Diabetes results from a late change in the autoimmune response of NOD mice

Alfano, Joiner, and Perry (1994) used analysis of variance and analysis of covariance in a sample of college students to demonstrate that attributional style mediates between shyness and depression, as opposed to depression mediating between shyness and attributional style. They suggested, however, the hypothetical possibility of a third model, in which negative attributional style, rather than shyness, is the independent variable, and shyness serves as a mediator between attributional style and depression. In the current study, this alternative model was tested by using LISREL to reanalyze the Alfano et al. data. Results showed that this third model, with shyness as the mediating variable, fit the model quite poorly, in contrast to the preferred model of Alfano et al., which fit quite well. However, variants of this model, in which shyness functionally precedes both attributional style and depression, fit the data better. The substantive implications of this model for counseling practice and the methodological cautions about interpreting cross-sectional data for temporal sequences are discussed.