Did you know that a single empty box on a marking rubric can make an apparently precise total unfair? A student may receive points for reasoning, evidence and accuracy, while the clarity row is left blank. If someone converts that blank into zero or quietly removes the row, the arithmetic produces two very different grades, neither of which may be justified.
A rubric is a scoring guide with named criteria and defined performance levels. An unrated criterion means there is no recorded judgement for that part of the work; it does not automatically mean zero performance or not applicable. The immediate repair is to mark the review incomplete, identify why the row is unrated, and have an authorised marker apply the stated assessment policy before calculating a final score.
Three labels that must never be merged
A zero is an actual score: the marker assessed the criterion and, under the rubric’s rules, awarded no points. Not rated means the marker has not made or recorded a judgement, perhaps because the evidence was overlooked or the marker lacked access to it. Not applicable means a criterion truly does not apply to that assessment under an authorised rule. A fourth possibility is missing student evidence, such as a required page never submitted; that situation must be evaluated under the task’s stated rules rather than automatically treated as a missing marking action.
All four cases may produce an empty-looking spreadsheet cell. That does not make them equivalent. A defensible score depends on the meaning of the cell, who may decide it and the scoring formula that students were told would be used. This is as much an evidence and fairness problem as an arithmetic problem.
Worked example: the fictional four-criterion project
Imagine a completely fictional class project with four equal criteria, each marked from zero to four points: reasoning, evidence, clarity and accuracy. Suppose the recorded marks are reasoning 3/4, evidence 4/4, clarity not rated, and accuracy 2/4. Assume the published rubric requires all four criteria for a final mark and provides no rule to waive clarity. The student’s work is available; the reviewer simply omitted that box.
The sum of known points is 3 + 4 + 2 = 9. The known maximum across those three scored rows is 4 + 4 + 4 = 12. A report can truthfully say, ‘The three assessed criteria total 9/12, and clarity is unassessed.’ That corresponds to 75% of assessed points only. It must not silently be labelled the student’s final percentage.
If somebody puts a zero in the missing box, the apparent total becomes 9/16 = 56.25%. That falsely asserts that clarity was actually assessed at zero. If somebody drops the box and reports 9/12 = 75% as the official grade, the person has effectively changed the weight of all three scored criteria. Neither action is permitted by the hypothetical rubric’s stated policy.
We can calculate a mathematical range, not a grade: if the missing clarity mark could be any whole score from 0 to 4, the overall total would lie from 9/16 (56.25%) to 13/16 (81.25%). That range reveals why the missing judgement matters. It does not tell us the true clarity mark or authorise replacing professional assessment with a midpoint.
What the reviewer should do, in order
- Check the rubric as published. Confirm each criterion, its scale and weight, any explicit N/A rules and who may assign scores.
- Inspect the work, not just the spreadsheet. Determine whether evidence for the unmarked criterion exists.
- Label the reason accurately. Use a distinct code such as ‘not yet assessed’, ‘evidence missing’ or ‘not applicable’, with a documented definition.
- Obtain an authorised review. Ask the original marker or designated moderator to judge the criterion according to the same descriptors used for classmates.
- Recalculate only when a valid score or formal exemption exists. Preserve the assessment history and avoid silently changing the earlier marks.
- Tell the learner what comes next. Provide clear next steps and learning feedback rather than a misleadingly exact provisional percentage.
How an actual zero differs from an absent score
Suppose the student included a clarity section and the marker used the rubric descriptor to conclude it met the zero-point level. Then 0/4 is an assessed decision and may legitimately be included, subject to the task’s policy. The learner should be able to see what evidence was considered, what the rubric required and how to improve. If the student never submitted a required section, the rubric or assignment policy may specify a zero or another consequence. The marker must follow that policy rather than guess one on the spot.
Conversely, a criterion can be genuinely inapplicable. For example, a peer-feedback activity might expressly waive a speaking criterion for written-only projects and specify how weights are redistributed. That exception must be decided in the rubric design or according to an approved procedure, not invented after looking at an individual’s scores. Otherwise pupils completing different tasks can receive misleadingly comparable percentages.
A short diagnostic for learners and markers
Show a student the four-row example and ask, ‘Is the blank a zero?’ If they say yes because Excel displays blank cells as zero in a sum, they are confusing a software calculation with a human judgement. Ask, ‘Is 9/12 a final result?’ If they say yes without reading the four-criterion rubric, they may be confusing observed coverage with the stated total. Ask, ‘Can you choose 2/4 for clarity because it is halfway between zero and four?’ If they say yes, they are inventing an assessment.
