Competency Validation: Assessment Lead

Proof that you lead assessment to standard in live work.

This is a candidate guide, not a taught course. It tells you what evidence from your live work to collect, how to structure your portfolio, how attestation works and how to prepare for the professional discussion. You already hold the knowledge from the prerequisite courses. Here you show that you apply it, with the responsibility of a lead, over time and with observable effect.

Framework alignment. This validation sits in the ICoBC Credentialing Skills Framework (ICSF) at skill ASMT "Assessment and evidence" (VRITE pillar: Rigor), level L4 "Assure", stage C "Competency". At this level you lead assessment strategy for significant programmes, design workplace-based competency assessment and manage integrity risks, including generative AI. The stage definition is: applying the capability in a real context, usually work, with the responsibility of the stated level, over time and with observable effect.

Who it is for

• Programme and micro-credential leads
• Learning designers and teaching staff
• L&D and corporate academy leads
• Professional bodies and sector associations

Prerequisites

• ICB-P02 Assessment Design for Credentials: Valid, Reliable, Fair
• ICB-A02 Assessment Integrity in the Age of Generative AI
• ICB-A03 Workplace-Based Competency Assessment

Learning outcomes

1. Show that assessment strategies you lead produce valid, reliable and fair decisions.
2. Show how you manage integrity risks, including generative AI.
3. Show how you develop and calibrate other assessors.

Workload. About 18 hours of candidate preparation and assessment. Plan the time across several weeks, because the portfolio draws on work you have already done and on moderation data that needs time to exist.

How it is assessed. You submit a portfolio from live assessment work with moderation data. You provide an attestation from a person who can confirm your role and your work. You take part in a 60-minute professional discussion with an assessor.

Badge. "Competency Validation: Assessment Lead (Competency, L4)", alignment code ICSF-ASMT-L4-C. The badge is valid for 3 years. The badge evidence is the portfolio, the attestation and the discussion record.

A note on what the badge means. The badge records that a panel judged your evidence against the criteria at the time of assessment. It does not prove competency in any other job or setting.

Two example settings. Worked examples in this guide use two contexts. The first is a university, college or training provider, with a programme lead called Amara. The second is a corporate academy or HR department, with an L&D lead called Jonas. All numbers in the examples are illustrative.

Portfolio guidance

What the portfolio is for

Your portfolio answers one question: does the assessment work you lead meet the standard of the ICSF L4 "Assure" level in live conditions? It is a curated set of evidence with your own commentary. It is not an archive of everything you have done.

An assessor reads your portfolio before the professional discussion. They look for three things.

• Scope and responsibility. You lead the strategy. You do not only deliver tasks that someone else designed.
• Live work. The evidence comes from real assessment of real candidates or learners, not from a simulation or a course exercise.
• Observable effect. Your decisions changed something that you can show: a revised task, a better agreement rate between assessors, a closed integrity gap, a fairer outcome for a group.

Keep the portfolio short and organised. Most candidates submit between 8 and 14 pieces of evidence, plus a two-page strategy summary and a short reflective commentary for each piece. Quality of selection matters more than volume.

The five parts in detail

Part 1: Context and your role (one page). Describe the programme or credential, the number of candidates assessed per year, the assessors involved, the stakes of the decisions and your exact role. Use verbs that show leadership: set, approved, redesigned, chaired, audited, trained. If a team shares the work, say which decisions were yours.

Part 2: Assessment strategy (two pages). Summarise the strategy you lead. Cover what the assessment claims to certify, which methods you chose and why, how the methods fit together, how you set standards and how you review results. Link each choice to validity, reliability and fairness.

Part 3: Evidence of decisions. Show that the strategy produces sound decisions in live work. This part holds your moderation data, standard-setting or cut-score records, appeals and their outcomes, and analyses of results by group.

Part 4: Integrity and AI governance. Show how you identify and control integrity risks, including generative AI. This part holds your risk register, your assessment design changes, your rules on AI use by candidates and by assessors, and the disclosure statement described in the next unit.

Part 5: Assessor development. Show how you develop and calibrate other assessors. This part holds calibration session records, feedback to assessors and the change in their agreement or decision quality over time.

Rules for every piece of evidence

• Date it. Evidence should come from live work in roughly the last three years. Older evidence needs a note on why it still applies.
• Anonymise it. Remove names and identifiers of candidates and colleagues. Follow your organisation's data protection rules and any applicable law.
• Label your own contribution. Mark which parts you wrote, decided or led.
• Add a short commentary. For each piece, write 100 to 200 words: what it is, why you chose it, what it shows and what you learned.
• Do not alter records. Submit real data as it stands. A modest, honest result with a clear improvement action is stronger than a polished result that looks too good.

