Disjunction effect
People wait to learn an outcome before acting — even when it wouldn't change the act.
What it means
The disjunction effect is a violation of the sure-thing principle: people who would choose the same action under every possible outcome of a pending event nonetheless postpone or decline that action while the outcome is still unknown. Classically, someone who would take a holiday whether they had passed or failed an exam still puts off booking until the result arrives — sometimes even paying a fee to wait — as though the resolved-but-irrelevant fact were needed to justify acting. The usual explanation is reason-based, nonconsequential thinking: under uncertainty no single rationale organises the choice. The effect is real but fragile: it replicates in some framings (notably the pay-to-know vacation scenario) yet repeatedly failed to appear in the original two-step-gamble paradigm, so it should be treated as a caution rather than a law.
Why it happens
The standard account is that people fail to think through the branches of a decision separately and then combine them. Shafir and Tversky called this nonconsequential reasoning: under a known outcome, a person constructs a clear reason to act (the trip is a reward for passing, or consolation for failing), and the reason drives the choice. When the outcome is unknown, neither reason is available yet, so no single rationale organises the decision and it feels premature. The felt need is not for the outcome itself but for the story that lets you justify acting. Waiting resolves the ambiguity and supplies the reason. On this view the effect is a byproduct of reason-based choice rather than a preference for information, which is why paying a fee to wait can look irrational even to the person doing it.
What the evidence shows
The original demonstrations used hypothetical two-stage gambles, a Hawaii-vacation problem, and a prisoner's dilemma in which players cooperated more when the partner's move was hidden than when it was known either way. The picture since has been mixed. Kuhberger, Komunska and Perner (2001) ran the two-step gamble four times, with real money and different payoffs, and found no reliable effect. Lambdin and Burdsal (2007) reported the same and argued some earlier support rested on a statistical artefact. A large pre-registered replication by Chandrashekar and colleagues (2021) recovered the effect in the pay-to-know vacation scenario but found it much weaker or absent in the gamble scenarios, succeeding in only a minority of tested cases. The phenomenon appears real but fragile and paradigm-specific.
Competing mechanisms and models
Reason-based choice is not the only candidate. A rival account is reluctance to think: uncertainty makes people postpone the effortful step of evaluating consequences at all, rather than actively constructing reasons. The two predict different things about how added deliberation changes behaviour, and tests have not cleanly settled the question. Separately, the effect became a showcase for quantum-probability models of cognition. Pothos and Busemeyer (2009) showed a Hilbert-space model can reproduce sure-thing-principle violations that an equivalent classical Markov model cannot. This is a claim about the mathematics of belief and interference, not about physics in the brain, and other researchers have since fitted the same data with ordinary probabilistic decision-process models, so the quantum framing remains contested rather than established.
Where it shows up and where it breaks down
The pattern surfaces wherever a choice is gated on a soon-to-resolve fact that would not actually change the choice: acting on a medical plan before a test result, signing a lease before a grant decision, committing to a project before a budget vote. Practically, the fix is to ask whether your action would differ across the outcomes; if not, the waiting buys nothing. But name the limit honestly. Genuine option value is real: sometimes the pending information does matter, or acting early forecloses a better move, and waiting is correct. The disjunction effect only applies when the states are truly irrelevant to the act, and the shaky replication record means it should be invoked as a caution, not treated as a law.
Examples
In experiments, students who would gamble again whether they had won or lost a first bet nonetheless declined the second gamble while the first outcome was still unknown.
A couple would move house whether or not the school application succeeds, yet they will not ring the estate agent until the letter lands on the mat.
A team agrees the reorganisation makes sense whichever way the funding decision goes, and still parks it until the budget is confirmed.
A patient would book the same overseas holiday whether a routine biopsy comes back clear or not, yet refuses to reserve anything until the pathology report arrives.
A student would enrol in the same graduate programme whether they pass or fail an optional qualifying exam, yet delays submitting the application until the exam score is posted.
First described in Amos Tversky & Eldar Shafir (1992).
Key references
- Chandrashekar, S. P., et al. (2021). Replication: Revisiting Tversky and Shafir's (1992) disjunction effect, with an extension comparing between- and within-subject designs. Journal of Economic Psychology, 82. www.sciencedirect.com/science/article/abs/pii/S0167487020301070
- Lambdin, C., & Burdsal, C. (2007). The disjunction effect reexamined: Relevant methodological issues and the fallacy of unspecified percentage comparisons. Organizational Behavior and Human Decision Processes, 103(2), 268-276. www.sciencedirect.com/science/article/abs/pii/S0749597806000409
- Kuhberger, A., Komunska, D., & Perner, J. (2001). The disjunction effect: Does it exist for two-step gambles? Organizational Behavior and Human Decision Processes, 85(2), 250-264. www.sciencedirect.com/science/article/abs/pii/S074959780092942X
- Pothos, E. M., & Busemeyer, J. R. (2009). A quantum probability explanation for violations of 'rational' decision theory. Proceedings of the Royal Society B: Biological Sciences, 276(1665), 2171-2178. royalsocietypublishing.org/rspb/article/276/1665/2171/72688/A-quantum-probability-explanation-for-violations
- Tversky, A., & Shafir, E. (1992). The disjunction effect in choice under uncertainty. Psychological Science, 3(5), 305-309. doi.org/10.1111/j.1467-9280.1992.tb00678.x
- Shafir, E., & Tversky, A. (1992). Thinking through uncertainty: Nonconsequential reasoning and choice. Cognitive Psychology, 24(4), 449-474. doi.org/10.1016/0010-0285(92)90015-T