The objective of the IPCC is to provide governments with scientific information and act as a guidance for decision-making and while developing climate policies. The IPCC reports are used by policymakers, managers, stakeholders, and the public, which have different levels of education, are trained differently, and interpret the results in terms of their goals, values, and agendas. Along with these differences the views on how to deal with uncertainty vary and can lead to misunderstandings (Rowe et al., 1994). Consequently, the IPCC uses a uniform, at first glance classifiable communication of uncertainty that can be understood by all readers. Uncertainty is divided into qualifiable probability and quantifiable likelihood, assigning the likelihood verbal quantifiers, such as “virtually certain”, “most likely” and “likely” (Mastrandrea et al., 2010).
But is the choice of verbal quantification over numerical quantification reasonable and does it make a difference to readers? This blogpost tries to give an overview of the respective advantages and disadvantages and outlines to what extent the choice of verbal quantifiers could influence decision-making.
While numerical likelihood quantifiers leave little room for vagueness, this is different using verbal quantifiers. The greater vagueness of verbal quantifiers allows greater flexibility, if the uncertainty of events or statements cannot be captured precisely (Teigen et al., 2003). For example, for ambiguous events, whose underlying uncertainties are so difficult to specify and measure, uncertainty cannot be determined numerically. With verbal quantifiers these uncertainties can be assigned to a range on the likelihood scale.
However, this vagueness also leads to problems: Understanding and translating verbal quantifiers into numerical quantifiers varies greatly from person to person. For example, the terms “possible” or “likely” are perceived variably in different contexts by diverse readers with different languages and cultures. For terms such as “almost certain” or “almost impossible”, studies find more precise perceived numerical translations, but they are still not consistent between individuals (de Bruin et al., 2000, Teigen et al., 2003). Budesco et al. (2014) discovered strong variations in the interpretation of likelihood quantifiers across 25 countries with 17 different languages. When using verbal quantifiers linguistic variability is not negligible, accordingly.
While the between-individual variance in interpretation is high, the awareness of this variance is underestimated, which increases the risk of miscommunication (Amer et al., 1994). This is not only the case for scientifically untrained people but can also be observed within scientists. For example, Teigen et al. (2022) showed that participants do not distinguish between “likely” and “most likely”. The judgments were not related to level of education, so there were no differences between the estimates of subjects with university degrees and low levels of education. Thus, the perception and estimation of verbal quantifiers does not seem to be learnable.
However, this vagueness in the interpretation of uncertainties is not random but show a pattern across individuals. For example, positive phrases are often interpreted as more probable than negative phrases technically expressing the same probability (Teigen et al., 2003). Thus, verbal quantifiers can function as a frame to direct attention to the positive possibility (‘likely’, ‘possible’) or to the lack of complete certainty (‘unlikely’, ‘doubtful’) (Teigen, 1988). According to Teigen (2014), the term “likely” is used 10-20 times more often in the 5th IPCC report than the term “unlikely”. Thus, the IPCC directs readers’ attention to what could happen, rather than informing them about events that are not as likely to happen.
Furthermore, the 2014 study by Budesco et al. showed that across the 25 countries, participants translating the verbal probability terms only achieved a 27% agreement with the IPCC standard. However, presenting participants with both the verbal and numerical quantifier increased agreement rates to 40%. This is still alarmingly low, but verbal terms can be accompanied by numerical values to reduce the discrepancy between meaning and understanding that arises from the reliance on verbal terms (Budesco et al., 2009). This combination can prevent misunderstandings and leaves less room for interpreting the uncertainty in one’s own favor.
But does the choice of verbal or numerical quantification influence the decision-making process? According to Liu et al. (2020), verbal quantifiers do not help to make faster, nor less stressful decisions than numerical ones. But with verbal quantifiers only, the decisions made are often less accurate and context-reliant. As soon as numerical quantifiers are added, the reliance on context is reduced.
In summary, the perception of probabilities is influenced by the communication style. Verbal quantifiers are more flexible, but also more vague than numerical terms and leave room for the reader’s own interpretations. Estimating numerical quantifiers is not possible for all kinds of events, but they enhance the clarity while reducing the room for misinterpretations. However, when verbal quantifiers are used, they can guide readers’ attention and frame their understanding. This can influence decision-making processes. Verbal quantifiers add a layer of directivity, but also increase the possibility of misunderstandings. Whether this is positive or negative, however, is something the IPCC and each actor must decide for themselves.
- Amer, T., Hackenbrack, K., & Nelson, M. (1994). Between-auditor differences in the interpretation of probability phrases. Auditing, 13(1), 126. ABI/INFORM Collection.
- Budescu, D. V., Broomell, S., & Por, H.-H. (2009). Improving Communication of Uncertainty in the Reports of the Intergovernmental Panel on Climate Change. Psychological Science, 20(3), 299–308.https://doi.org/10.1111/j.1467-9280.2009.02284.x
- Budescu, D. V., Por, H.-H., Broomell, S. B., & Smithson, M. (2014). The interpretation of IPCC probabilistic statements around the world. Nature Climate Change, 4(6), 508–512. https://doi.org/10.1038/nclimate2194
- de Bruin, W. B., Fischhoff, B., Millstein, S. G., & Halpern-Felsher, B. L. (2000). Verbal and Numerical Expressions of Probability: “It’s a Fifty–Fifty Chance”. Organizational Behavior and Human Decision Processes, 81(1), 115–131. https://doi.org/10.1006/obhd.1999.2868
- Teigen, K. H. (1988). THE LANGUAGE OF UNCERTAINTY. Acta Psychologica 68. 27-38.
- Teigen, K. H. (2014). When very likely is not so likely. Nature Climate Change, 4(6), 421–422. https://doi.org/10.1038/nclimate2256
- Teigen, K. H., & Brun, W. (2003). Verbal probabilities: A question of frame? Journal of Behavioral Decision Making, 16(1), 53–72. https://doi.org/10.1002/bdm.432
- Teigen, K. H., Juanchich, M., & Løhre, E. (2022). What is a “likely” amount? Representative (modal) values are considered likely even when their probabilities are low. Organizational Behavior and Human Decision Processes, 171, 104166. https://doi.org/10.1016/j.obhdp.2022.104166
- Mastrandrea, M. D. and Field, C. B. and Stocker, T. F. et al. (2010). Guidance Note for Lead Authors of the IPCC Fifth Assessment Report on Consistent Treatment of Uncertainties. IPCC Cross-Working Group Meeting on Consistent Treatment of Uncertainties. https://www.ipcc.ch/site/assets/uploads/2017/08/AR5_Uncertainty_Guidance_Note.pdf
- Smithson, M., Budescu, D. V., Broomell, S. B. & Por, H. H. Never say `not’: Impact of negative wording in probability phrases on imprecise probability judgments. Int. J. Approx. Reason. 53, 12621270 (2012).