Discriminating Targets among Distractors in a Virtual Shopping Environment with Different Rack Orientations: Testing a Model of Visibility
Scott Watamaniuk, Ph.D. (Advisor); Robert Gilkey, Ph.D. (Committee Member); John Flach, Ph.D. (Committee Member); Pratikh Parikh, Ph.D. (Committee Member)
Master of Science (MS)
Objective: This study measured observers’ abilities to identify letter targets distributed among number distractors in a virtual shopping environment. Head-turning behavior was also continuously recorded throughout each trial. The data were then used to test whether a model’s prediction for the duration of visibility needed for target detection in a virtual shopping environment (Parikh & Mowrey, 2014) generalize to the more realistic shopping task of identifying a target on a shelf. Currently, the model predicts the visibility of the locations of targets in traditional racks oriented 90° to the aisle (perpendicular) as well as racks oriented at 30°, 45, 135°, and 150° to the central aisle. Background: Exposure (whether a portion of the rack is seen) and intensity (how long that rack portion is seen) are the two variables of interest in the model. According to the analytical and computational models developed by Parikh and Mowrey (2014), traditional 90° racks in retail shopping environments result in lower exposure and intensity than racks at other angles. A previous study confirmed these model predictions with a simple target detection task (small red targets on empty grey racks) in a virtual environment. However, discriminating a target on a stocked shelf requires more time and is more representative of typical shopping behavior. Methods: The 24 participants completed 10 target discrimination trials as they were moved through a virtual shopping environment. Hypothesis: We hypothesized and found a significant effect of orientation on discrimination performance. Additionally, we hypothesized that the percentage of total targets correctly identified would be lower than the simple detection rate in Parikh and Mowrey (2014) but found mixed results. Model fit was first assessed via a d’ metric. The d’ values were generally low, but they were best at intensities higher than that needed for detection due to the additional time needed to identify the targets among distractors. However, the observed non-normal distributions of hits and false alarms make the d’ analysis difficult to interpret. Subsequently, a chi-square analysis was done. The chi-square analysis also showed evidence for higher intensities needed for discrimination than for detection in the 30°, 45°, and 90° rack orientations. Limitations and modifications needed for the model to achieve a better match to human discrimination performance are discussed.
Department or Program
Department of Psychology
Year Degree Awarded
Copyright 2020, some rights reserved. My ETD may be copied and distributed only for non-commercial purposes and may not be modified. All use must give me credit as the original author.
Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 License.