Bid and Proposal Quality Assessment for Sales, Consulting and Grant TeamsThe bid that lost was not the cheapest. It was the one that never answered the question.

Methodology: Thirty-six original exercises across five formats. The spine is a marking task: twelve extracts, each six sentences from a bid or proposal response to a stated buyer question, and the respondent marks every line a reviewer should comment on. Behind each line is a hidden class, a genuine defect that would cost the bid or not a defect at all, and thirty-five of the seventy-two lines are defects against thirty-seven that are not, so marking everything and marking nothing are both worthless by arithmetic rather than by a threshold. The other twenty-four exercises are eight single-choice judgements about what a buyer's evaluator is actually scoring, six binary claims about how competitive bids are evaluated, five match-the-following exercises pairing a weakness with the change that fixes it, and five ordering exercises on the order a response is built in. CONSTRUCT STATEMENT: this measures whether a person can tell the line in a bid response that would cost marks from the line that is doing its job, and what they know about how evaluators score, price and check compliance; it does not measure writing ability, sales skill, persuasiveness in a meeting, domain expertise in any sector, whether a real bid would be won, or anything about the person's character. RESPONSE INSTRUCTION: knowledge, declared once for the whole instrument: what is worth a comment here, never what the respondent would do. SCORING DESIGN: C6 signal detection. Marking a defect is a hit, marking a fine line is a false alarm; d-prime is the distance between the two rates in standard units and criterion c is where the respondent sets the bar, and both are reported because a person who flags everything and a person who flags nothing both land at d-prime zero and are told apart only by c. The log-linear correction of one half per cell keeps both finite. The hit condition is the classic flag-or-leave call rather than a naming condition, because a respondent drawing from the declared priors on this bank scores a d-prime of about plus one half, above the scale's zero, so no fixed habit beats a real reviewer by doing nothing. Every marking line carries a declared prior, the share of ordinary reviewers expected to flag it, and every choice exercise carries a declared prior per option; the matching and ordering exercises carry a declared per-pair and per-position prior. A modelled typical respondent is computed from those priors and marked on the same axis as the reader, and it is a model rather than a norm group: observed answer shares replace it once live sittings exist, and no percentile is claimed anywhere. The two error directions, defects left in and fine lines flagged, are counted apart and never netted. REFUSAL RULES: an unanswered extract leaves the numerator, the denominator and the expected term together; a sitting with fewer than six extracts marked is refused a d-prime and the refusal is printed where the number would have been; every competency has fewer than eight exercises or an assumed omega under .70 and so carries a three-way placement, no number and no percentile, with the refusal printed; the assumed omega is estimated from item count and an assumed inter-item correlation of .22 and printed unrounded. A careless-responding flag count is computed for the operator and never shown to the respondent as a judgement. Sources drawn on: Green and Swets, Signal Detection Theory and Psychophysics, and Macmillan and Creelman, Detection Theory: A User's Guide, for d-prime, criterion and the log-linear correction; Toulmin, The Uses of Argument, for the claim-data-warrant structure behind the evidence items; Cialdini, Influence, on social proof and authority as persuasion levers and their misuse as unsupported claims; Oppenheimer, Consequences of Erudite Vernacular Utilized Irrespective of Necessity (2006), on needless complexity and judged competence; Kimble, Writing for Dollars, Writing to Please, and the plain-language readability findings collected there; Flesch and Kincaid on readability measurement; Newman, Proposal Guide for Business Development Professionals, and Sant, Persuasive Business Proposals, on proposal evaluation practice and the compliance matrix tradition; Lewis, Bids, Tenders and Proposals: Winning Business Through Best Practice, on evaluator behaviour; Thai, International Handbook of Public Procurement, on scored criteria, pass-or-fail gates and clarification practice; Kahneman, Thinking, Fast and Slow, on substitution, answering an easier question than the one asked; and Haladyna, Downing and Rodriguez (2002) for the item-writing rules. Every exercise is an original work written for this instrument, and no real company, buyer, government department, tender portal or country's procurement law is named anywhere. APMP is a trademark of the Association of Proposal Management Professionals; this instrument is not affiliated with, endorsed by or derived from that body, its certifications, or any other commercial instrument or certification body.