Inter-group scalability and dimensionality of Bardis' familism items

Abstract

This article discusses the inter-group scalability and dimensionality of Panos D. Bardis' familism items. One of the major approaches to studying attitudes toward family involves respondents-completed attitude scales. One such instrument is Bardis' Familism Scale. The Bardis' scale consists of sixteen Likert-type statements scored in a standard five-point format. Scale reliability was determined using both test-retest and split-half techniques and validity was argued on the basis of scoring trends in known groups. As originally conceived, the items were designed, in part, to reflect Burgess and Locke's ideal-type of familism: strong in-group feelings, family goal emphasis, common property, mutual support and the desire for perpetuation of the family. Over the years, the Bardis scale has been a relatively important instrument. It has not only been employed as a variable in empirical studies, but it has been reviewed in analyses of objective and the familism concept. However, in spite of its relative influence, the Bardis scale has also been subjected to severe criticism on several grounds. Two samples were drawn on which to test the Bardis instrument for scalability and dimensionality. The first consisted of 117 undergraduate students in introductory sociology at a Mid-western state university. Roughly 90 percent of these students came from the metropolitan area within which the university is located.