Misuses of the multiple sufficient causal scheme as a model of naive attributions: A case of mistaken identity.

Abstract

Much current work in attribution theory presumes that in multicausal situations, naive attributions are often characterized by the multiple sufficient causal scheme. The present research was designed to examine whether the multiple sufficient scheme or an alternate model, the variant-effect scheme, better characterizes naive attributions. The variant-effect scheme is a generalization of Kelley's graded-effects scheme and, in contrast to the multiple sufficient scheme, reflects a belief in the conservation of causal energy. Five-to 12-yearolds received a series of social inference problems. In each problem, subjects gave base-rate estimates of an event and also made attributions of the likelihood that the event occurred after hearing that a related effect occurred in the presence of another cause. Three critical tests consistently disconfirmed use of the multiple sufficient scheme and supported use of the variant-effect scheme. The results further indicated increasing use of the variant-effect scheme with development. Preference for a variant-effect scheme over a multiple sufficient scheme suggests a picture of a naive attributor who is perceptually sophisticated but operationally naive-accepting causal simplicity rather than entertaining the possibility of causal indeterminacy.

When one model casts doubt on another: A levels-of-analysis approach to causal discounting.

“…Depending on how the supply of causal resources is manipulated, the problem may become computationally intractable. Kun, Murray, and Sredl (1980) provided a similar account of discounting; they drew an analogy to physics with the notion of conservation of causal force. Like Nisbett and Ross (1980), Kun et al argued that there is a limited amount of total causal force available to explain a given event.…”

“…The models described in this article represent a wide range of approaches to understanding causal attribution and discounting. We presented both descriptive and formal models that represent many different disciplines, including social psychology (e.g., Kelley, 1972b), developmental psychology (e.g., Kun et al, 1980), and cognitive psychology (e.g., Einhorn & Hogarth, 1986). Among the formal models, we have reviewed a diverse set of attempts to account for discounting, including subtraction models, ratio models, and connectionist accounts.…”

Investigations in spontaneous discounting

“…When one cause is known to be present, people discount alternate causes as also being present, even though these multiple causes are not necessarily mutually exclusive (Einhorn & Hogarth, 1986). The attribution and causal reasoning literature is full of studies demonstrating this phenomenon (e.g., Hansen & Hall, 1985;Kun, Murray, & Sredl, 1980;Morris & Larrick, 1995;Nisbett & Ross, 1980;Read, 1987;Read & Marcus-Newhall, 1993).…”

Discounting Attributions and Multiple Determinants

“…Subjects are told that one cause is present and are asked if a second cause also determined the outcome. The normal criterion for a violation of discounting with this procedure is the adding of the second cause by at least half of the subjects (Kun, Murray, & Sredl, 1980). In the present study, rating scales were also included to measure the perceived strength of either cause and the degree of compensation between the two causes.…”