1 Papers 1–7 Friday Morning Scene Processing Grand Ballroom, Friday Morning, 8:00–9:35 Chaired by Carrick C. Williams, Mississippi State University 8:00–8:15 (1) Encoding and Visual Memory: Is Task Always Irrelevant? CAR- RICK C. WILLIAMS, Mississippi State University—Although some aspects of encoding (e.g., presentation time) appear to have an effect on visual memories, viewing task (incidental or intentional encoding) does not. The present study investigated whether different encoding manipulations would impact visual memories equally for all objects in a conjunction search (e.g., targets, color distractors, object category distractors, or distractors unrelated to the target). Participants were pre- sented sequences of 12 real-world pictures for 637 msec each and were asked to search for prespecified targets (e.g., green apple), memorize all objects, search for specified targets while memorizing all objects, search for postidentified targets (e.g., How many green apples were there?), or memorize all objects with one object prespecified. Encoding task significantly affected visual memory, but only for targets and unrelated distractors, indicating differences in the encoding processes of visual memories. Finally, confidence ratings indicated that participants were able to accurately judge the veracity of their visual memories. 8:20–8:35 (2) When Does Memory Facilitate Perception (of a Scene’s Layout)? After One or Two Episodes. THOMAS SANOCKI & NOAH SUL- MAN, University of South Florida—Four experiments indicate that prior memory of a scene is necessary for a top-down spatial priming effect. We measured spatial processing within scenes that were immediately preceded by either a (same) scene prime or a control prime. The scene was either new or repeated on a trial. When new, scenes primes did not cause more accurate processing than did the control prime. However, after one or two repetitions, scene primes increased accuracy (sensitiv- ity) of spatial processing of the briefly presented scene target, relative to the control prime. Thus, facilitation of scene layout processing was not immediate (first primed trial) but required memory for at least one prior episode with the scene. In a fifth experiment, reaction time methods suggested that scene primes can cause a bias effect on the first trial with a new scene. 8:40–8:55 (3) The Influence of Scene Context on Parafoveal Processing of Objects. MONICA S. CASTELHANO, Queen’s University—Does scene context influence object recognition before the object is directly fixated? We examined this question using a modified boundary paradigm (Rayner, 1975). The participants’ task was to indicate whether a target object matched an object name presented earlier. Critically, objects were pre- sented on either a scene or a gray background. On each trial, a cue ap- peared, and, once fixated, an object preview would onset 4º or 10º away. The preview object could be identical to the target, of the same category (but with the same or a different shape), of a different category (but with the same or a different shape), or a control (rectangle). During the saccade toward it, the target object replaced the preview. The results revealed that, although there was no effect of preview category, same- shape previews presented at 4º produced a greater benefit on a scene than on a gray background. Two possible influences of scene context on parafoveal preview benefit will be discussed. 9:00–9:15 (4) The Role of Semantic Memory in Learning Contextual Regularities in Real-World Scenes. JAMES R. BROCKMOLE, University of Notre Dame, & MELISSA L.-H. VO, University of Edinburgh—When encoun- tering familiar scenes, observers can use episodic memory to facilitate the guidance of attention to objects appearing in known locations or configurations. We investigated whether and how memory for semantic contingencies that exist across different scenes is used to guide attention. Observers searched for letter targets embedded in different bedrooms. In a between-subjects manipulation, targets were always on bed pillows or randomly positioned. When targets were systematically located within scenes, search for them became progressively more efficient. Learning was abstracted away from bedrooms and transferred to a living room, where the target was on a sofa pillow. These contingencies were explicit and led to central tendency biases in memory for precise target positions. These results broaden the scope of conditions under which contextual cuing operates and demonstrate for the first time that semantic memory plays a causal and independent role in the learning of associations be- tween objects in real-world scenes. 9:20–9:35 (5) Visual Memory: Confidence, Accuracy, and Recollection of Specific Details. GEOFFREY R. LOFTUS, University of Washington, MARK T. REINITZ, University of Puget Sound, WILLIAM PERIA, University of Washington, & JULIE SEGUIN, University of Victoria, Wellington— How does the confidence/accuracy relationship differ when picture rec- ognition is based on general familiarity versus specific features? Pictures were presented for varying exposure durations, followed by an old–new recognition test. Observers provided a confidence rating for each old– new response and also indicated whether each response was based on the picture’s general familiarity or on one or more specific features in the picture. Feature-based responses produced higher confidence and higher accuracy. However, holding confidence constant, a feature-based response was less accurate than a familiarity-based response. The sci- entific conclusion is that confidence and accuracy are not based on the same internal events. The practical conclusion is that, contrary to com- mon sense and to normal behavior (see Bell & Loftus, 1988), an eyewit- ness’s accuracy in a legal setting should be discounted more when his or her recognition responses are based on memories for specific features than when they are based on general familiarity. Selective Attention I Constitution Ballroom, Friday Morning, 8:00–9:35 Chaired by Jeremy M. Wolfe Brigham and Women’s Hospital and Harvard Medical School 8:00–8:15 (6) Two Dissociable Decision Criteria in Visual Search Revealed by Varying Target Prevalence. JEREMY M. WOLFE, Brigham and Women’s Hospital and Harvard Medical School, & MICHAEL J. VAN WERT, Boston University—The frequency of targets in visual search (target prevalence) shapes search behavior. When targets are rare (1%–2% prevalence), observers use conservative response criteria, producing high miss rates. This might be just a version of a speed– accuracy trade-off, since low prevalence yields fast absent responses. We disprove this hypothesis by showing that very high target preva- lence (98%) shifts response criteria in the opposite direction, lead- ing to elevated false alarms, without leading to fast target-present re- sponses. Rather, the rare target-absent responses are greatly slowed. In a second experiment, prevalence was varied sinusoidally over 1,000 trials. Observers’ criterion and target-absent RTs tracked prevalence, whereas sensitivity (d ′ ) and target-present RTs did not. The results support a model with two criteria, both influenced by prevalence. One criterion governs perceptual decisions about each attended item. The other influences a quitting threshold that modulates RTs for target- absent responses. 8:20–8:35 (7) Reward-Induced Attentional Amnesia in Visual Search. BRIAN R. LEVINTHAL, Northwestern University, & ALEJANDRO LLERAS, University of Illinois, Urbana-Champaign (read by Alejandro Lleras)— Intertrial effects in visual search have been demonstrated in a wide va- riety of paradigms and are known to be remarkably robust. The distrac- tor previewing effect (DPE) is a particularly robust intertrial effect that emerges during oddball feature searches and reflects purely inhibitory attentional biases against one specific visual feature (or category). Pre- viously, we have argued that these inhibitory biases reflect an implicit assessment of the usefulness of the information encountered in recent Abstracts of the Psychonomic Society — Volume 14 — November 2009 50th Annual Meeting — November 19–22, 2009 — Boston, Massachusetts