Lawlor M, Cochran L
Br J Med Psychol. 1981 Mar;54(Pt 1):41-50. doi: 10.1111/j.2044-8341.1981.tb01468.x.
This study tested Bannister's (1963) serial invalidation hypothesis that repeated invalidation of a person's judgements loosens conceptual organization. That is, a person's constructs or dimensions of judgement will become more weakly and inconsistently related. Using a variant of Kelly's (1955) repertory grid methodology, thirty university students rated three sets of eight career roles on sixteen constructs. During the ratings of the second set, subjects received both evaluative commentary from the experimenter and objective evidence in the form of fake ratings. Both types of feedback were either invalidating or validating of subjects' judgements on the first eight constructs. The main finding was that invalidation loosened conceptual organization, but only for subjects who initially manifested strong relations among constructs. Subjects who began with weaker construct relations were not significantly affected by type of feedback.
本研究检验了班尼斯特（1963）的系列无效化假设，即对一个人的判断反复进行无效化会削弱概念组织。也就是说，一个人的判断结构或维度之间的关联会变得更弱且更不一致。采用凯利（1955）的 repertory grid 方法的一个变体，30 名大学生根据 16 个结构对三组共八个职业角色进行评分。在对第二组进行评分时，受试者既收到了实验者的评价性评论，也收到了以虚假评分形式呈现的客观证据。这两种反馈对受试者在前八个结构上的判断要么是无效化的，要么是确认有效的。主要发现是，无效化削弱了概念组织，但仅针对那些最初在结构之间表现出强关联的受试者。最初结构关联较弱的受试者没有受到反馈类型的显著影响。
