On the interpretation of below-chance responding in forced-choice tests.

none - methodological/statistical paper, no sample

Adult forced-choice symptom-validity testing in clinical neuropsychology; the statistical argument is test-agnostic and transfers directly to any multiple-choice subtest.

A statistical exposition, not a study. It is included because it is the only peer-reviewed source that states the rule this archive needs - what a score at or near chance does and does not license - and because it is a rule about arithmetic rather than about a population, so it does not inherit the clinical sample's limits. It carries no effect estimate and no design; grade D is the honest ceiling.

Two claims, and the second is the one this archive was missing. (1) A score in the guessing range is uninformative about ability: the examinee's responses carry no signal, so the score is evidence about the administration, not about the person. (2) "Below chance" is a statistical statement with a critical value, not a descriptive one, and clinicians and researchers routinely make "serious errors in communicating what is guessing and what is worse than guessing." A raw score merely under the chance MEAN is not below chance; it must fall outside the sampling distribution of a pure guesser. Most commercial forced-choice tests do not in fact rest their cutoffs on the chance comparison, which is why the distinction gets lost.

Not applicable - no estimates, no sample, no causal claim.

A statistical argument rather than an empirical finding; replication is not the relevant standard. The binomial result it rests on is not in dispute.