A primacy effect in subjective probability revision.

Peterson, Cameron R.; Ducharme, Wesley M.

doi:10.1037/h0024139

Cited by 102 publications

(43 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For example, DuCharme (1970) argued that subjects were biased against making responses that appear extreme on the response scale used. This paper describes an experiment directed at investigating the formally irrelevant aspects of sample order and method of presentation.With regard to order, primacy has been reported by Peterson and DuCharme (1967) and by Dale (1968) while Pitz and Reinhold (1968) and Shanteau (1970) have found recency . Slovic and Lichtenstein (1971) argue that which effect occurs could depend on the form of the sample data.…”

mentioning

confidence: 98%

See 1 more Smart Citation

Order effects and display persistence in probabilistic opinion revision

Edenborough

1975

Bull. Psychon. Soc.

View full text Add to dashboard Cite

Sixteen subjects performed a bookbag and poker-chip task. On one run ambiguous data preceded unambiguous data; on another run, unambiguous data came first For half the subjects only the current data item appeared; for the others, all preceding items were shown too. Final responses were higher when unambiguous data came last, which was interpreted as a recency effect. Whether the data persisted had no significant effect.It has been suggested that suboptimal effects in opinion revision tasks, specifically those of the bookbag and poker-chip type, arise by the influence of artifactual variables. That is, performance is affected by aspects irrelevant to the Bayesian norm. For example, DuCharme (1970) argued that subjects were biased against making responses that appear extreme on the response scale used. This paper describes an experiment directed at investigating the formally irrelevant aspects of sample order and method of presentation.With regard to order, primacy has been reported by Peterson and DuCharme (1967) and by Dale (1968) while Pitz and Reinhold (1968) and Shanteau (1970) have found recency . Slovic and Lichtenstein (1971) argue that which effect occurs could depend on the form of the sample data. They suggest that recency effects would generally be expected; primacy had only occurred where data had first supported one hypothesis and then switched to supporting the alternative. In this case, subjects discounted the later data.In the present experiment, an attempt was made to manipulate sample data so that they would not be discounted. Samples were made up of an ambiguous part and a part pointing clearly to one hypothesis. The order of these parts was varied. Under primacy a higher final response should be made when the unambiguous data come first. Recency would be indicated by higher terminal responses occurring with the ambiguous half first.Order effects might also be influenced by how the data are presented. A particular aspect examined here was the persistence of the data display, whether only the current item or all data so far are displayed. It would be predicted that a persistent, or cumulative display would tend to enhance the order effects. Primacy would be enhanced because the persistent display preserves information that could otherwise be forgotten. Thus, with both persistent and nonpersistent displays, primacy would imply higher terminal responses when the first This paper is sponsored by Paul Slovic, who takes full editorial responsibility for it.. 39 half of the sample was unambiguous, but this effect would be greater with the persistent than with the nonpersistent display. If recency were found, when the ambiguous half is presented first the terminal response should be slightly higher with the persistent display. The effect will not be marked as the more recent items should not, in any case, be forgotten. Similarly, with the unambiguous half first, terminal responses should be slightly lower with the persistent display under recency. METHODA PDP-S computer was programmed to generate seque...

show abstract

mentioning

confidence: 98%

“…With regard to order, primacy has been reported by Peterson and DuCharme (1967) and by Dale (1968) while Pitz and Reinhold (1968) and Shanteau (1970) have found recency . Slovic and Lichtenstein (1971) argue that which effect occurs could depend on the form of the sample data.…”

mentioning

confidence: 98%

Order effects and display persistence in probabilistic opinion revision

Edenborough

1975

Bull. Psychon. Soc.

View full text Add to dashboard Cite

show abstract

“…The estimates were more conservative for extreme probabilities relative to the correct Bayesian probability, but there were no major differences between the final estimates of subjects making sequential estimates (as the cards were drawn) versus those making a single, final estimate. Peterson and DuCharme (1967), using dice and colored chips found that sequential estimates tended to be slow to follow the correct Bayesian values, with estimates less resistant to moving up (towards 100%) than down. Finally, in an experiment not involving estimating probabilities, Bruner and Potter (1964) studied subjects' ability to recognize pictures of common objects that gradually came into focus.…”

mentioning

confidence: 99%

Simultaneous versus sequential information processing

2011

View full text Add to dashboard Cite

“…If Bayes' theorem is to be an acceptable model of human inference, then these variables should not markedly influence human inference. For example, it has been shown that while Bayes' theorem is indifferent to the order in which the conditionally independent data occur, human behavior does reflect this factor (Peterson & DuCharme, 1967). The present study is concerned with another in this class of variables: the a priori probability of the event or datum on the basis of which inference is made-usually designated P(D).…”

mentioning

confidence: 99%