Recognition Confidence Under Violated and Confirmed Memory Expectations
Recognition Confidence Under Violated and Confirmed Memory Expectations
复制标题
DOI：
10.1037/a0025687
复制
发表时间：
2012-05-01
影响因子：
4.1
通讯作者：
Dobbins, Ian G.
中科院分区：
文献类型：
--
作者：
Jaeger, Antonio;Cox, Justin C.;Dobbins, Ian G.
个人的记忆体验通常与周围其他人的记忆体验存在差异，平均而言，如果同伴推荐某个项目，那么该项目更有可能是熟悉的。如果观察员能够在作出承认决定时使用其他人的外部建议作为统计先验，这将是理想的，但目前还不清楚他们如何或是否这样做。此外，理解识别判断对这些外部线索的敏感性对于理解记忆一致性和目击者可辨认性现象至关重要。为了解决这个问题，我们检查了识别准确性和信心以下线索从外部来源（例如，“可能是旧的”），其预测即将到来的内存探测的可能状态。三个人出现了。首先，当参与者被无效提示和有效提示时，命中率和正确拒绝率都有所下降。第二，命中的信心一般高于正确拒绝的信心，无论线索的有效性。最后，也是最值得注意的，线索有效性与判断信心相互作用，有效性严重影响了正确拒绝的信心，但对命中的信心没有明显的影响。引导通知蒙特卡洛模拟支持一个双过程识别模型下，熟悉和回忆过程抵消严重抑制外部线索的影响，平均报告的信心。第三个实验使用源记忆测试了这个模型。正如预测的那样，由于源记忆在很大程度上受上下文回忆的影响，线索有效性再次不影响信心，尽管与识别一样，它明显改变了准确性。
Individuals' memory experiences typically covary with those of others' around them, and on average, an item is more likely to be familiar if a companion recommends it as such. Although it would be ideal if observers could use the external recommendations of others' as statistical priors during recognition decisions, it is currently unclear how or if they do so. Furthermore, understanding the sensitivity of recognition judgments to such external cues is critical for understanding memory conformity and eyewitness suggestibility phenomena. To address this we examined recognition accuracy and confidence following cues from an external source (e.g., "Likely Old") that forecast the likely status of upcoming memory probes. Three regularities emerged. First, hit and correct-rejection rates expectedly fell when participants were invalidly versus validly cued. Second, hit confidence was generally higher than correct-rejection confidence, regardless of cue validity. Finally, and most noteworthy, cue validity interacted with judgment confidence such that validity heavily influenced the confidence of correct rejections but had no discernible influence on the confidence of hits. Bootstrap-informed Monte Carlo simulation supported a dual process recognition model under which familiarity and recollection processes counteract to heavily dampen the influence of external cues on average reported confidence. A 3rd experiment tested this model using source memory. As predicted, because source memory is heavily governed by contextual recollection, cue validity again did not affect confidence, although as with recognition it clearly altered accuracy.