Spoken Instruction Accuracy Assessment for Phone, Radio and Shop-Floor WorkYou heard all of it. The question is how much of it reached your hands.

Methodology: Thirty-five original exercises across five formats: sixteen spoken instructions with written response sets, eight confidence probes on a six-point scale, four select-every-that-applies exercises, four keyed claims and three ordering exercises. The declared response instruction is knowledge and applied accuracy throughout - what the instruction you heard actually requires - and one instruction covers the whole form. Construct statement: it measures how much of a SPOKEN instruction reaches the action, which class of detail is lost when one is lost - a condition, a figure or the order of two steps - and whether the respondent knows which of their own answers are right. It does not measure hearing acuity and is not a hearing test, it does not measure English proficiency or accent, it is not a memory span or a clinical attention measure, and it says nothing about performance in any named job. Scoring is calibration scoring. The accuracy strand is chance-corrected against the authored answer priors carried on every option, and the confidence strand is scored as a Brier skill score against a reference respondent who is equally sure about everything, decomposed into calibration error and discrimination and reported as two separate figures because a person can be well calibrated and undiscriminating, or discriminating and over-confident. Half the spoken conditions deliberately do not fire, so applying every exception and ignoring every exception both land at chance as arithmetic rather than as a threshold, and half the confidence probes are worded as doubt and reverse-keyed so that agreeing with everything cannot look like calibration. Every spoken exercise carries a written equivalent; an exercise delivered in writing is recorded as such and left out of the listening figure rather than scored as listening. Constructs and sources: the phonological loop and the limits of holding spoken material (Baddeley and Hitch's working-memory model; Baddeley's later revisions); the serial position curve and the loss of late-arriving material under a competing task; closed-loop communication and read-back or hear-back as an error control in aviation and clinical handover (the read-back requirement in radiotelephony practice, and the SBAR and check-back literature in health care); the finding that confirmation of understanding is weaker than repetition of wording; gist versus verbatim memory for instructions (fuzzy-trace accounts of gist extraction); the calibration literature on confidence and accuracy coming apart, and the Brier score with Murphy's decomposition into reliability, resolution and uncertainty; acquiescence and balanced keying (Baumgartner and Steenkamp); and item-writing guidance from Haladyna, Downing and Rodriguez. All items are original works written for this instrument. No test, vendor, framework or standard is reproduced, named or implied, and no endorsement or affiliation exists or is suggested.