Worked example: A strategy summary from a college

Amara leads a 12-month micro-credential in digital skills at a training provider. Her strategy summary opens with the claim: "A holder can plan, build and test a small data workflow to the standard of a junior analyst." She then lists three methods: a workplace project observed by a line manager, a short oral review of the project decisions and a timed practical task. For each method she names the risk it controls. The oral review controls the risk that a candidate submits work they did not do themselves, including work produced with generative AI. She explains that the three methods together give evidence from different angles, so no single method decides the result alone. She closes with two review rules: a yearly look at pass rates by candidate group, and a rule that any task with less than 0.5 agreement between two assessors on a sample of 20 scripts is reviewed. The numbers are illustrative.

An assessor reading this can see the claim, the methods, the link to risk and a plan for review. That is the shape of good evidence at L4.

Practice activity: Draft your context page and strategy summary

1. Write your context page in under 300 words. Include the number of candidates, the number of assessors and your role in three verbs.
2. Write your strategy summary in under 500 words. Use this order: claim, methods, why these methods, standard-setting, review cycle.
3. Underline every sentence that describes a decision you made yourself.
4. Ask a colleague to read it and name the validity, reliability and fairness controls they can find. If they cannot name at least one for each, revise.
5. Keep the draft. You will add to it in the next module.

Self-check: Portfolio guidance

Selecting evidence

Build an evidence map

Before you choose artefacts, draw an evidence map. The map lists each outcome and criterion, and the pieces of live-work evidence that could show it. It helps you spot gaps and avoid overloading the portfolio with one kind of evidence.

Evidence for valid, reliable and fair decisions (Outcome 1)

Choose evidence that shows the quality of decisions, not only the quality of the design.

• Validity. A mapping from the assessment tasks to the claim the credential makes, and an example of a check that the tasks measure what they should. A record of a task you changed after a validity concern is strong evidence.
• Reliability. Moderation data. This can be double-marking samples, agreement rates between assessors, borderline reviews or the outcome of resolving disagreements. Include the method you used, the sample size and what you did with the result.
• Fairness. Evidence that you check outcomes for different groups of candidates, handle reasonable adjustments, and review appeals. Show what you did when you found a difference, not only that you looked.

Moderation data is a required item in the portfolio. Present it in a simple table with the sample, the agreement measure, the threshold you set and the action taken. State plainly if the result was below your threshold. Assessors value honest use of data more than a perfect result.

Evidence for integrity and generative AI (Outcome 2)

Generative AI is an ongoing integrity issue for all issuers. Detection tools are not a reliable basis for decisions, so do not build your evidence on them. Show instead how you design and govern assessment so that decisions rest on sound evidence.

Good evidence includes:

• A risk register that names integrity risks such as impersonation, unauthorised help, copying and undisclosed AI use, with a rating and a control for each.
• Design changes that reduce the risk, for example an observed workplace task, an oral follow-up on submitted work, or a task tied to local data that a general tool cannot know.
• Written rules for candidates on what AI use is allowed, how to disclose it and what happens if a rule is broken.
• Written rules for assessors and staff on AI use in their own work, such as marking support or feedback drafting, including what must stay under human decision.
• A case log, anonymised, that shows how you handled suspected misconduct, the evidence you used and the outcome. A decision that relied on a detection score alone is weak evidence.

AI disclosure statement. Your portfolio must include a short statement on your own use of AI tools in preparing it. Say which tasks, if any, you used a tool for, what you checked and that the judgements are your own. The statement is part of the evidence, and an honest, specific statement is expected. You remain responsible for every claim in the portfolio.

Evidence for developing and calibrating assessors (Outcome 3)

Show that you improve the work of other people.

• Calibration records. The date, the number of assessors, the sample marked, the agreement before and after discussion, and the agreed standard.
• Development activity. An induction plan, a guide or exemplar set you wrote, observation of assessors with feedback, or coaching notes.
• Change over time. For example, agreement between assessors across two or three cycles, or fewer overturned decisions on appeal.
• Assessor voice. A short anonymised comment or survey result from assessors, with your response.

Avoid presenting a training slide deck alone. A deck shows intent. Records of calibration and its effect show leadership.

Choosing and rejecting evidence

Use four tests for each candidate piece. Does it come from live work? Does it show your decisions? Does it have an observable effect? Can an assessor verify it, through the attestation or the data? If a piece fails two or more tests, leave it out.

