To the Editor.
—Rosenblatt et al1 conclude, as a result of correlational analysis, that the organization, location, and mission of US medical schools are closely related to the production of rural practitioners. These conclusions are strengthened and refined by factor analysis of the matrix of intercorrelations between 10 selected variables that they provide in Table 4 (p 1563).Three underlying factors account for 66% of the variance of this correlation matrix. Factor analysis is appropriate for this data set since the variables are intercorrelated (Kaiser-Meyer-Olkin measure of sampling adequacy = 0.69; Bartlett test of sphericity,P<.001). The percent of variance accounted for and eigenvalues of the first three factors are 35.8% (3.579), 16% (1.603), and 14.8% (1.48). The fourth factor accounts for less than 10% of the variance and has an eigenvalue less than 1.000. A three-factor model seems appropriate.The factor loadings (of.6000 or greater) of the study
Morrow RC. The Supply of Rural Physicians. JAMA. 1993;269(6):744. doi:10.1001/jama.1993.03500060043015