Engineering Drawing and Specification Reading Assessment for Manufacturing and Maintenance TeamsA drawing is a contract. The part that arrives wrong was usually made exactly as drawn.

Methodology: Forty-two original exercises across six formats: fourteen single-choice drawing calls, each a drawing feature, callout, tolerance, view arrangement or note carried entirely in words with one keyed reading; eight confidence probes on a 0 to 100 slider, each attached to one of those calls; six true-or-false claims about drawing conventions; six select-every-that-applies exercises asking what a described drawing leaves undetermined; four match-the-following exercises pairing a control or a note with what it governs; and four ordering exercises on how a drawing is read or an inspection is planned. Construct statement: this measures how accurately a person reads what an engineering drawing and its notes actually permit or require, and how well that person's stated confidence tracks whether each reading was right; it does not measure drafting skill, design judgement, machining or inspection skill, and it is not a qualification, a certification, an inspection authority or a permission to sign off a part. One response instruction is declared for the whole instrument and it is a KNOWLEDGE instruction: what does this drawing permit or require, never what the respondent would do. Scoring is C9, calibration scoring. Accuracy is scored on the thirty-four keyed exercises and corrected against declared answer priors: every keyed exercise carries an authored estimate of how ordinary respondents answer it, per option, per position, per match or per typical slider value, so zero on the accuracy scale means no better than the way people typically answer these exercises, and the priors will be replaced by observed shares once live data exist. An unanswered exercise leaves numerator, denominator and chance term together; an empty sitting scores exactly zero; a sitting with fewer than twenty keyed answers is refused a figure. Each of the five parts carries between five and eight keyed exercises, under the eight a figure needs and with an assumed omega below .70, so every part carries a three-way placement and the refusal of a number is printed in its own row. Calibration is scored separately and never combined with accuracy. Each probe asks for the chance, out of 100, that the answer to the call just given was correct, or, on four of the eight, the chance that it was wrong, which the scorer reverses; the doubt phrasing is an acquiescence control, and a respondent who gives the same number regardless of phrasing is flagged for review. Four probes are attached to calls declared hard, on which the keyed option carries a prior of .40 or less, and four to calls declared easy, with a prior of .55 or more, so that a real share of the probes sits on calls the respondent is likely to get wrong; a confidence probe attached only to easy calls is an agreement scale, not a calibration measure. The Brier score, the mean squared difference between stated probability and outcome, is decomposed by confidence bin into reliability, resolution and uncertainty after Murphy, and reliability and resolution are reported apart because a person can be well calibrated and undiscriminating, or discriminating and over-sure. The signed gap between mean confidence and accuracy is reported as the direction. A calibration figure needs all eight probes answered with their calls; under that the strand carries a three-way placement and the refusal is printed. The platform's own banded points on a probe compare the stated number with the declared prior on the call's keyed option; the report does not use those points. Every reported figure carries its standard error band from an assumed omega that is printed unrounded; no percentile appears anywhere because no norm group exists yet; a careless-responding flag count is computed for the operator and never shown to the respondent as a judgement. Both first-angle and third-angle projection are in use in the world and no exercise is keyed on which convention is correct; where projection matters the stimulus states the convention. Units are stated in every stimulus. Sources drawn on: Giesecke, Mitchell, Spencer and colleagues, Technical Drawing (projection, views and sectioning); French and Vierck, Engineering Drawing and Graphic Technology; Simmons, Maguire and Phelps, Manual of Engineering Drawing (drawing practice, title blocks and revision); the geometric dimensioning and tolerancing textbook literature, including Krulikowski, Fundamentals of Geometric Dimensioning and Tolerancing, Meadows, Geometric Dimensioning and Tolerancing, and Henzold, Geometrical Dimensioning and Tolerancing for Design, Manufacturing and Inspection (datum precedence, material condition modifiers, basic dimensions); Fischer, Mechanical Tolerance Stackup and Analysis (chain and baseline dimensioning); Busch, Fundamentals of Dimensional Metrology, and Curtis and Farago, Handbook of Dimensional Measurement (inspection planning and limits); Brier, Verification of forecasts expressed in terms of probability (1950); Murphy, A new vector partition of the probability score (1973); Lichtenstein, Fischhoff and Phillips, Calibration of probabilities: the state of the art to 1980 (overconfidence); Stankov and Crawford, Self-confidence and performance on tests of cognitive abilities (1997); Larkin and Simon, Why a diagram is (sometimes) worth ten thousand words (1987), and Chi, Feltovich and Glaser, Categorization and representation of physics problems by experts and novices (1981), on how experts read diagrams; Haladyna, Downing and Rodriguez (2002) for the item-writing rules; McDonald on omega; and Jacobson and Truax on the reliable change index. The geometric tolerancing concepts assessed here, such as datum precedence, basic dimensions and material condition modifiers, are common to the published standards in this field, of which ASME Y14.5 is one; that designation is a mark of the American Society of Mechanical Engineers, and this instrument is not affiliated with, endorsed by or derived from it and reproduces none of its text or figures. All exercises are original works written for this instrument. No real company, product, standards body or training vendor is named in any candidate-facing text. This instrument is not affiliated with any commercial instrument, training provider or certification body.