James S. Boster

Social position and shared knowledge: Actors' perceptions of status, role, and social structure

In this study of a university administration office, we explore the implications of variation among informants in their understandings of the structure of the group. Each office actor completed two similarity judgement tasks (pile sort and triad test) and two advice ranking tasks (personal and work advice) evaluating the other actors. We compare the patterns of judged similarity, the patterns of advice seeking, and the patterns of agreement among the actors on the four tasks. We find there is a consensus about the similarities of actors and that the structural position of actors influences their approach to the consensus. However, we also find that individuals who agree with each other are not necessarily those who are judged similar by other informants.

Estimating relational attributes from snowball samples through simulation

Abstract The difficulty of studying large networks is widely recognized by social network researchers. The fuzzy boundaries of networks and the high cost of data collection often make the exhaustive exploration of a social network impractical. Researchers have applied a variety of alternative procedures for sampling social networks, (such as snowball sampling) and have discussed the adequacy of the statistics for estimating a variety of network population parameters. However, little empirical work has been done on these estimates of social network parameters. With the use of computer simulations, this paper examines estimates of indegree centrality in snowball samples that vary in initial sample size, number of stages, and number of choices. The effects of varying sample parameters on estimated error is discussed.

Natural sources of internal category structure: Typicality, familiarity, and similarity of birds

Typicality ratings of 53 types of birds by University of California, Berkeley undergraduates (Rosch, 1975b) were more strongly correlated with the number of related species than with the frequency of the birds in the observers’ immediate environment or with the frequency of mention of the birds in written materials. Furthermore, the subjects made finer discriminations among the most typical birds (passerines) than among the less typical birds (nonpasserines), yet disagreed more often in identifying passerines and reported greater difficulty in judging the similarities among passerines. A single model explains both sets of results: passerines appear to be densely and continuously spread through the bird similarity space, whereas nonpasserines are more sparsely and discontinuously distributed, leading to the choice of passerines as both more typical and more difficult to categorize than nonpasserines.

Semantic structures and psychiatric diagnosis

This study examines psychiatric diagnosis using the concepts and methods of cognitive anthropology. Clinicians are viewed as cultural experts whose knowledge is represented in semantic structures, which are arrangements of diagnostic categories based on similarity of meaning. A test of similarity judgments of diagnostic categories was administered to a group of clinicians to derive two semantic structures of the categories included in axis I and axis II, respectively, of DSM-III. These structures were then compared with actual clinical diagnoses of concurrent disorders, and with the expected incidence of such diagnoses, in order to determine the possible influence of semantic similarity on the diagnostic process. The results suggest that semantic meanings of diagnostic categories influence the use of rule-out diagnoses more than positive diagnoses, and that this influence is more pronounced with respect to axis II than axis I.