Academic journal article Perception and Psychophysics

Attention Modulates Set Representation by Statistical Properties

Academic journal article Perception and Psychophysics

Attention Modulates Set Representation by Statistical Properties

Article excerpt

Recent reports have claimed that observers show accurate knowledge of the mean size of a group of similar objects, a finding that has been interpreted to suggest that sets of multiple objects are represented in terms of their statistical properties, such as mean size (Ariely, 2001; Chong & Treisman, 2003, 2005a, 2005b). In the present study, we directed visual attention to a single set member and found that mean estimations were modulated according to the size of the attended item, regardless of whether size was the relevant search criterion (Experiment 1) or not (Experiment 2). These findings suggest that observers do not always accurately average together the entire set, and that instead the average is either biased by the features of the attended item, or based on a short-cut strategy of extracting the mean of a smaller subset.

In our everyday visual experience, objects rarely occur in isolation; instead, they are often part of a set of multiple objects of the same category, such as a number of oranges in a fruit bowl or leaves on a tree. Recent findings have suggested that the visual system may represent such sets of similar items by abstracting a schematic summary in terms of its descriptive statistics, such as the mean or distribution. When observers are briefly shown displays consisting of sets of circles with varying sizes and are asked whether a subsequently presented single test circle was part of the previous set, they are most likely to indicate that the test circle was part of the set when its size is close to the mean size of the set, and least likely to indicate that the test circle was part of the set when its size is outside the range of the set (Ariely, 2001). By contrast, observers are no more likely to correctly identify an actual member of the previous set as they do a nonmember whose size is close to the mean of the set. These findings suggest that, when one is perceiving sets of similar items, the visual system does not represent each item in the way that it does a single item, but instead may represent the set in terms of its statistical parameters-in this case, mean size. This process has been termed set representation, and the processing of a set of similar objects by extraction of overall statistical properties of the set, rather than individual item properties, provides an efficient way of processing complex scenes.

The extraction of the statistical properties of a set of items appears to be a robust process, and it applies to many stimulus dimensions, including orientation (e.g., Dakin & Watt, 1997; Parkes, Lund, Angelucci, Solomon, & Morgan, 2001), motion speed (e.g., Atchley & Andersen, 1995; Watamaniuk & Duchon, 1992), and motion direction (Williams & Sekuler, 1984). Moreover, the accuracy of judgments of mean size is almost as good as the accuracy of size judgments for a single item (Chong & Treisman, 2003). Mean judgment accuracy also remains good under difficult perceptual conditions, such as brief set exposure duration, or the insertion of a delay between two sets, the means of which need to be compared. Increasing the number and density of the elements in a set also does not lead to a deterioration in mean judgment (Ariely, 2001; Chong & Treisman, 2005b).

However, when processing large sets of similar items, we often search for (and find) one particular member of that set, rather than distribute our attention equally across all members of the set. It remains unclear precisely how this process of attending to a single set member affects the representation of the statistical parameters of the set. In the present study, we therefore investigated how mean judgments are affected when attention is directed toward a particular member of the set, and to what extent the effect is a function of the features of the attended item. We instructed participants to estimate the mean size of a set of circles while they directed their attention to either a small or a large target set member and reported the location of the target, before asking them to make the mean size judgment. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.