Developing an Empirically Based Linguistic Probability Lexicon
Individual variation in the interpretation and application of linguistic probabilities can have an adverse impact on how decision makers reach conclusions, particularly for stochastic events such as climate change. A popular method to ameliorating the problem of comparing subjective probabilities across persons is to derive a standardized verbal probability-quantitative probability lexicon, to minimize the interpretative mismatch between the emitter and receiver of probability words. Most attempts in the literature have proposed ad-hoc lists of words and cutoff points. Using data from a recent international study (Budescu et al.,2014), this study attempts to reduce the heterogeneity of this inter-person variation by systematically analyzing four methods of outlier elimination and examining the degree of agreement between the verbal probability lexicons produced via these different methods. Comparison of the four methods showed that Quade's pair chart method showed greatest phrase differentiation across trims, but that the numerical boundaries between adjacent phrases remained largely invariant to trim within each method. Sample validation of our various lexicons with UK and Australian English-speaking samples showed that our lexicons and the IPCC lexicons exhibited similar levels of compliance.
Psychology|Public policy|Quantitative psychology
Ho, Emily H, "Developing an Empirically Based Linguistic Probability Lexicon" (2014). ETD Collection for Fordham University. AAI10185404.