Spatial Effects in Visual Selective Attention

In many visual selection experiments, Ss view displays of colored letters and numbers. They are instructed to attend to some stimulus dimension e.g., row location, color, class and are cued to report the items indicated by one value e.g., top row, red color on that dimension. Accuracy is always highest for row report. Since the items cued by row are spatially connected and easily coded for memory, and those cued by color or class spatially scattered and difficult to code, it cannot be concluded that row selection is more efficient than, say, color selection. In these experiments, selection criterion is held constant, and the spatial arrangement of the targets is varied. In Exp. I, Ss reported the identities of the 5 red letters appearing in a 5 x 5 matrix. Four types of target arrangements were tested. Four backgrounds, varying in degree of confusability with the targets, were combined factorially with the four target patterns. The effects of pattern and background and their interaction were highly significant. It is suggested that spatial arrangement per se is not crucial rather the target pattern serves to control the degree of background interference. Experiments II and III were detection analogs of the Letters background similarity condition of Exp.I. The Ss had to report whether an A or a T appeared among the red letters. The results for the pattern types were similar, indicating that there are spatial constraints on visual processing at a level low enough to be tapped in a detection task.