Worked example: Moderation evidence from a corporate academy

Jonas leads the academy that certifies team leaders inside a large employer. Eight line managers act as assessors of a workplace project. Jonas includes a one-page moderation summary. In spring, two assessors independently rated a sample of 16 projects against the rubric. The agreement on the pass or fail decision was 12 of 16, or 75 per cent. His threshold was 85 per cent. He lists the four disagreements, notes that three involved the criterion "handles conflict", and records that the descriptor was vague. He rewrote the descriptor with two worked examples and ran a calibration session with all eight assessors. In autumn, a new sample of 16 showed agreement of 14 of 16, or 88 per cent. All numbers are illustrative. The summary also notes that one assessor remained below the group on two criteria, and that Jonas arranged observation and coaching.

This piece shows live data, a threshold, a decision, an action and a measured effect. It also shows honesty about a remaining gap.

Practice activity: Build your evidence map

1. Draw a table with the three outcomes as rows and the columns "Candidate evidence", "Date", "Your role" and "Gap".
2. List at least two candidate pieces for each outcome.
3. Apply the four tests to each piece. Cross out any piece that fails two or more.
4. Mark each remaining gap and decide how you will close it, for example by running a moderation sample this term.
5. Choose the final 8 to 14 pieces. Check that moderation data, an AI risk and disclosure item and a calibration record are present.
6. Write the 100 to 200 word commentary for three pieces.

Self-check: Selecting evidence

Preparing for the professional discussion

How attestation works

An attestation is a signed statement from a person who knows your work and can confirm it. It does not replace your evidence. It confirms that the evidence is yours, that the context is as you describe and that your role is accurate.

Choose an attester who has seen your leadership of assessment first hand. This can be a senior manager, a chair of an examination board, an external examiner, a quality lead or a client of your assessment service. The attester must not be a person who reports to you, and should not be a close relative or friend.

The attestation should state:

• The attester's name, role and organisation, and how long they have known your work.
• Their relationship to you and to the assessment work.
• That they have read the portfolio summary and consider it a fair and accurate record.
• Which specific claims they can confirm, for example that you chaired the calibration sessions or led the redesign of the task.
• The date and their signature or a verifiable electronic confirmation.

The assessor may contact the attester to check the statement. Tell your attester in advance, and give them the portfolio summary and the criteria. An attester who cannot confirm specific claims gives weak support, so ask them to be concrete.

What happens in the professional discussion

The professional discussion lasts 60 minutes. An assessor talks with you about your portfolio and your practice. It is a structured conversation, not a presentation. The assessor has read your portfolio, so you do not repeat it. The discussion tests whether the portfolio is your work and whether you can explain, justify and reflect on your decisions.

The assessor will usually cover these areas:

• Your role. What you decided yourself, and what the team decided.
• Valid, reliable and fair decisions. How you know your decisions are sound, and what you did when the data showed a problem.
• Integrity and AI. How you assess the risks, what controls you chose and why, and how you handle a suspected breach without relying on detection tools.
• Assessor development. How you calibrate others, handle an assessor who disagrees with the standard and measure improvement.
• Reflection. What you would do differently, and where you are still developing.
• Authenticity. Any question that checks the portfolio is your own work, including your AI disclosure.

The assessor records the discussion, and the record becomes part of the badge evidence. It will be used to judge your performance against the criteria in the section "Assessment and badge".

How to prepare

Preparation is mostly rehearsal of explanation. You should be able to explain any item in your portfolio in under two minutes and answer follow-up questions on it.

1. Re-read your portfolio as if you were the assessor. Mark every claim you would challenge.
2. Prepare a one-page note with the key facts: numbers of candidates and assessors, your thresholds, the dates of key events and the outcome of each action. You may refer to it if the assessor agrees.
3. Prepare for the "why" question. For each design choice, write one sentence on why you chose it and one on what you would change.
4. Prepare two integrity scenarios. For example, a candidate with a polished submission and a weak oral response, or an assessor who uses a drafting tool to write feedback on candidates' work. Think through your evidence, your steps and your decision.
5. Prepare a calibration story. A real case where assessors disagreed, what you did and the effect.
6. Rehearse aloud with a colleague. Ask them to interrupt and ask "How do you know?"
7. Check the practicalities. Time, platform, recording consent and any adjustments you need.

Be specific and honest. If you do not know something, say so and explain how you would find out. If your data shows a weakness, discuss it. A candidate who can describe a limit and a plan is stronger than one who claims no limits.

