Castela Marta, Erdfelder Edgar
Department of Psychology, School of Social Sciences, University of Mannheim.
J Exp Psychol Learn Mem Cogn. 2017 Feb;43(2):205-225. doi: 10.1037/xlm0000299. Epub 2016 Sep 15.
The recognition heuristic (RH) theory predicts that, in comparative judgment tasks, if one object is recognized and the other is not, the recognized one is chosen. The memory-state heuristic (MSH) extends the RH by assuming that choices are not affected by recognition judgments per se, but by the memory states underlying these judgments (i.e., recognition certainty, uncertainty, or rejection certainty). Specifically, the larger the discrepancy between memory states, the larger the probability of choosing the object in the higher state. The typical RH paradigm does not allow estimation of the underlying memory states because it is unknown whether the objects were previously experienced or not. Therefore, we extended the paradigm by repeating the recognition task twice. In line with high threshold models of recognition, we assumed that inconsistent recognition judgments result from uncertainty whereas consistent judgments most likely result from memory certainty. In Experiment 1, we fitted 2 nested multinomial models to the data: an MSH model that formalizes the relation between memory states and binary choices explicitly and an approximate model that ignores the (unlikely) possibility of consistent guesses. Both models provided converging results. As predicted, reliance on recognition increased with the discrepancy in the underlying memory states. In Experiment 2, we replicated these results and found support for choice consistency predictions of the MSH. Additionally, recognition and choice latencies were in agreement with the MSH in both experiments. Finally, we validated critical parameters of our MSH model through a cross-validation method and a third experiment. (PsycINFO Database Record
再认启发式（RH）理论预测，在比较判断任务中，如果一个对象被再认而另一个未被再认，则会选择被再认的对象。记忆状态启发式（MSH）扩展了RH，它假设选择并非受再认判断本身的影响，而是受这些判断背后的记忆状态（即再认确定性、不确定性或拒绝确定性）的影响。具体而言，记忆状态之间的差异越大，选择处于较高状态的对象的概率就越大。典型的RH范式无法估计潜在的记忆状态，因为不知道这些对象之前是否被经历过。因此，我们通过将再认任务重复两次来扩展该范式。根据再认的高阈值模型，我们假设不一致的再认判断源于不确定性，而一致的判断很可能源于记忆确定性。在实验1中，我们将2个嵌套的多项式模型拟合到数据中：一个明确形式化记忆状态与二元选择之间关系的MSH模型，以及一个忽略一致猜测（不太可能）可能性的近似模型。两个模型都得出了趋同的结果。正如预测的那样，对再认的依赖随着潜在记忆状态的差异而增加。在实验2中，我们重复了这些结果，并发现了对MSH的选择一致性预测的支持。此外，在两个实验中，再认和选择潜伏期都与MSH一致。最后，我们通过交叉验证方法和第三个实验验证了我们的MSH模型的关键参数。（《心理学文摘数据库记录》
