Information Technology Competence Level Assessment for Role Placement and Team Capability PlanningFourteen decisions, five rungs of responsibility. Find out which rung your IT judgement is actually standing on.

Methodology: Forty original exercises across six formats: fourteen single-choice level calls, each printing the same five-rung ladder of responsibility as its options; seven select-every-line exercises; seven true-or-false claims; five matching exercises with more entries on the right than on the left; four ordering exercises; and three numeric estimates on a slider. DECLARED RESPONSE INSTRUCTION: one instruction is declared for the whole instrument and it is a KNOWLEDGE instruction. Every exercise asks at which level of responsibility a decision belongs, or what a described situation supports. No exercise asks what the respondent would do, would prefer, or would feel, and no self-report scale appears anywhere in the instrument. CONSTRUCT STATEMENT: this measures at which level of scope and responsibility a person's judgement about technology work actually operates, across planning and architecture, building and delivery, running and support, enabling and advisory, and managing and governance; it does not measure technical depth in any language, platform or tool, does not test anybody's code, does not measure seniority, pay, experience or how well a person would perform in a named job, and it is not a qualification or an entitlement to a role or a grade. SCORING DESIGN: chance-corrected proximity scoring. The fourteen level calls form one vector. That vector is ipsatised within the person, z = (call minus the person's own mean) divided by the person's own standard deviation, and every role level's key vector is ipsatised identically, so somebody who reads the ladder high and somebody who reads it low are compared on the shape of their calls rather than on where they anchor. The distance to each role level is the sum of the absolute differences between the two standardised vectors. That distance is then corrected against the EMPIRICAL null: the expected distance of a respondent whose answer to every situation is drawn from that situation's declared answer prior, estimated by a seeded Monte Carlo of six thousand draws inside the report builder, and never against a uniform draw across the five rungs. Zero on the printed scale therefore means indistinguishable from somebody answering the way people typically answer; one hundred means the shape of the level's own vector exactly. The band around each match is a seeded item bootstrap over the fourteen calls, so the width of the band is the item-sampling error of a fourteen-call strand rather than an assumption pulled from nowhere. Two role levels whose ninety-five per cent bands overlap share one rank and are printed as a tie group, and the report prints how many levels sit reliably above and below each row instead of treating rank one as a recommendation. The five areas are scored separately as keyed accuracy on all forty exercises, eight per area, each exercise corrected against its own declared prior: the share declared on the keyed option for a single-choice or true-or-false exercise, the expected F1 of a respondent ticking each line at its declared selection share for a select-all exercise, the mean declared pair share for a matching exercise, the mean declared position share for an ordering exercise, and the expected banded credit under the declared value distribution for a slider. On select-all exercises the declared shares are SELECTION shares, one per line, and they are not a distribution over the lines; on single-choice and true-or-false exercises they are a distribution and sum to one, with index zero meaning True. No binary prior is set at an even split, because an even split on a two-option exercise is a coin wearing a prior's clothes and the correction it feeds is a coin toss. REFUSAL RULES: an unanswered exercise leaves the denominator and the chance term rather than scoring zero, so a short sitting is short rather than bad. An entirely empty sitting scores exactly zero and is marked not reportable. A sitting with fewer than twenty-four of the forty exercises answered is not reported at all, and the refusal is printed where the figures would have been rather than beside them. The shortlist needs at least nine of the fourteen level calls answered. The flatness test runs on the RAW presses, always: a respondent who chose the same rung on every level call they answered has a vector with no shape, and the shortlist is refused in the reader's own words, which say that one repeated level call is one call however the scorer re-orients it. An area with fewer than eight exercises, or with an omega below .70 on the unrounded value, carries a three-way classification and no number. Every reported figure carries its standard error and both its sixty-eight and ninety-five per cent bands. No percentile appears anywhere, because no norm group exists for this instrument yet. FRAMEWORK ANCHOR AND SOURCES: the structure of five areas of technology work and five levels of responsibility follows the freely published European e-Competence Framework, issued as European Standard EN 16234-1 by CEN, the European Committee for Standardization, and available at no charge; the framework and the standard are the property of CEN, and this instrument is not affiliated with CEN, is not endorsed by CEN, and is not a conformity assessment against EN 16234-1. The five area names and the five level descriptors used here are our own words, written for this instrument. The European multilingual classification of skills, competences and occupations, published by the European Commission, is used as a secondary check that the areas described are recognisable across labour markets, again with no affiliation or endorsement implied. The measurement sources are: Cronbach and Gleser, Assessing similarity between profiles (1953), for why a raw profile distance confounds elevation, scatter and shape; Cattell, r-p and other coefficients of pattern similarity (1949), for the pattern-similarity tradition this scorer sits in; Fischer, Standardization to account for cross-cultural response bias (2004), and Fischer and Milfont, Standardization in psychological research (2010), for within-person standardisation and its effect on subgroup differences; Efron and Tibshirani, An Introduction to the Bootstrap (1993), for the resampled band; McDonald, Test Theory: A Unified Treatment (1999), and Flora, Your coefficient alpha is probably wrong (2020), for reporting omega rather than alpha; Jacobson and Truax, Clinical significance (1991), for the reliable-change threshold printed for a re-sitting; Chi, Feltovich and Glaser, Categorization and representation of physics problems by experts and novices (1981), and Ericsson, The influence of experience and deliberate practice (2006), for the finding that experts and novices differ in how they represent a problem rather than only in how fast they solve it, which is what a level call measures; Dreyfus and Dreyfus, A five-stage model of the mental activities involved in directed skill acquisition (1980), for staged descriptions of practice; Kane, Validating the interpretations and uses of test scores (2013), for stating the claim the report is allowed to make; Haladyna, Downing and Rodriguez, A review of multiple-choice item-writing guidelines (2002), for the item-writing and cue-control rules; Baumgartner and Steenkamp, Response styles in marketing research (2001), for the response-style risk that the within-person standardisation removes; and Gollwitzer and Sheeran, Implementation intentions and goal achievement (2006), for the if-then action the report ends on. All forty exercises are original works written for this instrument. No commercial instrument's items or name are used or implied, no vendor certification, training provider, product or competitor is named anywhere in the instrument or the report, and no single jurisdiction's law is required to answer any exercise.