Correction: Recollection-Based Retrieval Is Influenced by Contextual Variation at Encoding but Not at Retrieval

The equations in the Experiment 1 section under the subheading “Results and Discussion” are incorrect. The first three paragraphs of that section, containing the corrected equations, are as follows: We present the data corrected for the estimates of recollection and familiarity: We first calculated, for each participant, the proportion of Remember (R) and Know (K) responses by dividing a given response by the total number of non-guess responses, that is, by the total number of responses excluding guess responses. To illustrate, given that each experimental condition contained 20 items, the R proportion was calculated as: PðRÞ 1⁄4 R 20 G where R and G are the raw number of R and G responses. The subtraction of G responses from the total number of trials was performed after post-test debriefing revealed that G responses were mainly due to experimental noise. For example, participants typically responded with G when they pressed the "old" button by mistake, or when they could not decide when an item was old or new. Overall G rates were very small (M = .016, SD = .035 in the encoding-manipulation block;M = .038, SD = .071 in the retrieval-manipulation block), and were equally distributed over the experimental conditions (t(47) = 0.275, p = .785 in the encoding-manipulation block; F(3,141) = 0.828, p = .480 in the retrieval-manipulation block) (cf., [46]). Thus, given that each experimental condition contained 20 items, Proportion Rs and Ks were calculated as: