Description
Present the same risk to matched groups as a frequency or as a probability and compare their decisions, because format alone can change how people act on identical information.
In the classic study, forensic psychologists and psychiatrists read the same case and decided whether to discharge a patient. Told that 20 out of every 100 similar patients commit violence, 41 per cent refused discharge; told that similar patients have a 20 per cent chance, 21 per cent refused. Frequency formats seem to call up images of actual people, making risk feel more real; experts were affected as much as anyone. As a research method this is a between-subjects format experiment: hold the information constant, vary only its presentation (counts, percentages, icon arrays, verbal labels) and measure decisions, perceived risk and comprehension. In UX it applies wherever an interface states a risk, error rate or confidence level.
Attributes
- Discipline
- Information Architecture & Content
- Loss type
- Siloed
- Stage
- Design, Evaluation
- Data
- Quantitative
- Effort
- Weeks
Steps
- Define the risk information and the decision or judgement it should inform.
- Create equivalent versions differing only in format: frequency (20 in 100), probability (20 per cent), and optionally icon arrays or verbal labels.
- Randomly assign participants to one format each, recruiting from the real audience including experts where relevant.
- Measure the decision, perceived risk and a comprehension check.
- Compare outcomes across formats with appropriate tests and effect sizes.
- Choose the format that leads to accurate understanding and the intended level of caution, and document why.
Sources
- Cognitive Psychology (Goldstein, 2019)
- On Second Thought: Outsmarting Your Mind's Hard-Wired Habits (Herbert, 2010)
Also known as
- Frequency versus probability risk format
- Natural frequency format test
- Risk communication format testing
Categories
- Content
- Information Architecture & Content