Cosmopolitan-Local: A Reanalysis of the Construct

Abstract

Most typologies of professional workers have followed the unidimensional polar cosmopolitan-local model proposed by Alvin Gouldner. This model views the latent role identity of the professional worker as constituted of two mutually exclusive foci, the wider generalized profession on the one hand and the local, situationally specific work setting on the other. A review of the original formulation and Aubsequent applications of this model, in research, leads to a reformulation of the concepts and an empirical test. The new perspective is that the two latent role-identity types, cosmopolitan and local, are two independent foci or dimensions and hence not intrinsically contradictory as proposed in the previous model. This perspective logically implies a minimum of four typologies, the two pure forms. the situation where neither of the two orientations is extant, and a mix of the two dimensions reviewed as striationally determined in its character. Measurement of the dimensions was based upon a five item Likert instrument for each dimension. These items were randomly distributed among 44 other Likert items in an 83 item questionnaire distributed to all full time faculty at three major midwestern universities, and one Canadian university. The average return rate was 59.9% with a usable respondent N of 1949 for the four institutions. The r es earch design involved predicting the outcomes of submitting responses for the ten items in a pooled format to principle components factor analysis. Because the interaction between the two dimensions was expecte4 to be a consequence of external variables unique to each study site, the data were factor analyzed by site population. It was predicted that the ten measurement items would load on two factors clearly delineating the two dimensions, and that where local unique factors were effective, a third factor would emerge displaying the consequential interactive mix among the items. The findings were correspondent to the predicted outcome. The first factor, representing the total dimension in each case, exhibited an average loading for all five items and at all four sites of .6425 and accounted for 47.65 percent of the explained variance. The second factor in each instance represented the cosmopol.Ltan dimension and exhibi ted an average loading of .5020 and accounted for 41.91 percent of the variance explained. For two of the populations a third factor also emerged constituted of a unique mixture of items in each case and exhibiting average loadings of .4800 and .5500 with a percent of explained variance of 19.03 and 21. 73 respectively. The conclusion drawn from this study is that the reformulation of the two concepts of cosmopolitan and local latent role-identity as independent foci of identity is more isomorphic with reality and hence potentially more useful.