Endel Tulving (1985) drew a distinction between Remembering and Knowing, spurring a great deal of research on the memorial experiences of recollection and familiarity and their contribution to various phenomena in memory. More recently, studies have used this distinction to situate our understanding of the processes that contribute to the testing effect-or, the benefit of retrieval practice to later memory (see also Tulving, 1967). Using retention intervals of approximately 15 min or less between initial and final testing, several studies have found that initial testing magnifies estimates of recollection but not familiarity, regardless of whether a testing effect is revealed in overall recognition performance (Chan and McDermott, 2007). However, the efficacy of prior testing in enhancing memory has been shown to change over time, as have estimates of recollection and familiarity. Thus, the mechanisms that underlie the quintessential testing effect-one that occurs in overall recognition or recall over longer delays-are still uncertain. To investigate this issue, in two experiments, subjects studied word lists, took 3-letter stem cued-recall tests on half of the studied words, and completed a final recognition test in which estimates of recollection and familiarity were obtained via confidence (Experiment 1) or Remember-Know-New (Experiment 2) judgments. Critically, final recognition tests occurred either immediately, 1 day (Experiment 1 only), or 4 days after initial learning. At all retention intervals and in both methods of estimating recollection and familiarity on the final test (i.e. receiver-operating characteristic and remember-know analyses), initial testing magnified estimates of both recollection and familiarity. These findings suggest that the testing effect can result from changes in both processes and pose issues for theories of the testing effect that consider an exclusive role for recollection.

中文翻译：

---

熟悉度在支持测试效果方面的作用。

恩德尔·图尔文（Endel Tulving，1985）在“记住”和“知道”之间作了区分，促使人们对回忆和熟悉的纪念性经历及其对记忆中各种现象的贡献进行了大量研究。最近，研究已经利用这种区别使我们对有助于测试效果的过程的理解，或者是取回实践对以后的记忆的好处（另见Tulving，1967）。使用初始测试和最终测试之间大约15分钟或更短的保留间隔，几项研究发现，初始测试会放大回忆的估计值，而不是熟悉程度，而不管测试效果是否在总体识别性能中有所体现（Chan和McDermott，2007）。但是，事实表明，先前测试在增强记忆力方面的功效会随着时间而改变，以及对回忆和熟悉程度的估计。因此，仍然无法确定作为典型测试效果基础的机制（在总体识别或长时间记忆中出现的一种机制）。为了研究这个问题，在两个实验中，受试者研究了单词列表，对一半的单词进行了3个字母的词干提示回忆测试，并完成了最终的识别测试，其中通过置信度获得了对回忆和熟悉程度的估计（实验1 ）或“记住新知识”（实验2）的判断。至关重要的是，最终识别测试是在初次学习后立即，1天（仅限实验1）或4天进行的。在所有保留时间间隔上，以及在两种评估最终测试的记忆力和熟悉度的方法中（例如，接收者的操作特征和记住的分析），初始测试会放大回忆和熟悉程度的估计。这些发现表明，测试效果可能来自两个过程的变化，并且给测试效果的理论带来了问题，这些理论考虑了回收的唯一作用。