Moderating Inconsistent Marks in a Biology Department
A secondary school Biology department has seven teachers and 210 Form 4 students. The end-of-term test was common, but teachers marked their own classes using different conventions. Teacher A's class mean is 73; Teacher B's class mean is 51. Teacher B's students include 18 students within two marks of the promotion cutoff of 50. A parent has appealed, claiming that Teacher A 'gives marks easily'. The head of department has five working days before results go to the school council and must decide whether to adjust marks, re-mark scripts, or defend the existing results. Two teachers are already hostile to any moderation, saying it questions their professionalism.
1. What evidence would you collect before deciding whether the two sets of marks are comparable? Explain what each piece of evidence can and cannot show.
Read the worked answer
Before adjusting anything, the head of department needs evidence about marking, not just about means.
1. Mark distributions, not only means. Compare the spread, median, and pass rate for each class. A mean difference of 22 points could reflect teaching, class composition, or marking. Distribution alone cannot tell which.
2. Item-level scores for common questions. For each item, compare the proportion correct in Teacher A's and Teacher B's classes. If Teacher B's class performed similarly on objective items but lost marks on extended-response items, marking convention is a likely cause. If Teacher B's class also scored lower on objective items, the difference may be partly real.
3. A sample of scripts re-marked blind. Select scripts from both classes, remove names, and have two teachers independently mark the same five or six scripts against the official mark scheme. This can show scoring drift on specific criteria. It cannot prove that one teacher is incompetent.
4. The mark schemes and any annotations. Check whether teachers used the same acceptable answers, the same tolerance for units, and the same penalties. A difference in conventions is a system problem, not necessarily a professionalism problem.
5. Class composition and attendance. Compare prior attainment and absence rates. This cannot excuse unfair marking, but it prevents a false accusation if Teacher B's class genuinely started lower.
2. Design a five-day moderation process that is fair to students and defensible to hostile teachers.
Read the worked answer
Day 1 – Establish the facts. The head of department collects mark sheets, mark schemes, and a random sample of eight scripts from each teacher, stratified across grade bands. Names are removed and scripts are coded.
Day 2 – Blind re-mark. Two teachers who did not mark the sampled scripts re-mark them against the official scheme. The head of department also marks them. Discrepancies are recorded by criterion, not by teacher name.
Day 3 – Moderation meeting. Present the evidence as a department problem: 'On the extended-response item, we applied the scheme differently.' Avoid naming Teacher A as generous or Teacher B as harsh. Agree on anchor scripts: one high, one middle, one borderline. Re-mark any script within three marks of the cutoff, especially the 18 borderline students in Teacher B's class.
Day 4 – Decide the adjustment rule. Prefer re-marking to statistical scaling. If scaling is unavoidable, apply it only where evidence shows a systematic criterion difference, and record the reason. Do not add the same number of marks to every student, because that distorts the ranking.
Day 5 – Report and communicate. Produce a short moderation note: sample size, criteria reviewed, decisions, and any changes. Send it to the school council. This gives hostile teachers a procedural defence: the issue is consistency of marking, not their integrity.
3. Draft the core message to the appealing parent and to the department. How do you explain moderation without admitting arbitrary mark inflation?
Read the worked answer
To the parent: 'We have reviewed the end-of-term Biology results because several classes were marked by different teachers. We did not simply raise or lower marks. We re-marked a sample of scripts against the same official mark scheme and checked every script close to the promotion cutoff. Your child's script was included in that review. The outcome is based on the written answers and the published criteria, not on which teacher taught the class. If your child's mark changes, you will receive the revised mark and a short written reason. If it does not change, the original mark stands because the script met the criteria as marked.'
To the department: 'The review is not an accusation. It is a check on whether our common test was marked with common standards. The sample shows that we interpreted the extended-response criteria differently, especially on units and on partial credit. We will agree anchor scripts and re-mark the borderline scripts. This protects all of us: a parent cannot credibly claim that one teacher is generous if we can show a documented, criterion-based process. It also protects students from being penalised for a marking convention they never saw.'
The message should avoid saying 'marks were inflated'. It should say that comparability was not guaranteed and has now been checked. That is accurate, defensible, and does not require the head of department to pretend that a 22-point mean difference is automatically evidence of cheating.