Methodology and definitions

What the data is

Every figure in this index comes from one source: responses to the free assessments published at leadersadapt.com/assessments and leadersadapt.com/ai-readiness-assessment. Responses are submitted by the respondent, scored by the assessment engine in the browser, and written to the Leaders ADAPT contact record with the scores for each dimension.

No third party panel, no purchased data, no survey vendor. No response is solicited or paid for.

The edition window

The Q3 2026 edition covers completions recorded between 1 July 2026 and 19 September 2026 inclusive. A completion is dated by its recorded completion timestamp where one exists, and otherwise by the date the contact record was created, which for assessment respondents is the same event.

Editions are quarterly and do not accumulate. Each edition reports its own window so that a change between editions is a change in the population, not an artefact of a growing pile.

The four instruments

Each instrument is scored independently. A respondent may complete one, several or all four. Scores are not comparable across instruments, because the scales and the number of dimensions differ. Within an instrument, dimension scores are directly comparable to each other, and that is how every comparison in this index is framed.

The dominant type or style is the highest scoring dimension for that respondent. Where two dimensions tie exactly, the assessment records the first in its internal scoring order. This is a limitation of the instrument, it affects a small number of records, and it is not corrected after the fact.

The reporting floor

No cut is published with fewer than 30 responses. Where a cut falls below the floor, the chapter is published with the count and no figures, as Chapter 4 is in this edition.

The floor is 30 because below it a single respondent moves a percentage by more than three points, which is larger than most of the differences this index reports. The floor is set in advance of each edition and is not adjusted to make a cut publishable.

Where a reportable cut is close to the floor, the chapter says so and asks the reader to weight the rank order more heavily than the exact percentage. Chapter 3, at n = 66, carries that warning.

Rounding and arithmetic

• Percentages are rounded to one decimal place. Rounded shares may not sum to exactly 100.
• Means and medians are rounded to one decimal place on the instrument's own scale.
• No weighting, imputation or outlier removal is applied. Every valid completion in the window is counted.
• A record is valid if it carries a score for the instrument's first dimension. Partial completions that never reached scoring are not written to the record and therefore never enter the sample.

What this index does not claim

• It is not representative of leaders generally. Respondents chose to take a free leadership assessment on a website about executive coaching and AI advisory. That selection is likely to favour leaders who are already reflective about how they lead.
• It is not observer validated. Every score is a self report. Where a chapter discusses the gap between self perception and how a team experiences a leader, that gap is named as unmeasured, not estimated.
• It carries no demographic cuts. Role, company size, industry, region and tenure are not collected by the assessments, so the index cannot break results down by any of them. Adding one or more of these is the most likely change to a future edition, and it would be announced in the change log before the data was collected.
• It is not a longitudinal study. Editions are independent quarterly snapshots of whoever responded, not a tracked cohort.
• It does not measure outcomes. Nothing in this index links a score to company performance, and no such link should be inferred from it.

Privacy

Only aggregates are published. No individual response, name, email address, employer or free text answer appears in this index or in any chart, table or downloadable file derived from it. No cut below 30 responses is published, which also prevents a small group being identifiable by combination. Respondents are not contacted because of a result.

How to cite the index

Figures, charts and tables may be republished with attribution and a link to the canonical page. No permission request is needed.

Change log

Any change to the floor, the window rule, the scales or the instruments will appear here with its date before it affects a published figure.