Dyck R J, Rule B G
J Pers Soc Psychol. 1978 May;36(5):521-9. doi: 10.1037//0022-3514.36.5.521.
In two studies, men were defeated on either 17%, 50%, or 83% of reaction time trials, received aversive noise, and could ostensibly retaliate by delivering shock to their partner. The noise level delivered was described in Experiment 1 as typical of most other people (high consensus) or atypical of most other people (low consensus) and in Experiment 2 as from a partner who knew (high foreseeability) or did not know (low foreseeability) the kind and level of stimulation controlled by the switches delivering reinforcement to the recipient. Hypotheses were based on the notion that retaliation increases as more personal causality is attributed to a provoker and that more personal causality is inferred in highly foreseeable--or low consensus--50% defeat conditions. As expected, greater differences in aggression between high and low consensus and between high and low foreseeability were displayed in the 50% defeat condition than in the other defeat conditions. Anticipated differences in inferences were obtained.
在两项研究中，男性在17%、50%或83%的反应时间试验中失败，会听到厌恶的噪音，并且表面上可以通过电击其伴侣进行报复。在实验1中，所发出的噪音水平被描述为大多数其他人的典型水平（高一致性）或非大多数其他人的典型水平（低一致性），在实验2中，噪音被描述为来自知道（高可预测性）或不知道（低可预测性）为接受者提供强化的开关所控制的刺激类型和水平的伴侣。假设基于这样的观念，即报复会随着更多个人因果关系归因于挑衅者而增加，并且在高度可预测——或低一致性——50%失败的情况下会推断出更多个人因果关系。正如预期的那样，与其他失败条件相比，在50%失败条件下，高一致性与低一致性之间以及高可预测性与低可预测性之间在攻击性上的差异更大。获得了预期的推断差异。
