UBC Theses and Dissertations
A two model description of attitudinal choice processes for subjects with high, medium and low involvement in three social issues Wood, Keith
The research reported here used Coombs' (1964) theory of data and evidence drawn from attitude change research to construct two models which, if correct, would describe the attitudinal choice and judgmental processes of, for the first model, an uninvolved S and, for the second model, a highly involved S. Both models were dependent on two of Coombs' (1964) eight classes of data, Petrusic's (1966) findings using single stimulus response latencies and the evidence from Sherif and Hovland (1961) and Ager and Dawes (1965) that a judge's attitude will affect his judgment of favourabiIity of alternative positions on a social issue. The test of the models occurred when single stimulus response latencies were collected from Ss who were required to accept or reject a position and then indicate if the position was more-pro or less-pro than his ideal position on issues of high, medium, and low involvement. The accept-reject task was, according to Coombs' (1964) formulation, Qllb data and the more-pro, less-pro task was QIIa(c.) (categorization relative to an ideal point). The data did not follow the predictions of the models for any of the four Ss used. Thus, our major hypothesis that an individual who is not involved in a social issue will judge alternative positions according to our first model (J-scale model) and that an individual who is highly involved in a social issue will judge alternative positions according to our second model (I-scale model) was rejected. Additional classes of data collected did, however, replicate and confirm the findings of Petrusic (1966). We were able to show that the latency data collected, whatever the attitudinal choice process involved, was reliable and orderly for each S over each issue. (Except for one S on one issue where a speed orientation was clearly shown). The order Iiness of the data was shown by the ability of QIIa(c.) and QIlb inferred orderings to predict the ordering of QIa (preference ordering) data.
Item Citations and Data