Classical attribution theorists developed models of causal attribution that reflected their belief that people were primarily interested in attribution accuracy. These models did not consider contextual factors such as relationships and societal norms which resulted in the emergence of several empirical puzzles many of which are related to the use of consensus information. This study investigates whether the puzzle of the differential treatment of consensus information can be solved if it is assumed that people are primarily concerned with social features of the attribution setting rather than strict attribution accuracy. This study experimentally tests the role of key aspects of the social context such as the impact of social strategies in Kelley’s model of attribution to explore whether some of its empirical anomalies could have their origins in the social aspects of attribution in research contexts. The study found that participants were 2.63 times more likely to provide ‘inaccurate’ responses when there was a risk that the accurate answer would be socially disruptive. Findings from this study suggest that participants prioritise the implications of the social context over attribution accuracy.

Thesis (M.Soc.Sc.)-University of KwaZulu-Natal, Pietermaritzburg, 2009.