How this assessment was built, scored and bounded

Several of the best-known instruments in this field are commercial products. They are sold under licence, their items are copyright, their scoring keys are trade secrets, and the serious ones restrict administration to practitioners holding specific qualifications. Those restrictions exist for good reasons and we take them seriously.

IVYBootcamps has not licensed any of them, does not claim to have, and does not reproduce their items, their scoring or their branding anywhere in this product. Where a proprietary instrument measures a construct that matters to career guidance, we measure that construct ourselves. That is possible because the constructs are public science — Jung's typology, the five-factor model, Holland's hexagon and the five-index structure of cognitive ability are all published, cited and freely available to build on. What is proprietary is not the idea; it is the particular set of questions somebody else wrote and the norms they gathered.

So the items in this battery were written by our team, the scoring keys are ours, and the norms are ours. The one exception is the five-factor section, where the items come from the International Personality Item Pool: its authors released it into the public domain expressly so that instruments like this one could use it, and we credit it rather than pretending we wrote it.

The practical consequence for a family is worth being direct about. Our instruments do not carry decades of published reliability and validity evidence, because they are new. What they carry instead is a stated construct, an item bank you can ask to see, deterministic scoring, published norms, and validity checks that run on every single submission. We would rather offer that and say so than borrow someone else's evidence by implication.

Nothing about a student is inferred. Class, board and stream are captured at intake, before a single item is rendered, because every downstream choice depends on them: the item bank, the norm table, the career map and the institution list are all keyed on that tuple. A Class 11 Commerce student served the Science bank produces a report that is worse than no report at all.

Three grade bands, three board groups and the streams available to each band give 27 distinct forms. Every item declares which bands, boards and streams it applies to, and a build-time check walks all 27 forms and refuses to release the application if any form is incomplete, if an item leaks into a stream it did not declare, if a key is duplicated, if an agreement scale carries no reverse-worded items, or if the assembled sitting falls outside 45 to 60 minutes. Non-crossover is enforced by the build, not promised in a document.

Classes 7 and 8 sit the Explorer edition. They are not asked to choose a stream, because in the Indian system that decision does not exist until after Class 10; the Explorer edition ends by recommending one instead.

A percentile is a statement about a reference group, and it is meaningless without one. Ours is the student's own grade band: a Class 7 student and a Class 12 student who answer the same number of reasoning items correctly have not demonstrated the same thing, and comparing them as though they had would be the single easiest way to mislead a family.

Each raw score is converted to a standard score against the mean and spread for that scale in that grade band, then expressed as a percentile. Percentiles are clamped to the range 1 to 99. Telling a fourteen-year-old they are at the zeroth percentile of anything does harm and buys no measurement precision.

Percentiles are then grouped into five interpretive bands, which is what the report leads with. The wording is deliberate: nothing is labelled “below average”, because a parent reads this document alongside their child and because a trait that is still forming is more accurately described as developing than as deficient.

This is the limitation a family is most entitled to know about, so it gets its own section rather than a footnote.

Provisional norms, version v1. Percentiles are referenced to a modelled distribution for this grade band, not yet to a measured cohort. Interpret bands and relative ordering with more confidence than exact percentile points.

In plain terms: our current means and spreads are derived from the published distributions of the underlying constructs and from the difficulty we designed each item to, rather than from a measured cohort of Indian students sitting this exact battery. Every norm row records its own source, the counsellor sees it, and it is printed on the report.

What follows from that is a rule about how to read the numbers. Bands and relative ordering — that this student's spatial reasoning sits well above their processing speed, that investigative interest is their strongest theme — are the parts that are robust to imperfect norms. An exact percentile point is not. Treat a 63rd percentile and a 68th percentile as the same finding.

Once roughly five hundred completed assessments exist per grade band, the norms will be rebuilt from the observed distribution and the norm version will be incremented. Every historic score keeps pointing at the norm version it was produced under, so nothing already issued to a family silently changes meaning when that happens.

This is a single-rater, unsupervised, self-report battery, which is the least reliable modality available. There is no parent form and no teacher form. That makes the validity indices load-bearing rather than decorative: they are how a counsellor knows whether to trust the profile in front of them.

Four checks run on every submission.

This is a design decision, not a disclaimer bolted on afterwards. The clinical instruments that cover adjacent ground are restricted to qualified professionals, normally require parent and teacher informants alongside the student, and are normed specifically to support diagnostic judgements. Screening an unsupervised teenager for anything of that kind through a career product would be both outside our qualifications and unsafe, so the domains this instrument covers were chosen to be educationally actionable and nothing more: attention during study, getting started, organising work, regulating impulse and emotion, learning strategy, and working with others.

A low score on any of them means a specific habit is currently costing the student marks, and the report responds with something concrete to do about it. It means nothing else.

If a student, a parent or a teacher is concerned about attention, mood, learning or wellbeing, that belongs with a qualified clinician, and our counsellors are instructed to say so plainly rather than to interpret this profile beyond what it can carry.

Consent to sit the assessment is captured separately from consent to receive marketing, and separately again from the terms of use. A psychometric result about a minor is a different category of data from a mailing-list opt-in and is treated as one.

Responses and scores are stored so the profile can be scored, the report produced and the counselling session prepared. They are visible to the student, to the counselling team working on that student's case, and to nobody else. They are not sold and not shared with third parties. Access is enforced at the database rather than in the interface, so it holds regardless of which screen someone reaches.

Every number in the report is produced by deterministic arithmetic over the student's answers. No language model is involved in producing a score. One model call writes the explanatory prose about numbers that have already been fixed, and it cannot change them.

To ask what is held about a student, to request a copy or to ask for it to be deleted, write to hello@ivybootcamps.in.