Congruence Bias

The Full Picture

Wason gave 29 students the number sequence 2-4-6 and asked them to discover the rule by proposing test sequences. Most hypothesized 'even numbers ascending by 2' and tested only confirming examples: 4-8-10, 6-8-12, 20-22-24. All received positive feedback. None tested 1-3-5 or 10-7-3. The actual rule was simply 'any ascending sequence'—but students never discovered it because they only tested instances congruent with their hypothesis. Klayman and Ha reframed this in 1987 as a 'positive test strategy'—not irrational bias but a default heuristic where organisms test cases they expect to succeed rather than cases they expect to fail. The distinction matters because positive testing is often efficient when the hypothesis space is large and errors are expensive. Biology runs on positive testing. A squirrel that learned 'oak acorns are nutritious' confirms this daily by returning to oaks. It doesn't test buckeye nuts to falsify the hypothesis that only oak acorns are safe—because a single disconfirming test of an actually toxic nut could be fatal. The asymmetry in testing costs makes congruence bias adaptive: confirming that a known food source still works costs nothing; testing an unknown source costs everything. Honeybee scouts exhibit the same pattern at the colony level. A scout that found a productive flower patch performs waggle dances recruiting more bees to that patch—a positive test of 'this source works.' Scouts rarely investigate whether a competing patch might be better until their current source begins to decline. The colony's foraging strategy is built on repeated confirmation of known sources, not systematic falsification of alternatives. New Caledonian crows reveal where congruence bias breaks. These crows are among the few animals that test alternative hypotheses when solving tool-use problems—bending wire into hooks, selecting sticks of appropriate length through trial and error. They engage in something closer to Popperian falsification. The difference: tool use involves a physical environment where failed tests are cheap (the stick doesn't fit, try another) rather than potentially fatal (the berry is poisonous). The business version is A/B testing that only tests variations of the current approach. Companies confirm that their existing strategy works in slightly different configurations but never test fundamentally different strategies—because fundamental tests risk the current revenue stream. Falsification is Popper's ideal, but positive testing is ancestral pragmatism: when the cost of a failed test is existential, organisms confirm what works rather than risk discovering what doesn't.