Mayer J D, Bower G H
J Pers Soc Psychol. 1986 Sep;51(3):473-92. doi: 10.1037//0022-3514.51.3.473.
Although personality traits are commonly assumed to be represented in memory as schemata, little research has addressed whether such schemata can be learned from observation. Subjects in three studies classified 60 person instances into group members and nonmembers as defined by the instances' match to a complex personality prototype. To simulate learning of fuzzy categories, each person instance provided conflicting cues to group membership. Learning for instances' group membership was excellent across studies. In Study 1, frequency of cues indicating group membership was greatly overestimated among nongroup instances. In Study 2, schema-consistent memory bias was revealed for person instances. In Study 3, schemata of consistently positive (or negative) traits were learned faster than arbitrary schemata. The findings implicated frequency sensitivity of memory (Estes, 1986), and a model of probabilistic cued-memory retrieval was developed to account for the effects. The findings were then discussed in relation to everyday cognitive performance.
尽管人们通常认为人格特质在记忆中是以图式的形式呈现的，但很少有研究探讨这种图式是否可以通过观察来学习。在三项研究中，受试者根据与复杂人格原型的匹配程度，将60个个体实例分为组成员和非组成员。为了模拟对模糊类别的学习，每个个体实例都提供了相互矛盾的组成员线索。在各项研究中，对实例组成员身份的学习效果都很好。在研究1中，非组成员实例中表明组成员身份的线索频率被大大高估。在研究2中，发现了个体实例的图式一致记忆偏差。在研究3中，始终积极（或消极）特质的图式比任意图式学得更快。这些发现暗示了记忆的频率敏感性（埃斯蒂斯，1986），并开发了一个概率线索记忆检索模型来解释这些效应。然后结合日常认知表现对这些发现进行了讨论。
