Twenty Out of Twenty: Moderating Internal Practical Marks
At a higher secondary school in Cachar, 34 second-year students appear for the Biology practical. The school's internal component is 20 marks, split as project 8, practical record 6 and viva 6. The subject teacher submits marks in which 22 students score 18 or above, and 6 score the full 20. The external examiner, due next week, spot-checked nine records: four have no labelled diagrams, two project reports contain identical paragraphs set in the same font, and one project is on a topic outside the syllabus. The principal tells the teacher to 'adjust so that no child suffers'. The teacher still holds the students' rough notebooks with dated lab entries, and the attendance register shows that one of the two students with the matching reports attended all eight practical sessions while the other attended three.
1. Should the teacher submit the marks as they stand? Argue the case on assessment principles and on procedural risk.
Read the worked answer
No. The marks as they stand cannot be defended on validity, reliability or fairness. Validity is breached where a record with no labelled diagrams still earns near-maximum marks in the record component, because the component is meant to certify a drawing and labelling skill that the evidence does not show. Reliability is breached by the clustering: 22 of 34 above 18 and six perfect scores is not a distribution produced by a criterion applied to 34 pieces of work; it is a distribution produced by goodwill. Fairness cuts both ways, since a student who drew and labelled everything is treated identically to one who did not.
Procedural risk is concrete. Internal marks are open to verification and moderation by the board and by the visiting external examiner, and nine records already inspected include multiple defects. If the discrepancy surfaces after submission, the whole school's internal marks in the subject become suspect and every honest candidate is penalised. 'Adjust so that no child suffers' is not a marking criterion, and acting on it knowingly certifies evidence that does not exist.
2. Design a moderation procedure the teacher can complete in one week for all 34 students, describing the steps and the record to be kept.
Read the worked answer
First, publish the criteria for each of the three components before anything is re-marked, with observable descriptors: for the record, a stated minimum number of labelled diagrams and completed dated entries; for the project, a stated list of acceptable topics and a requirement of a methods section and a results section; for the viva, three questions with a mark scheme.
Second, re-mark blind. Remove names, shuffle the 34 records, and mark each component against the descriptors. Where possible have a second Biology teacher mark a sample of ten independently and compare; a difference of more than 2 marks in a component triggers a joint re-mark of the whole set.
Third, allow a short remedial window: students whose records are deficient may complete missing diagrams or entries, with the date of completion noted, and be marked on the completed work. This is the legitimate way to ensure 'no child suffers'.
Fourth, keep a moderation sheet: original mark, moderated mark, component-wise reason, teacher signature and date, plus the sample re-marked by the second teacher. A documented process protects both the students and the teacher if the board calls for the records.
3. The two matching project reports must be resolved. What is a fair and defensible outcome, and what evidence must be recorded?
Read the worked answer
Identical paragraphs in identical font establish that the text is shared; they do not establish who wrote it or whether both students understood it. The two cases must therefore be handled separately, and on evidence of process rather than on the text alone.
For the student with all eight sessions attended and dated rough notes showing drafts, a plausible reading is that the other student copied from them. That student should not lose marks for someone else's copying; their project should be assessed on its merits, with the matching text noted in the moderation sheet as not their fault.
For the student with three attendances and no drafting trail, the project cannot be certified as their own work. The fair route is not an automatic zero: hold a short viva on the project's method and findings. If the student can explain and defend the work, award marks for the project at a reduced level reflecting the missing process evidence; if not, award marks only for whatever supervised work exists, and record the outcome with reasons.
Evidence to file: copies of both reports, the attendance extract, the rough notebooks with dates, the viva questions asked and the student's responses, and the final decision with the teacher's signature. Both students should be told the outcome and the reason, so that the process is transparent rather than punitive.