Summative Assessment and Moderation

Purpose Summative assessment in project‑based learning (PBL) is the formal judgement of learner achievement at the end of a sustained project. It must produce defensible marks that can be reported within CAPS, demonstrate standards attainment, and inform iteration of future PBL designs. Moderation ensures those judgements are valid, reliable and fair.

Principles to apply

• Validity: assess what CAPS requires — knowledge, skills and values targeted in the project — not peripheral activities.
• Reliability: different assessors should reach comparable judgements; use clear criteria, exemplars and moderation to reduce variance.
• Fairness and accessibility: make reasonable accommodations for diverse learners; document adaptations.
• Authenticity and integrity: ensure work is the learner’s own; require evidence and authentication.
• Manageability: use assessment tasks and evidence that teachers can reasonably collect, mark and moderate.

Decide what is summative

• Products: artefacts, written reports, portfolios, models, posters.
• Performances: oral presentations, demonstrations, practical tasks, role‑plays.
• Process evidence (selectively): final iteration of research journals or reflection that demonstrates mastery.
  Select only the end‑point evidence that best represents the CAPS assessment standards for that Learning Area and Grade.

Designing rubrics for summative judgement

1. Choose rubric type

   • Analytic rubric: recommended for PBL. Breaks assessment into distinct criteria with separate marks (improves reliability and feedback).
   • Holistic rubric: quicker but less diagnostic; use only for short, single‑dimension tasks.

2. Define criteria (3–6 criteria typical)

   • Align each criterion explicitly to CAPS Assessment Standards/Learning Outcomes. Example criteria: Knowledge & understanding; Application & inquiry; Product quality/technical skill; Collaboration & contribution; Communication & presentation.
   • Use observable, measurable descriptors.

3. Set performance levels and descriptors

   • Use 4 levels (Exemplary/Proficient/Developing/Beginning) or 5 levels if preferred.
   • Write specific, differentiated descriptors for each level and each criterion (avoid vague language such as “good” without evidence).

4. Weight criteria

   • Assign percentage/value to each criterion to reflect relative importance and CAPS weighting. Ensure total = 100. (Example weights shown below.)

5. Provide exemplars

   • Add one or two annotated exemplars for each level to calibrate markers and support learner understanding.

Example analytic rubric (product + presentation) — total 100

• Knowledge & understanding (25)

  • 21–25: Demonstrates comprehensive, accurate application of relevant concepts; makes insightful connections to context.
  • 16–20: Accurate and relevant application with some linking to context.
  • 11–15: Basic understanding; partial inaccuracies; limited contextual links.
  • 0–10: Misconceptions or insufficient evidence.

• Research & inquiry process (20)

  • 17–20: Rigorous, appropriate inquiry; critical sources; well‑documented methods and reflection.
  • 13–16: Sound inquiry and documentation; some limitations.
  • 9–12: Limited inquiry; weak documentation.
  • 0–8: Little or no evidence of inquiry.

• Product/design quality and technical skill (30)

  • 26–30: Highly effective, refined final product; high technical accuracy and craft.
  • 19–25: Good product with minor technical flaws.
  • 12–18: Functional product with clear weaknesses.
  • 0–11: Poorly executed.

• Collaboration and contribution (15)

  • 13–15: Leadership and equitable teamwork; documented individual contribution.
  • 10–12: Reliable collaborator; clear contribution.
  • 7–9: Uneven participation; limited contribution documented.
  • 0–6: Little or no contribution.

• Presentation & communication (10)

  • 9–10: Clear, fluent presentation; excellent response to questions.
  • 7–8: Clear presentation; acceptable Q&A.
  • 4–6: Weak clarity; limited engagement.
  • 0–3: Unclear or incomplete.

Converting rubric scores to CAPS reporting

• Score each criterion, sum to a raw mark out of 100.
• Convert the project mark to the formal assessment mark as per your subject’s term allocation in the school assessment timetable (e.g., if project counts 60 marks of a 100‑mark formal task, scale accordingly). Ensure overall formal assessment complies with your subject’s CAPS assessment programme and school assessment policy.
• Record the final mark and the evidence referenced (product, observation record, video, learner reflection) in the learner’s file.

