Communicating risk: making denominators visible
In hydropower, healthcare, viticulture and Alpine tourism, managers must explain what an indicator actually allows people to anticipate. The management question is specific: how should a risk be presented so that the team distinguishes the frequency of a signal from the probability of the problem it is supposed to indicate?
What the literature establishes
Gerd Gigerenzer and Ulrich Hoffrage (1995, Psychological Review) show that, in Bayesian reasoning tasks, presenting information as natural frequencies can facilitate correct answers compared with a probability format. These frequencies describe nested counts drawn from a common reference population, rather than percentages with different denominators. The advantage therefore involves more than simply replacing the percentage symbol with a number of cases.
The mechanism: preserving the relationship between groups
Ulrich Hoffrage, Samuel Lindsey, Ralph Hertwig and Gerd Gigerenzer (2000, Science) extend this approach to the communication of statistical information, particularly in medicine. Interpreting a positive result requires distinguishing people who actually have the condition from those who show the same signal without having it. Presenting nested counts makes this distinction more accessible and reduces the reconstruction required of the recipient.
The premature conclusion
One might conclude that percentages should be banned from all management communication. These studies support a narrower conclusion: for certain conditional inferences, an appropriate frequency structure helps people connect the data correctly. They demonstrate neither that every raw count is clearer nor that a better answer to an exercise automatically leads to a better operational decision. (our executive and employee training programmes)
What the evidence does not guarantee
A readable format cannot correct a poorly chosen reference population, an unreliable indicator or outdated observations. Applying these findings to a hydropower risk, a vine disease or an Alpine route closure requires checking whether the categories and observation period are relevant. Routine practice remains vulnerable when rates are placed side by side without specifying who, or what, belongs in each denominator.
A practical check in Sion
For SHR’s “Communication” programme in Sion, an exercise worth testing is to select a documented indicator from participants’ work: a hydropower alert, a healthcare screening result, vine disease detection or a warning affecting an Alpine tourism activity. Using the same data, prepare a percentage version and a natural-frequency version with identical populations and periods, then randomly distribute them between two groups of participants. Ask each person, without assistance, to identify the reference group and estimate the share of signals that actually correspond to the problem; compare the proportion of correct answers against an answer key established before the exercise. This local check tests understanding of the message, not actual risk reduction. To go further: explore the Communication training in Sion in the canton of Valais, or browse our executive and employee training programmes in Switzerland.
In pictures: Communication in Sion
- communication
- Sion
- research
- training sion