Abstract of Meeting Paper

Society for Risk Analysis 2002 Annual Meeting

Explanatory Power of the Psychometric Paradigm for Individual Risk Perceptions.* N. Bronfman and L. A. Cifuentes, Pontificia Universidad Católica de Chile

Although the psychometric paradigm has high explanatory power with aggregated data, poorer results are obtained at the individual hazard level, because the attributes influencing the perceptions may vary from hazard to hazard. To test this, we analyzed the results of a survey including 54 hazards and 10 attributes conducted in Santiago, Chile, at the aggregate and individual hazard level. Using OLS regression models for the perceived risk, we identified which of the attributes had the highest predictive power for each individual hazard and for groups of hazards. As expected, the models for each individual hazard had less explanatory power than for the aggregate analysis. For perceived social risk, the percentage of explained variance fell from 75-90% for the aggregate analysis to an average of 26-29% (range of 4 to 55%) for each individual hazard. The results for personal risk were similar. The explanatory power was higher for some groups of hazards. The variance explained for the 6 environmental risks analyzed was in average 34% (23-52%), and for the 7 technological hazards was 29% (19-55%). However, for the 6 transport hazards the average explained variance was only 19% (8-44%). The attributes with higher predictive power varied from hazard to hazard, with catastrophic potential, severity of consequences, and number of exposed people being significant predictors for perceived social risk for more than 50% of the hazards. None of the other attributes was significant for more than 10% of the hazards. Personal effect was the only attribute consistently associated with personal perceived risk.

The authors would like to acknowledge the financial support of Superior Education MECE Program of the P. Catholic University of Chile, and Conicyt through Research Project Fondecyt 1020118.

*Best Paper Finalist.


Go to . . .

2002 SRA Annual Meeting Table of Contents
2002 SRA Annual Meeting Author Index
Main Abstracts Menu Page
RiskWorld Home Page