Worked example: A strong answer on an AI integrity scenario

The assessor asks Amara: "A candidate submitted an excellent report, but in the oral review could not explain two of the main decisions. What do you do?"

Amara answers: "First, I do not treat the report as proof of misconduct, and I would not use a detection score to decide. I check the facts we do have. I compare the report with the candidate's earlier work and the supervised task. I ask a second assessor to review the oral recording. Then I invite the candidate to a second conversation, tell them what concerns me and give them a chance to explain how they produced the work. Our rules allow disclosed AI help for drafting but require the candidate to explain every decision. If the candidate cannot show they met the criteria, the decision is that the evidence is not sufficient, and the outcome follows our published procedure, including appeal. In my portfolio, case 3 follows these steps, and I changed the task afterwards so that the oral review sits earlier in the process."

This answer names evidence, avoids reliance on detection tools, protects fairness, links to a written rule and shows a design change. Notice that she points to a portfolio item.

Practice activity: Rehearse the discussion

1. Write six questions you expect, two for each outcome.
2. Write a two-minute answer for each, with one specific example from your portfolio.
3. Hold a 30-minute rehearsal with a colleague who plays the assessor. Ask them to follow up each answer with "How do you know?" at least once.
4. Record yourself or ask for notes. Mark every answer in which you used "we" without saying what you did.
5. Revise your one-page note.
6. Send your attester the portfolio summary and the criteria, and agree a date for the signed statement.

Self-check: Preparing for the professional discussion

Assessment and badge

The task

The assessment has three parts. You complete all three.

1. Portfolio from live assessment work with moderation data. Submit the structure from the module "Portfolio guidance": context and role, assessment strategy, evidence of decisions, integrity and AI governance, and assessor development. Include 8 to 14 pieces of evidence, each with a short commentary, a moderation data summary and your AI disclosure statement.
2. Attestation. Submit a signed statement from an eligible attester, as described in "Preparing for the professional discussion".
3. Professional discussion. Take part in a 60-minute discussion with an assessor, who records it.

Before you submit, complete the final knowledge check below. It helps you confirm that you are ready. Your assessment decision rests on the portfolio, the attestation and the discussion.

Assessment criteria

The panel judges your evidence against these criteria. Each criterion is rated "Not yet", "Meets" or "Exceeds".

Pass rule

To earn the badge you must:

• Reach "Meets" or "Exceeds" on all criteria A to H. A rating of "Not yet" on any criterion means the result is not yet a pass.
• Include moderation data from live work in the portfolio.
• Include a signed attestation that the assessor has verified.
• Include a complete AI disclosure statement.
• Complete the 60-minute professional discussion.

If you receive "Not yet" on some criteria, you receive written feedback and may resubmit the missing evidence and take a short follow-up discussion.

Badge and evidence submitted

On a pass, you earn "Competency Validation: Assessment Lead (Competency, L4)", alignment code ICSF-ASMT-L4-C. The badge is valid for 3 years. The evidence attached to the badge is:

• The portfolio
• The attestation
• The discussion record

The badge records a judgement made against the criteria at the time of assessment. It does not prove competency in any other job or setting.

Final knowledge check

Sources and further reading

Status: November 2026. Confirm every title and publisher against the current official source before you rely on it.

Frameworks and standards

• ICoBC, Credentialing Skills Framework (ICSF) and the VRITE quality pillars. Available from ICoBC, info@icobc.net.
• SFIA Foundation, Skills Framework for the Information Age (SFIA). SFIA 9 is the current published version. SFIA 10 is in consultation, and the SFIA Foundation targets release around Q4 2027. Status: November 2026. Check the SFIA licence terms before you reproduce descriptors in any course material.
• AERA, APA and NCME, Standards for Educational and Psychological Testing. Joint publication of the three bodies, a standard reference on validity, reliability and fairness.
• ISO/IEC 17024, Conformity assessment: General requirements for bodies operating certification of persons. International Organization for Standardization.
• European Commission, Europass and the European Learning Model, for the description of learning outcomes and credentials in Europe.

Topic notes

• Generative AI and assessment integrity. Established and ongoing for all issuers. Detection tools are not a reliable basis for decisions. Rely on assessment design, oral verification, clear rules and evidence-based case handling. Status: November 2026.

Related ICoBC courses

• ICB-P02 Assessment Design for Credentials: Valid, Reliable, Fair
• ICB-A02 Assessment Integrity in the Age of Generative AI
• ICB-A03 Workplace-Based Competency Assessment