Study 2 ------- The protocol for the first study in this Registered Replication Report on the Verbal Overshadowing Effect included a potentially important discrepancy from the original procedure of Study 1 in Schooler & Engstler-Schooler (1990). Specifically, the order of the description task and the crossword puzzle filler task were reversed, meaning that there was a delay between the description phase and the recognition phase in the replication study but not in the original. That ordering mimicked Study 4 of the original paper, which showed a roughly comparable overshadowing effect to the original Study 1. But, some verbal overshadowing studies that included a delay between the description phase and the test phase showed a reduced overshadowing effect. This followup study is designed to verify the size of the verbal overshadowing effect using the original task ordering from Study 1 of Schooler & Engstler-Schooler (1990). It will allow an assessment of the verbal overshadowing effect with the original ordering of Study 1 and will also allow a comparison between studies to determine whether ordering affects the size of the verbal overshadowing effect. The protocol is identical to that of the first RRR study except that the order of the description task and crossword puzzle filler task are reversed. Participants first view the bank robbery video. They then compete the 20-minute crossword puzzle filler task. Next they either describe the bank robber (experimental condition) or name countries and capitals (control condition). Finally, they complete the recognition task. Other than this change in ordering, the protocol for the followup study is identical to that of the first study. Our laboratory made the following changes to our implementation of the design: 1. Sample Size: Our final sample will consist of between 80 (at least 40 in each condition) and 110 participants (55 in each condition). To reach 110, we expect to run between 110 and 120 participants. 2. Stopping Rules: We will test at least 40 participants in each condition. If we finish collecting data from 80 participants and more participants are available from the participant pool, we will continue testing additional participants until April 15 or until we reach 110 participants, whichever comes first. 3. Experimenters: Jessie Pappagianopoulos is studying abroad, so she will not be collecting data this semester. To take her place, I recruited another undergraduate student, Matthew Attaya, to work in the lab. Matthew is a Junior Psychology major. He will be trained to collect, analyze, and summarize data. Results and Discussion A chi-square test of independence was calculated comparing the percentage of correct identification in the experimental and control conditions. A significant relationship was found, χ²(1) = 7.257, p = .007. The target face was correctly selected by 29.5% of participants in the experimental condition and 58.5% of participants in the control condition. A chi-square test of independence was calculated comparing the ratio of the percentage selecting the wrong face (misidentification) to the percentage indicating “not present” (miss) across the experimental and control conditions. No significant relationship was found, χ²(1) = .536, p = .464. In the experimental condition approximately 41.9% of the errors were misidentifications compared to 52.9% of the errors in the control condition. We also examined the confidence ratings of participants using a 2 (Condition: Experimental vs. Control) x 2 (Accuracy: Correct vs. Incorrect/Miss) analysis of variance (ANOVA). There was no main effect of Condition F(1, 81) = .790, p = .377. The main effect of Accuracy was not significant, F(1, 81) = 1.300, p = .258. Finally, the interaction between condition and accuracy was not significant, F(1, 81) = .402, p = .528. Overall, our results support the verbal overshadowing effect (Schooler & Engstler-Schooler, 1990). We found that verbally rehearsing an image impairs later recognition of that image. However, in our previous attempt to replicate this study (i.e., RRR Study 1), we did not find statistically significant support for the verbal overshadowing effect.