Run a relevant WorkProof.
Participants complete a defined, versioned track before outcome data is interpreted.
MANDORKA VALIDATION PROGRAM
Authority is not claimed.
It is measured.
The central question is simple: does performance in a versioned WorkProof relate to meaningful differences in real work quality? We want employers willing to test that question with us.
Pre-registered hypotheses plannedVersion-controlled assessmentsNo fake significancePublish limitations
VALIDATION LOOP
Score → work outcome → calibration.
Measure real job-relevant outcomes.
Partners define observable quality criteria appropriate to the role instead of vague manager impressions.
Compare signal with outcome.
We examine whether total and dimension scores carry useful information and report uncertainty.
Improve the instrument.
Poor items, weak dimensions and unstable versions should be revised rather than defended.
Show what survived.
Mandorka will distinguish observed findings from hypotheses and publish important limitations.
Validation is continuous.
New tracks and major versions require new evidence. Historical results do not automatically validate future assessments.
FOUNDING RESEARCH PARTNERS
Bring a real workflow, not applause.
We are looking for hiring or operations teams that can define a real quality outcome and let Mandorka compare it with assessment performance. Participation does not imply endorsement, and negative findings are still useful.
Remote operations teams
Support, executive assistance, sales ops, finance ops, project ops, research, e-commerce and adjacent AI-assisted work.
Observable outcome
A concrete quality metric, work-sample rating or structured supervisor evaluation that can be compared consistently.
Public endorsement
A research partner can participate without becoming a testimonial or public case study.
REQUEST A VALIDATION PILOT