Learning Transfer Assessment: Applying a Rule in an Unfamiliar DomainKnowing a rule and knowing where it stops are two different skills.
Seven rules, stated in the exercise, applied at three surface distances — and seven cases that look exactly like them and are not.
Learning agility is sold as a questionnaire. A self-rating of adaptability is a rating of self-image.
Everybody says they want people who can learn. What almost nobody measures is the thing learning is for, which is being able to use what you know somewhere you have not been. The instruments in this category are self-report: people rate how curious, adaptable and open they think they are, and those ratings are about self-image rather than about transfer.
This sitting has a key. Seven rules are stated plainly, one sentence each, inside the exercise that uses them, so nothing here rewards having heard of them before. Then each rule is put to work at increasing distance from where it was stated. The base-rate rule that explains why most flags from a screening test are wrong is asked again about a fraud filter, and again about a school's early-warning system. Nothing about the rule changes. Only the words around it do.
Half the cases are traps, and that is the design rather than a trick. Seven of the fourteen look exactly like the rule and fail one of its conditions: the thing being looked for is common rather than rare, the sample was not chosen by its outcome, the steps do not feed one another, nothing is riding on the number. Somebody who applies the rule everywhere and somebody who refuses it everywhere score the same, and both score at chance. That is arithmetic and not a threshold, and it is the only way to tell somebody who has a rule from somebody who has a habit.
The two directions are counted separately and named in plain words: carrying a rule past its edge, and failing to carry it at all. They need opposite advice — one person needs to say the condition out loud, the other needs to look for the parts rather than for a familiar situation — and a single accuracy figure hides which one you are.
Twenty-six further exercises cover the anatomy of a rule and its conditions, what a surface difference is and what a structural one is, how to check a transfer before it costs anything, and what to do with a rule of your own that has stopped holding.
What you walk away with
Chance-corrected against the authored answer priors, drawn as a band with the 68 and 95 per cent spans written out in plain words.
Near, different, and nothing-in-common — plus the boundary set on the same scale, so consistency is visible rather than averaged away.
Two scores computed apart on a balanced set, because being accurate with a rule and knowing where it stops are genuinely different abilities.
Carried past the edge, and left behind. Counted separately on seven items and seven, with opposite advice for each.
Each with what it gets you and what it costs when it runs unchecked, so no sentence reads as generic praise.
Chosen from your own lean, naming a specific situation and a specific behaviour rather than a quality.
Inside your report
Illustrative sample — your report is generated from your own responses.
Even at every distance, or strong near and falling away, are the same average and a completely different reading.
Eight items warrant applying the rule and seven warrant withholding it, so applying everywhere and refusing everywhere both land at chance.
| The rule | What has to be true first |
|---|---|
| Most positives from an imperfect test are wrong | The thing looked for has to be rare |
| An extreme reading is followed by a milder one | The reading has to have noise in it |
| A measure that becomes a target stops measuring | Something has to depend on the number |
The condition is the half people drop, and dropping it is what turns a rule you know into an answer that is wrong.
Built for
- Employers hiring for how quickly somebody picks up an unfamiliar domain
- L&D teams who want a keyed reading rather than a self-rating
- Graduates and career changers moving into a field they have not worked in
- Anybody who has confidently applied a rule where it did not hold
Seven rules, three distances, seven cases that only look right
40 exercises across five formats · about 45 minutes · a transfer figure with both spans, four strips of cells, and the boundary set scored on its own.
₹1,199 (incl. GST) · assessment and full report, nothing further to pay
Frequently asked questions
No. Every rule is stated in one plain sentence inside the exercise that uses it. That is deliberate: an instrument that rewarded having heard of a principle would measure exposure rather than transfer, which is a different thing and a much less useful one.
By balancing the set. Seven of the fourteen transfer items warrant applying the rule and seven warrant withholding it, so applying everywhere and refusing everywhere both land at chance as arithmetic rather than as a threshold. The two error directions are then counted apart.
One strip per surface distance, one cell per exercise, each carrying its marks as a printed number as well as a shade. A person who is even across all three distances and a person who is strong nearby and falls away can have the same average, and they need completely different advice.
No, and the report says so directly. It measures whether you can apply a stated principle where the surface is unfamiliar and whether you recognise a case that fails its condition. It says nothing about general intelligence, and it explicitly does not measure noticing that a rule applies when nobody has stated it.
About forty-five minutes for forty exercises. ₹1,199 in India, inclusive of GST, or US$11.99 elsewhere, one-time, for the sitting and the full report.
Each one takes a single capability, puts you inside the situations where it is actually tested, and scores your choices against published evidence — with a report designed for that capability alone, not a template. They span hiring, compliance, education, operations and personal skill.
Browse the catalogue →Methodology: Forty original exercises across five formats: fourteen keyed single-choice transfer items across three surface distances, eight select-every-that-applies items on the anatomy of a rule, six match-the-following exercises pairing rules with their conditions, six ordering exercises, and six true or false claims. Every rule used is stated in the exercise that uses it. Construct statement: it measures whether somebody can apply a stated principle in a setting whose surface is unfamiliar, whether they recognise a case that resembles the principle but fails its condition, and whether they can narrow a rule of their own that has stopped holding. It does not measure general intelligence, prior knowledge of any of the principles used, curiosity, openness, willingness to learn, or fitness for any particular role. Declared response instruction: knowledge and judgment throughout. Every exercise has an answer that follows from the rule as stated. Scoring: CHANCE-CORRECTED GRADED CREDIT against the item's own declared marginals. Every option carries an authored prior for the share of respondents expected to choose it, drawn from the misconception the option encodes, and the reported figure is the share of the distance between a respondent answering at those priors and full marks. Zero therefore means no better than a respondent answering the way people typically answer, not no better than a coin. Seven of the fourteen transfer items warrant applying the rule and seven warrant withholding it, so applying everywhere and withholding everywhere both land at chance by arithmetic. The two error directions - carrying a rule past its edge and failing to carry it at all - are counted separately, reported separately, and never averaged; the absolute figure decides the verdict. An unanswered exercise leaves both sides of the ratio. Reliability is reported as McDonald's omega rather than alpha, stated as an assumption until observed data replaces it, and any part too short to carry a number receives a three-way placement with the arithmetic printed. Construct grounding, drawn across sources rather than from one framework: analogical transfer and the finding that people fail to use a solution they have just learned when the surface differs (Gick and Holyoak, 1980; 1983); structure mapping as the account of what actually transfers (Gentner, 1983; Gentner and Markman, 1997); the distinction between surface and structural similarity in problem categorisation (Chi, Feltovich and Glaser, 1981); near and far transfer and the conditions under which far transfer occurs at all (Barnett and Ceci, 2002; Perkins and Salomon, 1992); the base-rate fallacy and the arithmetic of positive predictive value under low prevalence (Kahneman and Tversky, 1973; Meehl and Rosen, 1955); regression to the mean and its misattribution to intervention (Galton, 1886; Campbell and Kenny, 1999); selection on the dependent variable (Berkson, 1946; Geddes, 1990); the distortion of a measure once it carries a consequence (Goodhart, 1975; Campbell, 1979); the theory of constraints and throughput governed by the binding step (Goldratt, 1984; Little, 1961); overgeneralisation and the developmental evidence that a rule is first applied too widely and then narrowed (Marcus et al., 1992); McDonald's omega in place of alpha (McDonald, 1999); and the standard error of measurement as the reason a band rather than a point is reported (AERA, APA and NCME Standards, 2014). All items are original works. No item, scale name or report section is taken from any commercial instrument, and no affiliation with or endorsement by any vendor is claimed or implied.