Premium
Probabilistic Connotations of Carcinogen Hazard Classifications: Analysis of Survey Data for Anchoring Effects
Author(s) -
Spedden Sarah E.,
Ryan P. Barry
Publication year - 1992
Publication title -
risk analysis
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.972
H-Index - 130
eISSN - 1539-6924
pISSN - 0272-4332
DOI - 10.1111/j.1539-6924.1992.tb00710.x
Subject(s) - anchoring , statistics , consistency (knowledge bases) , probabilistic logic , benchmark (surveying) , sequence (biology) , hazard , mathematics , econometrics , psychology , social psychology , chemistry , genetics , biology , geography , discrete mathematics , geodesy , organic chemistry
A study was undertaken to test for anchoring effects when numerical probabilities were elicited for carcinogen hazard designations. Subjects were asked to assign probabilities to both probable and possible. The sequence was randomly varied so that half of the subjects evaluated probable first and half evaluated possible first. While there was no consensus on the numerical probabilistic meanings assigned to these verbal expressions, in general, probable was assigned a higher probability than possible and there were specific values that were assigned frequently, indicating some consistency in interpretation. There appeared to be a fairly constant scaling factor between the probabilities assigned to the designations. Anchoring was manifested in two ways: a smaller difference between the designations when they were evaluated in sequence than when they were evaluated in isolation, and assignment of readily accessible “benchmark” values such as 50% and 75%.