threat showed good generalizability across countries and categorization dimensions, but their effects varied as a function of participant and target groups' status. Identification predicted stronger bias mainly when the participant belonged to a higher status and the target belonged to a lower status group. Symbolic threat predicted stronger bias mainly when the target was a minority group member. Conflict predicted bias only in few cases, and not only the strength but also the direction of the effects varied across countries, dimensions, and target and participant groups. These findings help to clarify the limits of generalizability of established determinants of intergroup bias and highlight the need for new explanations of social-cognitive processes among minority group members. (PsycInfo Database Record (c) 2020 APA, all rights reserved).

中文翻译：

---

多重分类和群际偏见：检验三种群际关系理论的普遍性。

对群体间偏见的研究通常集中在社会分类的单一维度上。然而，在现实生活中，人们知道其他人的多个组成员身份，并使用这些信息来形成对他们的态度。本研究测试了识别、感知冲突和感知符号威胁的预测能力，以解释在多个分类设置中社会分类的各个维度上群体间偏见的强度。我们进行了一项因子调查实验，在来自 4 个国家的不同样本中处理了 9 个社会分类维度（n = 12,810 次观察，代表 103 个社会群体的 1,281 名参与者）。研究的维度是年龄、性别、种族、宗教、居住地、教育、职业、收入和 1 个国家特定维度。这种方法允许探索跨分类、上下文和目标群体维度的既定偏见决定因素的普遍性。识别和象征性威胁在国家和分类维度上表现出良好的普遍性，但它们的影响因参与者和目标群体的地位而异。当参与者属于较高的地位而目标属于较低地位的群体时，识别预测了更强的偏见。象征性威胁主要在目标是少数群体成员时预测更强的偏见。冲突仅在少数情况下预测偏见，而且影响的强度和方向因国家、维度、目标和参与者群体而异。这些发现有助于阐明确定的群体间偏见决定因素的普遍性限制，并强调需要对少数群体成员的社会认知过程进行新的解释。（PsycInfo 数据库记录 (c) 2020 APA，保留所有权利）。