A new smartphone was advertised with a 24-hour battery life. To verify the claim, the buyer checked the charge level each evening at plugging-in time and typically found it sitting at around 20 percent. Confirmation, apparently. The battery lasted.
What she never tested: how quickly it drained while running navigation, camera, or video. She never let the battery run low during a typical day of use. She never tracked it hourly across different usage patterns. She ran tests at the moments most likely to show a good result, and found a good result.
The battery, under real conditions, frequently died before 6 p.m.
What It Is
Congruence bias is the tendency to test hypotheses exclusively through direct testing - gathering evidence likely to support the current belief - rather than designing tests that could falsify it.
When evaluating a belief, individuals exhibiting congruence bias are more likely to select tests that confirm their current hypothesis rather than tests that could falsify it. This is more specific than confirmation bias, which describes a general tendency to prefer consistent evidence. Congruence bias operates at the point of test design: which test do you run, and which do you deliberately not run?
The result is a testing process that functions more like a search for reassurance than an honest enquiry.
The Mechanism
Congruence bias tends to arise from a preference for low-effort, confirmatory testing strategies. Confirmatory tests feel natural - they match the expectation, they are easy to design, and they produce the outcome that reduces uncertainty without requiring revision. Falsification requires more effort: you have to imagine what would prove you wrong, design a test capable of doing so, and then actually run it.
There is also a psychological cost to finding disconfirming evidence. If your hypothesis is wrong, you have wasted whatever was invested in it. Confirmatory testing shields against that possibility, at the cost of accuracy.
This preference is largely unconscious. People who exhibit congruence bias typically believe they are testing their hypothesis rigorously. The confirmatory structure of the tests is not usually noticed.
How It Differs From Confirmation Bias
Confirmation bias and congruence bias are related but distinct.
Confirmation bias describes the tendency to favour, seek out, and remember information that confirms existing beliefs - applied to how incoming evidence is filtered and weighted.
Congruence bias is more specific: it describes the test design itself. The person does not receive mixed evidence and discount the disconfirming parts. They design a test regime that never generates the disconfirming evidence in the first place. The problem occurs upstream of interpretation.
Congruence bias is not identical to confirmation bias - it is a specific manifestation focused on hypothesis testing, and it typically operates in situations where someone is actively investigating something rather than passively receiving information.
Where It Shows Up
In product evaluation. A consumer testing a product tends to use it in the conditions most favourable to success - and concludes it works well. Edge cases, high-load conditions, and adverse circumstances go untested.
In engineering and software. A developer who believes a bug is caused by a specific function may only test that function with inputs likely to reproduce the error. Other modules that could also produce the same output are not investigated, because the current hypothesis feels sufficient.
In business reasoning. An analyst supporting a preferred conclusion designs analyses likely to support it - choosing comparison periods, benchmarks, or data slices where the story holds - without running the same analysis on periods or slices where it might not.
The Falsification Move
The corrective for congruence bias is active: design the test most likely to disprove the hypothesis, and run it before concluding the hypothesis is correct.
This is not natural. It requires deliberately constructing the failure case. But a hypothesis that survives a genuine falsification attempt carries more weight than one that has only ever been confirmed.
A useful question: "If this hypothesis were wrong, what would I expect to see - and have I looked for that?"
See Congruence bias in everyday decisions
Pick a life context to see how this bias can show up outside the textbook.
The Promotion Pitch That Only Looked for Wins
A team leader builds a promotion case by highlighting only successful project outcomes while disregarding feedback about missed deadlines and skill gaps, resulting in an overconfident recommendation.
Scenario
Maya, a senior engineer, is asked to nominate a teammate for a lead role during the quarterly promotion review packet meeting. She believes the teammate is ready because they delivered two high-visibility features on time: Feature Atlas and Feature Nova. In the preparation meeting, she pulled the sales numbers for Feature Atlas and Nova, which looked good, and shared positive client comments. She did not request the teammate's recent peer review notes, which noted tendency to dominate discussions and miss deadlines, nor did she look at the sprint burndown charts that showed frequent task spillover. During the quarterly promotion review packet, she filled out the achievements section but skipped the required 'areas for improvement' part. When a junior colleague suggested checking those sources, Maya said the existing data is sufficient and moved forward with the nomination. The leadership team later discovered the overlooked concerns, causing a delayed promotion and a need for additional coaching. After the oversight, Maya created a simple checklist that forces her to ask for disconfirming evidence-such as peer-review notes and burndown trends-before any future nomination. This illustrates how congruence bias distorts promotion decisions when leaders only seek confirming evidence.
Where The Bias Enters
Maya seeks confirming evidence of readiness by selecting data that supports her hypothesis that the teammate is qualified, while avoiding or overlooking information that could falsify it, such as constructive feedback or performance trends that indicate gaps.
Decision Check
Before finalizing the recommendation, ask: What specific evidence would show the teammate is not ready, and have I actively sought that information?
This scenario is illustrative. It explains the pattern and does not claim a documented public case.