The AssessAll Blog
Assessment science, decoded weekly.
Measurement, hiring, proctoring, and the future of skills — written for people who make talent decisions.
The Same Leadership Programme Scored d = .69 and d = 1.22. Only the Evaluation Design Changed.
One Situational Leadership programme measured four legitimate ways produced effects from d = .69 to d = 1.22, and a control scale that should not have moved shifted too. A comparison of the three training-measurement designs and when each is right.
Read articleThe Mean Is a Choice, Not a Summary: Turning Individual Scores into a Team-Readiness Decision
Mean, minimum and maximum aggregation of members' ability scores each predicted team performance at .21 to .29, so the meta-analysis cannot pick your rule. An illustrative walkthrough of the five decisions behind one defensible team-level capability claim.
A 360 Moves Ratings by 0.15 of a Standard Deviation: Six Myths About Multisource Feedback
Meta-analysis puts improvement after 360-degree feedback at d = 0.05-0.15, and a third of feedback interventions make performance worse. Six claims about multisource feedback tested against the evidence: panel sizes, rater self-selection, cross-source disagreement, and development vs decision use.
Your Assessment Centre Measures the Exercise, Not the Competency. Here's What to Report Instead.
Meta-analysis puts exercises at 34% of assessment centre rating variance and competencies at 22%, and a 2026 development-centre study found perfect assessor agreement on scores that matched nothing external. What is actually defensible to report back to a participant.
The Rater Is the Instrument: Why Calibrating Assessors Beats Redesigning the Rating Form
Only 21-25% of variance in performance ratings comes from the person being rated. Frame-of-reference training moves that number - trained assessors tracked actual performance in 54.6% of rating variance vs 33.5% untrained. How to run a calibration session that works.
Testing Is Not Just Measurement: What 272 Effect Sizes Say About Quizzing After Training
Practice testing beats restudying by g = 0.51 across 272 effect sizes, and spaced retrieval beats massed by g = 0.74 after bias correction. Why the end-of-course quiz is mistimed, what retrieval practice does not do, and a seven-step design for L&D.
The Level 3 Decision: How to Prove Training Actually Changed Behaviour
A walkthrough of the three measurement designs L&D teams choose between when asked to prove training changed on-the-job behaviour — and what the meta-analytic evidence says about each.
Your Competency Framework Is a Vocabulary, Not a Measurement System
Competency frameworks name the right things and measure almost none of them. The evidence on self-ratings (r = .29), administrative manager ratings (interrater reliability .45), and what it takes to retrofit a real evidence layer onto a framework you already own.
The Accidental Manager: Why Your Best Performer Is Your Riskiest Promotion
82% of managers were never trained for the job — and most were never assessed for it either. Why promoting on past performance fails, and how scenario-based readiness assessment before the promotion fixes management's most expensive selection mistake.
Only 8% of Companies Have Reliable Skills Data. Here's How to Build a Real Skills Inventory
88% of organizations admit they can't see their workforce's skills, and half hire externally for skills they already have. Why self-reports and manager judgment fail, and how assessment-based skills inventories fix the visibility gap.
Performance Is Not Potential: Building a Leadership Pipeline on Evidence, Not Instinct
Only about one in ten organizations rate their leadership bench as strong, and HiPo quit-intent has jumped to 21%. Why nomination-based talent reviews keep failing, and how scenario-based, evidence-first HiPo assessment fixes the pipeline.
Training Without a Baseline Is an Anecdote: The Case for Pre/Post Assessment
Only about a third of companies measure L&D impact beyond completion rates. Here is how a disciplined pre/post assessment design turns training spend into evidence — baselines, parallel forms, delayed retention tests, and competency-level learning gains.
Want this thinking applied to your organisation?
Talk to the team behind the platform.
solutions@bodhih.com