Research Transparency
Methodology
A complete description of how responses are collected, grouped, and counted — so anyone can examine the rankings and understand exactly how they were produced.
Why This Page Exists
The credibility of this research depends on one thing: the rankings must reflect what participants actually said, not what we think they should say.
This page documents every decision made between a participant's submission and the published Top 10. It is intended to be readable by HR professionals, leadership researchers, and anyone else who wants to evaluate the methodology before drawing conclusions from the results.
The grouping rules and this methodology statement are versioned. When anything changes, the version number increments and the change is recorded here. The current version is v1.3.0.
Step 1
How Responses Are Collected
Participants are asked to think of the best leader they have ever had and name three characteristics that made that leader unforgettable. They may optionally identify the type of leader they are describing (manager, coach, teacher, parent, etc.) and share a brief story.
Participation is voluntary and anonymous. No name or contact information is required. Participants who choose to provide their name or email do so at their own discretion.
Every response is stored verbatim in the database exactly as submitted. Nothing is corrected, paraphrased, or interpreted before storage. The original text is never modified.
Step 2
How Characteristics Are Grouped
Raw responses are grouped at query time — not at storage time — using a fixed set of rules applied consistently to every submission.
Verbatim preservation
The original submitted phrase is always preserved. Grouping is a read-only operation applied when calculating rankings. It does not alter the stored data.
Grammatical variants only
A group combines spelling variations and grammatical forms of the same root word — noun, adjective, adverb, and gerund forms. For example: honest, honesty, and being honest are grouped under Honest. These are the same characteristic expressed differently, not different characteristics.
Conceptually distinct characteristics stay separate
Words that are related but carry different meanings are not grouped together. Integrity, Honest, Trustworthy, Ethical, Transparent, and Accountable are separate entries even though they are thematically connected. Participants chose their own words; we respect that choice.
Ungrouped phrases rank independently
Any phrase that does not match a defined group variant appears exactly as submitted (with initial capitalisation) and ranks on its own. No phrase is discarded or excluded from the rankings.
Canonical label
Each group has a canonical label chosen for readability — the most natural standalone form of the root word. The label is not necessarily the most frequently submitted variant. All original phrases are shown in the alias drawer on the Top 10 page.
Step 3
How Rankings Are Calculated
Rankings are determined by the percentage of unique participants who named each characteristic — not by the total number of times it was mentioned.
Each participant is counted once toward a characteristic, even if they described it in more than one of their three responses. For example, if a participant submitted both 'honest' and 'honesty,' they count as one participant toward the Honest characteristic, not two.
The percentage shown is: (number of unique participants who named this characteristic) ÷ (total number of participants) × 100, rounded to one decimal place.
The number of mentions is also shown separately. This is the total number of times the characteristic appeared across all submissions, including multiple mentions by the same participant. The distinction matters: percentage of participants is the primary ranking metric; mentions is supplementary context.
In the event of a tie in participant percentage, characteristics are ranked by total mentions. Further ties are resolved alphabetically.
Data Integrity
What We Do Not Do
We do not modify submissions
Original responses are never edited, corrected, or deleted. The database record is permanent.
We do not weight responses
Every participant's response carries equal weight regardless of their background, role, or how they found the initiative.
We do not curate the rankings
The Top 10 is determined entirely by the data. We do not promote, suppress, or reorder characteristics to reflect any particular leadership philosophy.
We do not combine related concepts
Grouping decisions are based solely on whether phrases are grammatical variants of the same root word — not on whether they are conceptually related or commonly associated in leadership literature.
Test entries are excluded
Submissions flagged as test or administrative entries are excluded from research results. They remain in the database but are filtered out before rankings are calculated. Original participant responses are never deleted.
Complete Grouping Map
Every Group and Its Variants
The table below lists every defined group and the exact phrases that map to it. Any submitted phrase not appearing in this table ranks independently, exactly as written.
Inspecting the Groupings
Every ranked characteristic on the Top 10 page includes an expandable alias drawer showing the original phrases submitted by participants that were grouped under that label. This allows anyone to verify the grouping decisions independently.
Change Log
Methodology Versioning
When the grouping rules or counting methodology change, the version number increments and the change is recorded here. The current version is returned in the API response alongside every set of rankings.
2026-11-02
Separated 'forward thinker' from Forward-thinking — now its own independent group. Separated 'active listener' and 'active listening' from Good listener — now their own independent Active listener group. Database migration applied: added is_test boolean column (DEFAULT false) to survey_submissions. All existing submissions unaffected. Test submissions are now excluded from rankings at the database query level.
2026-11-02
Tightened grouping map to strict grammatical variants only. Removed non-grammatical phrases from Selfless, Calm, Team player, Forward-thinking, Principled, Knowledgeable, Good listener, Visionary, Mentor, and Coach groups. Added Composed as a separate group. Added isTest flag to schema.
2026-11-02
Separated Transparent from Openness/Open. Separated Authentic from Genuine. Separated Reliable from Dependable. Separated Persistent from Perseverance. Rankings ranked by percentage of unique participants.
2026-11-02
Initial methodology. Grammatical variants only. Conceptually distinct characteristics kept separate. Ungrouped phrases rank independently.