What Can Be Learned about Attention from Studying Its Development

AbstractThis paper stresses the importance of developmental research in evaluating theories of visual attention. Although researchers have long used theories developed with the help of college - age subjects to better understand children, there has been a reluctance to use developmental data to better understand theories of attention. I try to show that developmental studies can and do provide a unique vantage point from which to assess these theories. Three research steps are proposed: (1) theoretically - important differences and similarities between age groups are established, (2) theoretical constructs are mapped to these age differences/similarities, (3) data are collected to examine the relation between the constructs and age. Several examples of the use of this strategy are summarized. The variable of age is shown to play a role in testing theories of Gestalt grouping, perceptual organization, spatial orienting, and attentional filtering.How can research in the development of attention inform our understanding of the psychology of attention? I welcome the opportunity to discuss this question explicitly in a paper, largely because this is the way that I have often implicitly framed my research questions. I hasten to add, however, that I admit to this with some trepidation since I have not always felt encouraged to ask this question.I was trained in a graduate program that did not have a separate division called "Developmental." Instead, studies of the development of various processes and functions were carried out by professors who gathered together under such umbrellas as "Perception and Cognition," "Neuroscience," "Social Psychology," and "Learning." Right or wrong, the implicit assumption was that development was not a topic to pursue in its own right. Rather, one always studied the development of something. In my case, this was the development of perception and attention in humans. The first piece of research I could call my own in graduate school was a developmental exploration of a perception phenomenon that my advisor, Joan Girgus, had recently published with Stanley Coren (Coren & Girgus, 1980). They had shown that there are reliable distortions in the perceived distance associated with the traditional displays of Gestalt grouping. Adult subjects consistently underestimated the distance between dots within Gestalt groups and overestimated distances between dots that fall in different Gestalt groups.My project involved testing for the presence of these distortions in a total of 100 subjects - 20 each in five age groups between the ages of 5 and 24 years. The results were very clear. Although all age groups were equally accurate in estimating the distances between dots in control figures, the subjective distortions of distances in the Gestalt displays were much larger for the younger subjects.Now the traditional developmental approach to these data would be to consider their implications for theories of normative perceptual development. For instance, one implication is that human observers are better able to attend selectively to the task - relevant dots with increasing age, and conversely, to successfully ignore the configurational aspects of the Gestalt displays. I don't want to belittle this aspect of the data. I believe it is necessary and important to outline the normal course of perceptual development with these sorts of tasks. However, the particular angle on these data that interested me was their potential for shedding light on a long - standing debate within the mainstream of perception. What causes perceptual grouping per se? Accounts of perceptual grouping span a full range of possibilities, from those that rely on sensory or "hard - wired" mechanisms such as spatial filtering (Ginsgurg, 1978; Uttal, 1975), to those that attribute grouping to preattentive mechanisms in the early stages of visual processing (Julesz, 1975; Kahneman, 1973; Neisser, 1967), to those that appeal to "intelligent" or "constructive" mechanisms in later stages of processing (Gregory, 1978; Hochberg, 1982). …