Trap/Cognitive Bias/No. 0511
Information Bias
Information bias is the tendency to seek information that cannot improve a decision. In decision research, Jonathan Baron, Jane Beattie, and John Hershey described it in 1988: people want more facts even when every possible answer leads to the same choice.
- Evidence
- Useful, modest evidence
- Read
- 6 min
- Links
- 9 connections
01You've seen this when…
- in life
You spend ten minutes comparing hourly rain forecasts. Each puts rain well above the threshold at which you take an umbrella.
- at work
Before approving a mandatory security patch, your team requests another usage report. The patch goes ahead under every possible result.
- out in the world
A council commissions another survey about a broken water pump. The pump must be replaced whatever residents say, and the repair waits for the report.
02The idea
A decision gives research a job. New information earns its place when it can improve the choice: change the option selected, adjust the timing or scale, or reveal a useful safeguard.
Information bias appears when people keep seeking facts whose possible answers leave that decision unchanged. The information may be accurate, interesting, and relevant to the general subject. Its value for the particular choice can still be zero.
Decision theory calls the expected benefit of learning something its value of information. A test can reduce uncertainty about what is happening while leaving the best action exactly where it was. Learning and deciding overlap, but each has its own goal.
The distinction from confirmation bias is useful. Confirmation bias steers the search toward evidence that supports an existing belief. Information bias concerns whether the search can improve the decision at all. Someone can collect balanced, reliable evidence and still fall into this trap.
03Why it happens
Several motives can make extra research tempting. These are plausible explanations; their importance varies across settings.
- Learning feels like progress. Reading a report produces an immediate reward: something previously unknown becomes clear. The information gap theory of curiosity helps explain that pull. The reward can persist after the decision has enough support.
- Uncertainty feels unfinished. A remaining unknown can keep demanding attention. For someone uncomfortable with ambiguity, another test offers the prospect of relief, even when the available actions stay the same.
- Research signals diligence. A manager can show the additional spreadsheet, interview, or expert opinion. Those visible efforts may be easier to defend than a judgment that the existing evidence is sufficient.
- Searching postpones commitment. Making a choice exposes someone to responsibility and possible regret. Another round of analysis gives them a concrete task while the commitment waits.
These motives can coexist. A team may enjoy learning, feel accountable for mistakes, and lack a stopping rule.
04A worked example
Consider an invented diagnostic exercise. A patient has one of three fictional conditions: A, B, or C. Their probabilities are 80%, 10%, and 10%. Each has a separate treatment that works only for that condition. The treatments have equal costs, and successful or unsuccessful treatment has the same consequences across conditions.
A test always comes back positive for B and negative for C. For A, positive and negative results are equally likely. The test takes time and money.
What it looks like The test promises a better-informed treatment choice. It perfectly separates B from C, removing one possible diagnosis.
What’s actually going on Either result leaves A as the best treatment choice. Among positive results, the probabilities become 80% A and 20% B. Among negative results, they become 80% A and 20% C. Treatment A has an 80% chance of working before the test and after either result. The test changes the remaining alternatives while leaving expected treatment success fixed.
What would have helped Write the treatment choice beside each possible test result before ordering it. Both branches lead to A, exposing the test’s zero value for this immediate choice. A separate decision about follow-up care would need its own analysis.
This simplified exercise isolates the logic of the bias. Actual medical tests can affect monitoring, prognosis, later treatment, and many other decisions.
05How to spot it
06What to do instead
- Name the decision precisely. Write down the choice, its deadline, and the available options. Deciding whether to install a patch differs from deciding when to install it or how to notify customers. Information may help one of those choices while adding nothing to another.
- Map answers to actions. Before requesting a report or test, list its plausible results and the action each would trigger. A small decision tree makes this visible. Identical actions across all branches are a strong signal to stop and examine the request.
- Price the possible improvement. If a result would change the action, estimate its likelihood and the benefit of that change. Compare that expected gain with the research cost and the opportunity cost of waiting. A study can have positive information value and still cost more than it delivers.
- Set a stopping rule before searching. Decide what evidence would be sufficient, what finding would reverse the choice, and when the decision must be made. Satisficing helps when the goal is to find an option that meets clear requirements. Keep searching when the remaining uncertainty could materially change the outcome.
07When it isn’t wasted research
Information can help with implementation or a later decision even when today’s choice stays fixed. A safety inspection might leave the decision to close a building unchanged while determining which repairs come first. Include those benefits in the analysis.
Research can also serve curiosity, enjoyment, documentation, or a legal duty. Each is a separate goal with its own value. The bias applies when someone treats extra information as improving a particular choice while its possible results leave that choice unchanged.
The answer-to-action check depends on a reasonable model of the situation. Overlooked options, unreliable assumptions, or unexpected findings can make supposedly irrelevant information useful. High-stakes decisions deserve a careful check of those assumptions before research stops.
There is also a naming collision: in epidemiology, information bias refers to systematic errors in measuring or recording data, related to measurement error. This entry covers the decision-making bias.
08Roots
In 1988, Jonathan Baron, Jane Beattie, and John Hershey put fictional illnesses into diagnostic problems. Invented disease names helped remove the influence of participants’ medical knowledge, giving the researchers control over the probabilities and what a test could reveal.
They were examining how people reason about diagnosis and information. Ordinary medical practice often links a clearer diagnosis with better care. The hypothetical problems let the researchers separate those benefits: a test could distinguish possible illnesses while leaving the preferred treatment unchanged. Some participants still wanted the test. The authors described this tendency as information bias.
The idea connected a psychological observation with an older decision-theory question: how much should someone pay to learn something before acting? It later reached broader audiences through Baron’s Thinking and Deciding. Its practical lesson travels easily from diagnostic tests to market research, comparison shopping, and project reviews: give each information request a decision it can improve.
09How solid is this?
The original experiments support the tendency in simplified diagnostic tasks. Evidence for its prevalence and size in everyday decisions is limited; the decision-theory test for information value is firmer than broad claims about excessive research.
10Connections
- Often confused with Measurement Error, Confirmation Bias
- Countered byDecision Tree, Satisficing, Value of Information, Optimal Stopping, Two-Way vs. One-Way Doors
- See alsoInformation Gap Theory of Curiosity, Opportunity Cost
11Origin and sources
Jonathan Baron, Jane Beattie, and John Hershey described information bias in their 1988 paper on heuristics and biases in diagnostic reasoning.
- [1]Baron, J., Beattie, J., & Hershey, J. C. (1988). Heuristics and biases in diagnostic reasoning: II. Congruence, information, and certainty. Medical Decision Making, 8(1), 25–30.
- [2]Howard, R. A. (1966). Information value theory. IEEE Transactions on Systems Science and Cybernetics, SSC-2(1), 22–26.
- [3]Baron, J. (2008). Thinking and Deciding (4th ed.). Cambridge University Press.
Suggest an edit· Updated 2026-10-02