Authentication and evidence

• Require supporting evidence to confirm authenticity: process journals, drafts with dated teacher comments, logbooks, witness statements, photos, audio/video.
• Keep a signed declaration from learner (and group mates where relevant) about their contribution.
• For group projects, include an individualised assessment component (e.g., individual reflection, interview or a short test) to differentiate marks.

Internal moderation — recommended process Moderation is not optional. It provides internal quality assurance before marks are final.

1. Pre‑moderation (recommended for high‑stakes tasks)

   • Review tasks, rubrics and marking memoranda before learners begin.
   • Confirm alignment with CAPS and fairness, and check accommodations.

2. Standardisation meeting (before marking)

   • All markers discuss rubric, examine exemplars and agree on application.
   • Mark a small set of anonymised exemplar scripts/products together and discuss discrepancies until agreement.

3. Marking

   • Markers use annotated rubrics. Keep evidence of decisions (comments, timestamps, exemplar comparisons).

4. Post‑mark moderation (internal moderation of marks)

   • Sampling: recommended sample is at least 10% of scripts/products or a minimum number (e.g., 5–10 per class), including top, bottom and mid range. Where class sizes are small, sample at least 2–3 items from each performance band. Schools may follow DBE or district guidance for exact sampling.
   • Moderator reviews sampled work, compares marks to rubric descriptors and flags discrepancies.

5. Moderation meeting and adjustments

   • Convene a meeting between marker/s and moderator to reconcile differences.
   • Adjust marks only after evidence-based discussion and document reasons for adjustments.

6. Record keeping and sign‑off

   • Preserve moderated exemplars, annotated rubrics and a moderation form signed by marker and moderator.
   • Enter adjusted marks in the official mark book and retain supporting evidence for verification.

Sample internal moderation form (fields)

• Subject, Grade, Project title
• Task date, submission date
• Marker name and signature, Moderator name and signature
• Sampled learner numbers/names
• Summary of moderation findings (consistency, rubric application, authenticity)
• Adjustments recommended (list learner IDs and mark changes)
• Date and final sign‑off

Inter‑teacher and cluster moderation

• Where possible, involve colleagues from other classes or schools for moderation meetings or cluster moderation sessions. This strengthens standardisation across grades and schools.

Reliability and inter‑rater calibration

• Use anchored exemplars and conduct periodic cross‑marking exercises.
• If multiple assessors are involved, calculate simple agreement rates (percentage agreement) and discuss inconsistent cases.

Dealing with group assessment

• Separate group product assessment from individual accountability. Use peer assessment, individual reflections, and short oral viva or tests to individualise marks. Moderators must see both group products and individual evidence.

CAPS record‑keeping and reporting

• Ensure the summative project is recorded within the subject’s formal assessment plan and reflected in the school file for moderation and district verification.
• Keep clear evidence that links each rubric criterion to the relevant CAPS assessment standards and Learning Outcomes. Use a mapping table for auditability.

Common pitfalls and how to avoid them

• Vague descriptors: write specific, observable descriptors linked to CAPS.
• No exemplars: provide annotated examples for each performance level.
• Poor sampling for moderation: sample across the full ability range.
• Over‑reliance on group marks: include individual evidence in the summative judgement.
• Failure to pre‑moderate high‑stakes tasks: pre‑moderate to avoid rework and unfair tasks.

Practical checklist before finalising summative decisions

• Does each assessment criterion map to a CAPS Assessment Standard?
• Is the rubric clear, weighted and accompanied by exemplars?
• Have you documented authentication evidence?
• Was pre‑moderation and/or standardisation carried out?
• Has a representative sample been moderated and signed off?
• Are marks converted appropriately into the formal assessment record and the year plan?
• Are reasonable adjustments documented for learners with special needs?

Quick templates (copy and adapt)

• Analytic rubric template: columns for Criteria | Weight (%) | Exemplary descriptor | Proficient descriptor | Developing descriptor | Beginning descriptor | Score.
• Moderation form: fields for task details, sample list, findings, adjustments, signatories.

Final note Summative assessment of PBL must be transparent, defensible and auditable. Clear rubrics, careful mapping to CAPS, pre‑marking standardisation and a rigorous internal moderation process are essential to ensure that PBL outcomes are recognised reliably as formal learner achievement.