The AssessAll Blog
Assessment science, decoded weekly.
Measurement, hiring, proctoring, and the future of skills — written for people who make talent decisions.
Screen the Applicant or Certify the Workseeker? Two Places to Put the Same Assessment
A South African experiment with 6,891 workseekers found a credibly shareable skill certificate raised employment 17% and weekly earnings 34%; the identical scores, unshareable, moved employment not at all. Employer-side screening vs candidate-side certification, and when each is right.
Read articleThe EU Moved the Hiring-AI Deadline to 2 December 2027. A Practitioner FAQ on What Already Applies
The Digital Omnibus pushed the AI Act's high-risk obligations for recruitment systems to December 2027 — but prohibited practices, transparency, AI literacy and the GDPR already apply. Eleven questions assessment teams are asking, answered.
Dropping the Degree Requirement Changed Fewer Than 1 in 700 Hires: A Practitioner FAQ
Indeed put 19.3% of US postings at bachelor-or-higher in November 2025, up from 16.6% two years earlier, while Burning Glass and HBS found skills-based hiring touched under 1 in 700 hires. Eleven questions on what to assess once the credential filter comes off.
Fewer Than 3% of Applicants Admitted Using AI. Your Problem Is Variance, Not Cheating.
A 2026 study of 5,675 applicants found under 3% self-reported generative AI use, and use alone showed no systematic link to job-fit scores. The measurement risk sits elsewhere: five chatbots answered a teamwork SJT with a fifth of the human spread. Six decisions behind an AI-use policy.
Internal Mobility vs External Hiring: You Are Comparing Two Incomparable Kinds of Evidence
External hires were paid ~18% more and performed worse for two years in Bidwell's 5,260-worker study - but tenure carries .07 validity and top salespeople make measurably worse managers. Why the two candidate pools are judged on non-comparable evidence, and a seven-step fix.
Map Once, Read Anywhere: How to Make Your Skills Data Portable
ESCO has 13,939 skills, Lightcast 34,000+, O*NET 277 descriptors - and none of them record how well a person performs a skill. A seven-step walkthrough for attaching level, evidence type, date and identifiers so skills data survives leaving your system.
Skills Don't Have a Half-Life. They Have a Use Rate.
The five-year half-life of skills is a metaphor, not a finding. What 1,352 effect sizes on skill decay actually show: loss tracks non-use at roughly 0.06-0.08 SD per month, with variance so large that any single organisation-wide interval is guesswork.
Codified Knowledge Got Cheap. Tacit Knowledge Did Not. What That Changes About What You Assess
Stanford's payroll data shows a 19% employment gap for young workers in AI-exposed occupations, with declines concentrated in codified-knowledge roles while tacit-knowledge work grew. What the split does - and does not - mean for how you assess.
Tamper-Proof Is Not Trustworthy: What a Verifiable Credential Actually Proves
W3C Verifiable Credentials 2.0 and Open Badges 3.0 have solved credential provenance — but a signature says nothing about the rigour of the assessment behind the claim. The three layers of a credential, the seven evidence fields almost nobody publishes, and when a credential is genuinely enough.
The Agent-to-Agent Hiring Loop: What Survives When Both Sides Are Bots
Candidates' AI agents now apply to jobs that employers' AI agents screen — and every generated signal is losing value. Why detection is a dead end, and how live, proctored, scenario-based verification becomes hiring's trust layer.
The EU AI Act's Hiring Deadline Just Moved to 2027. Don't Waste the Extra 16 Months.
The Digital Omnibus pushed the EU AI Act's high-risk hiring-AI deadline from August 2026 to December 2027 — but bans, transparency duties and enforcement powers are already live. What HR and assessment teams should do with the deferral.
The 62% Premium for a Skill Nobody Verifies: Assessing AI Fluency in 2026
AI skills now carry a 62% wage premium and 144% posting growth, yet measured AI literacy averages 51.7/100 and self-ratings correlate inversely with real ability. How to actually assess AI fluency.
The Skills-Based Hiring Gap: Why 85% Say It and 0.14% Do It
Employers overwhelmingly claim skills-based hiring, yet degree-free hires barely moved and degree requirements are rising again. The gap is missing infrastructure: assessment-first funnels, integrity signals, and verifiable credentials.
From Marks to Skill Passports: Rethinking How We Prove Ability
A mark says you passed an exam once. A Skill Passport says what you can do, with evidence an employer can verify. Here's the shift from scores to verified competency records.
Want this thinking applied to your organisation?
Talk to the team behind the platform.
solutions@bodhih.com