Recognition of Prior Learning, Explained: An RPL Certificate Is an Assessment Result, Not a Service Record

Recognition of Prior Learning (RPL) is an assessment process that certifies the formal, non-formal and informal learning a person already holds, measured against a published qualification standard. It is assessment, not credit for service: India's national RPL guidelines state that "the length of time and years of experience are not calculated." An RPL certificate is a measured result, not a tenure claim.

That distinction matters more in India than almost anywhere, and it is the one most hiring managers miss when an RPL certificate lands on their desk.

Why India built an RPL system at all

The ILO's India Employment Report 2024 puts around 90 per cent of Indian workers in informal employment, and finds that only 4 per cent of youth have accessed formal vocational training. Of roughly 120,000 establishments surveyed, only about 25,000 offer apprenticeships (ILO, 2024).

So the modal Indian worker has real accumulated skill and no document that says so. RPL closes that gap by assessing what is already there rather than training it again. The scale is substantial: across three phases of the Pradhan Mantri Kaushal Vikas Yojana, the Comptroller and Auditor General records an RPL target of 55.80 lakh candidates, 65.87 lakh trained and 54.28 lakh certified — 97.28 per cent of target (CAG Report No. 20 of 2025).

What the rules actually require

India's RPL guidelines are more demanding than the certificate's reputation suggests. The design asks for:

• Evidence of prior learning at intake
• An experience letter, with a self-declaration affidavit plus a passed aptitude test permitted where no letter exists
• Document verification by the assessment agency, not the training partner
• At least 70 per cent mapping of the candidate's experiential learning outcomes to the qualification's outcomes before direct assessment is allowed; below that threshold, upskilling comes first
• Measurement, not interview impressions
• Every candidate assessed on practical skills, with theory online and practical in person
• Batch sizes of 20 to 30
• At NSQF levels 6.5 to 8, practical assessment carries 80 to 100 per cent of the weight
• A named, trained assessor
• Assessors and proctors certified through Training of Assessors programmes; Master Assessors through Training of Master Assessors for the highest levels
• An attached evidence trail
• Every assessment proctored and supported by digital and video evidence, geotagged

Read as a specification, that is a defensible assessment system: it front-loads eligibility evidence, refuses to let tenure substitute for demonstrated competence, and binds every result to an identifiable assessor and a retrievable record.

Why years of experience could never have been the measure

The guideline that drops tenure from the calculation is not bureaucratic fussiness. It is the psychometrically correct call.

The largest meta-analysis of pre-hire work experience — Van Iddekinge, Arnold, Frieder and Roth, Personnel Psychology (2019) — found corrected correlations with outcomes of r = .06 for job performance (44 studies, 11,785 participants), r = .11 for training performance (21 studies, 8,176 participants) and r = .00 for turnover (32 studies, 11,676 participants). Even experience specifically relevant to the target job's tasks reached only r = .07 with job performance. The authors' conclusion is unusually direct: organisations should be cautious about selecting on such measures "unless more positive evidence emerges" (open-access record; doi:10.1111/peps.12335).

In practical terms, "eight years as a steward" carries almost no information about how well someone will do the job. Years are a biographical fact, not a capability estimate. A system that converts years into a certificate manufactures a signal that does not exist; one that assesses the competence those years may have produced is doing real work.

Where the audit found the evidence trail breaking

A specification means something only if the record behind each certificate matches it. India's performance audit of PMKVY, tabled in December 2025, found that at volume it often did not.

• Eligibility data largely unrecorded or non-compliant. Of 60,68,523 candidates certified in job roles requiring education above class 9, 8,09,046 (13.33 per cent) did not meet the minimum qualification and 6,77,807 (11.17 per cent) had no education information recorded at all.
• Technical prerequisites unmet on record. Of 1,23,533 candidates certified in job roles requiring a prior technical qualification, 1,05,493 (85.40 per cent) were recorded as having no technical qualification — no education, basic literacy only, or general schooling alone.
• The experience field itself. For job roles requiring at least one year of prior experience — Food and Beverage Service Steward (86,447 certified), LED Light Repair Technician (16,990) and Dairy Farmer/Entrepreneur (70,639) — the data model could record work experience but it was not being populated, so the criterion could not be enforced. Audit noted the field was still not in use under PMKVY 4.0 as of October 2024.
• Unattributable results. Assessor contact details were null or "none" in 3,27,220 batches covering 93,01,810 certified candidates; trainer ID was null or zero in 2,30,758 batches covering 61,12,030 certified candidates.
• Missing evidence artefacts. Geotagging was absent in 84 per cent of on-spot physical inspection cases, and only 13 per cent of short-term training batches since 2018 complied with the Aadhaar-based attendance requirement (PRS summary).

The Ministry's recorded response was that programme implementing agencies were responsible for pre-screening candidates, and that PMKVY 4.0 would add automated checks on minimum age and education.

Be precise about what this does and does not establish. It does not show that 54 lakh certified workers cannot do their jobs; many demonstrably can, and the programme's purpose — formal recognition for informally acquired skill — is legitimate and valuable. The finding is narrower and, for an employer, more consequential: for a large share of certificates the record cannot demonstrate that the eligibility conditions defining the construct were ever checked. A certificate's entire product is its record. When the record is incomplete, the inference an employer wants to draw is unsupported — which is exactly the standard the AERA/APA/NCME testing standards and EEOC selection guidance apply to any assessment used for a decision.

How to read an RPL certificate in a hiring funnel

Five fields are the certificate. Everything else on it is decoration.

1. The qualification pack and NSQF level. This tells you the construct and the difficulty band. "Certified" without a QP reference tells you nothing.
2. The awarding body and assessment agency. Verifiable against the regulator's registers and Skill India Digital.
3. The assessment route. Direct assessment, or assessment after upskilling? Below 70 per cent outcome mapping the candidate should have been upskilled first — so the route is informative.
4. The date. A 2017 certificate against a since-revised QP is a different claim from a 2025 one.
5. The result type. RPL is criterion-referenced: it is a pass against a standard. It does not rank candidates, so do not use it as a ranking variable.

Then do two things. Treat the certificate as an eligibility filter — it legitimately removes the burden of re-testing foundational literacy and numeracy at volume. And re-measure, briefly, the tasks you are actually paying for: a 20-minute work sample or AI-graded scenario on the two or three operations that define the role. That beats a longer general aptitude battery, and on pay-as-you-go credits it costs about ₹30 per candidate.

If you certify prior experience yourself

The audit's findings are a design brief for any employer, sector body or training provider running its own recognition process:

• Make the eligibility field mandatory and machine-validated at enrolment. If prior experience is the construct, an unfilled experience field is a void certificate — and finding that out in an audit two years later is too late.
• Store assessor identity with every result. A score no one can attribute is not a score.
• Keep integrity evidence attached to the result itself, not in a parallel folder: proctoring record, flag reasons and an integrity band travelling with the score, the way an AssessAll Skill Passport carries each attempt's evidence alongside its outcome.
• Re-score a random sample of past results against the current standard each year. Reliability decays quietly; audits do not.

The takeaway

An RPL certificate is only as good as the assessment record behind it, and India's own audit is the clearest evidence available of what happens when that record is optional. Read it for what it legitimately establishes — a criterion-referenced pass on a named standard — then spend twenty minutes measuring the handful of tasks that will decide whether the hire works.