EXPLORATION of status systems within American communities has become a preoccupation of social scientists. Rossi has grouped the methods used for this task into subjective and objective types.' The first type records the assessments made by community residents of each other, on the assumption that in the last analysis expresses conventional definitions of relative position. The objective approach derives a community hierarchy from differential possession of such tangible attributes as income or occupation. The present study concerned with the subjective approach to status measurement, examining particularly the consistency with which different raters assign ranks. Corollary to this analysis, an effort made to throw light upon three topics: 1) the influence of the extent of acquaintance and the status of the judge upon the distribution of families he able to rate, 2) the effects of the judge's status upon the correlation of his ratings with those of the same families by other judges, and 3) the influence of the ratee's traits upon his chances of being known to the judges. While a total of 24 judges were used, the present analysis centers upon the four (called Q, U, V, and W) who endeavored to rate every family.2 The proportion of the total 1500 (white) families in the community who were rated by these four judges ranged from 64 to 91 percent. Thus if only one judge had been used, one-tenth to nearly two-fifths of the families would not have been placed at all. Moreover, the prestige was perceived quite diversely by the four judges (Table 1). Mr. U has a marked middle emphasis, while Mr. V regards the various classes as being of about the same size; and both of these judges assign many more families to an upper level than do Mr. Q or Mr. W. That these diversities in perspective are not wholly a reflection of selective acquaintance demonstrated by the second panel of Table 1 which deals only with the 743 families rated by all four judges. To rely upon a single judge's distributions of the families in the community (or even, as will be shown, an average of four or more ratings) for correlation with other aspects of community structure would inject a large element of unreliability into the study. Nor there any conclusive basis, even after we have all four sets of judgments to guide us, for preferring one or two sets of ratings. Perhaps the distribution bell-shaped, as Mr. W concludes; perhaps, however, a community of 1500 families has a skewed distribution (as viewed by Mr. U). The fact that occupations in the community are distributed rather evenly among the several levels does not warrant reliance upon the similar distribution supplied by Mr. V. Two major hazards jeopardize any choice among these different estimates of distribution: 1) the unknown distribution of the unrated families, 2) the varying degrees of inconsistency among rankings at different segments of the scale. In examining inconsistency among the judges the first step was to plot the six inter-rater comparisons of judgments. Since space does not permit a display of these tables, the two extreme comparisons are summarized. The closest agreement was between Mr. Q and Mr. W: they agreed on the placement of 33 percent of the families, * Revision of a paper read before the sixteenth annual meeting of the Southern Sociological Society, Chattanooga, Tennessee, March 27, 1953. P. H. Rossi, Latent Structure Analysis and Research on Social Stratification (doctoral dissertation, Columbia University, 1951), p. 88. 2 In order to measure consistency among raters it was necessary to have all ratings upon a common basis. Accordingly it was stipulated that the families were to be grouped in seven categories. (It was pointed out that people commonly use such terms as: They are 'the cream of the town,' They are just good substantial people, and They are at the bottom of the ladder, in talking about community position or social class. Each judge was asked to place the families in some one of the seven groups.) The objective of this study not to determine what the class system of the community is but the extent to which judges agree in their assessments of status.
Read full abstract