My Climate Risk Interdisciplinary Learning Group
8 April 2024; 13:00-14:00 BST (GMT+1)
Presenter: Marie Juanchich
Biography
I am a behavioural scientist with a PhD in socio-cognitive psychology. Over the past ten years, my work has been aimed at empowering people to make better decisions. I am especially interested in the way we can improve climate change uncertainty communication to improve climate related decisions.
Paper to be presented
Title: Dimensions of uncertainty communication: What is conveyed by verbal terms and numeric ranges.
Author: Karl Halvor Teigen
Link to paper: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC9660216/
Link to University of Essex page: Marie Juanchich | University of Essex
Session Highlights:
The April MCRILG session explored the topic of communicating uncertainty, particularly in relation to climate change, by taking the perspective of socio-cognitive psychology. This theme is particularly relevant to the MCR community as we grapple with the challenge of communicating climate science information, with its inevitable associated uncertainty, to communities and decision makers. A naïve choice of words related to risk can have large (often unintended) effects on how people perceive and understand the information, and therefore impact how they act (or not) on it. Using the correct language is not just a stylistic preference, but is at the heart of effective communication.
Professor Marie Juanchich from the Department of Psychology at the University of Essex took us on a fascinating intellectual journey diving into a review paper by Karl Halvor Teigen on “Dimensions of uncertainty communication: What is conveyed by verbal terms and numeric ranges.” In this paper K. H. Teigen, a prominent Norwegian psychologist who has done seminal work on decision making and probability judgments, provides us with some key tools in the evidence-based language “toolbox” for talking about uncertainty. K. H. Teigen was also Prof Juanchich’s mentor and collaborated with her on part of the research she presented. Further, as she shared with us, he is someone who she really appreciates intellectually and as a friend.
While almost every aspect of life is uncertain and we all have experience of it, Prof Juanchich was quick to show us that being able to communicate about uncertainty is actually not trivial at all. The critical issue is that verbal statements about uncertainty can be understood very differently by different people. This is due to a combination of prior beliefs, context of the communication, interpretation of intentions, cultural background and more. For example, the statement “it is likely to rain today” which canonically represents a 50% probability of occurrence, can be interpreted as 80% chance by person living in a usually rainy country, while a person in a usually dry country may understand it as 30% chance of occurrence.
One route taken by scientists, like in the Intergovernmental Panel on Climate Change (IPCC), is to create a taxonomy that regulates the use of probability words (e.g., unlikely = 10-33%, likely = 66-80%). However, this approach needs to be rooted in careful and solid evidence, to avoid it backfiring. Elements that need consideration include:
- People’s interpretation matters: the every-day use of words such as “likely” and “possible” may be very different to what the taxonomy states.
- Context matters: For example, a severe or generally more frequent event can be perceived as more likely (than the actual probability word says).
- (Perceived) intentions matter: As a practical example we explored the interpretation of the sentence “Maybe this is a mistake”. We were asked if the word “maybe” represents (A) true uncertainty or (B) a way to be nice while being quite certain about the mistake. The answers in the room were almost evenly split between the two interpretations. And it was noted that associated perceived probability with the mistake would be also different (A= 50% and B=80-100%)
- Words can nudge attention, and thus decision in a certain “direction”: the expression “there is a chance” attracts attention towards occurrence (positive direction) while “it is unlikely” attracts attention towards non-occurrence (negative direction). While these expressions are meant to represent the same probability, they nudge towards different decisions. It has been found that “positive” words encourage action more than negative. Interestingly, the IPCC uncertainty lexicon for probabilities below 66% has negative directionality, thus may encourage inaction.
- Probability words are also typically associated with specific outcome values when they characterise quantities (e.g., how much it might rain). For example, “unlikely” is typically associated with “fringe” values, taken from beyond the range of possible values on both ends of the value range. While “it is certain” is usually associated with minimal values. Hence probability words indirectly indicate the magnitude of the outcome that might happen.
Prof Juanchich also talked about the use of ranges of values to represent uncertainty. This is for statements about values, rather than about occurrence (yes/no). It turns out that ranges also have interesting interpretations. People tend to believe that smaller uncertainty ranges of values represent “more competent” advice and that people that deliver it are “more confident”. This is despite larger ranges are more likely to be correct, and in general there is no connection between range size and probability (if not explicitly provided).
After this “tour” of uncertainty language and its interpretation, Prof Juanchich closed with a few take away messages. She reiterated that probability words are very useful when numerical precision is not warranted. However, their choice needs to be made carefully and based on evidence, as every word will nudge towards an interpretation. We need to be aware of the power of the words we use, and choose them to align with our intended goals. For example, directionality is an important element to consider. Ranges of values can be a tool too, but one must be aware that they may convey more uncertainty than intended, so to be used effectively they need to be associated with a probability.
Prof Juanchich engaged in lively discussions with participants. Interventions included reflections on the lack of uncertainty in people’s discussions, especially around politics, responding to which Prof Juanchich noted that people don’t like uncertainty and being wrong, so it is not surprising that most stay away from framing their knowledge as uncertain. Another interesting discussion point was on whether the findings of this research are replicated across cultures. Our speaker noted that, while the findings presented spanned languages (from European countries to Japan), there is a lack of more extensive cross-cultural research, often because of resource constraints, so we cannot be certain, and this warrants further research.
Written by Elena Saggioro. Reviewed by Prof Juanchich.