For a teacher, the diagnostic is operational: can the reviewer trace every recorded value back to a criterion, observation and descriptor? Can another authorised reviewer understand why any cell is missing? If the answer is no, repair the record before reporting. A beautiful spreadsheet formula is not a substitute for traceable marking.
Practice ladder: separate arithmetic from marking decisions
Guided: record the known information
Return to the fictional 3, 4, blank, 2 example. Have the learner write ‘known assessed points: 9 out of 12 assessed; final result pending a fourth judgement’. Have them find the possible total from 9 to 13 out of 16, clearly marked possible range only.
Pair practice: a different number of criteria
Another invented rubric has four criteria, each out of five. Scores are 2, 3, 4 and not rated. Known assessed points: 9 out of 15. The possible complete total is from 9/20 = 45% to 14/20 = 70% if the missing criterion could score 0–5. The correct administrative conclusion, under a complete-rubric requirement, is still ‘assessment incomplete’, not a guessed percentage.
Independent: distinguish not applicable from not rated
Invent two short scenarios. In the first, the marker accidentally skipped an ‘evidence’ row, even though the evidence exists: request assessment. In the second, the task instructions expressly remove an oral-presentation criterion for a written-only alternative and prescribe the remaining weights: follow the approved alternative. Explain in one sentence why the same blank symbol should never be used for both without explanation.
Transfer: audit an anonymous class spreadsheet
Before computing class averages, code each criterion as ‘scored 0’, ‘scored positive’, ‘not assessed’ or ‘formally not applicable’. Verify the count and reason for every nonnumeric entry and that any approved reweighting was applied consistently. If scores are temporarily missing, report the number of complete records rather than allowing a default formula to hide them.
Responsibilities for a practical repair
A learner can politely ask which rubric row was marked and how the recorded grade was calculated. A parent can ask for the task’s actual scoring policy and encourage the child to use the feedback rather than argue from an unexplained percentage alone. Teachers can make ‘not rated’ visually different from zero, include a completion check before submitting grades and record short evidence-based justifications. A moderator can sample inconsistent cases and calibrate marker interpretations, protecting both fairness and learning.
Observable progress means students can correctly label missing information, compute only what is known, identify what further decision is required and explain why imputation changes the meaning. Teachers can measure improvement by a lower rate of unreviewed rubric gaps and by consistent handling of genuinely identical cases—not by inflating marks or guaranteeing scores.
Limitations and boundaries
Schools, examinations and institutions have different grading, moderation, absence and submission policies. This article is a reasoning example, not a statement of any particular Singapore examination rule. In high-stakes assessment, the applicable official procedure and confidentiality requirements control the decision. Do not copy peer marks or disclose other pupils’ work as a way to settle a dispute.
A rubric may use unequal weights, qualitative levels rather than numeric points, or a minimum required competence in one criterion. In those cases, the simple arithmetic here cannot be reused without adapting the formal scoring rule. Software may also encode a blank, zero and ‘not applicable’ differently. Check the system before using its calculated totals.
Frequently asked questions
Can the teacher temporarily report a score based on only three criteria?
They may describe the assessed portion explicitly for formative feedback, if local policy permits, but must not present it as the final four-criterion grade without an authorised scoring rule.
Should an absent assignment section always receive zero?
No universal answer applies. The task instructions and school policy govern how unsubmitted work is treated. Do not confuse missing student evidence with a marker’s failure to complete a review.
What if the teacher truly cannot assess one criterion?
Document why, consult the authorised moderation or assessment procedure, and state any permitted alternative transparently. Do not fill the cell with a guessed value to make a spreadsheet complete.
How do I improve after a missing or low score?
Seek criterion-specific feedback with observable examples and practise the skill, such as presenting an evidence paragraph more clearly. The goal is better independent performance, not simply a filled-in cell.
Relevant reading and source notes
Two verified companions examine rubrics that accidentally score teamwork twice and calibrating peer reviewers who score differently. The eduKate Education Hub provides connected learner and parent routes; the Clementi mathematics reference models diagnosis and checked progression. As an example of an institution emphasising adherence to rubrics and clear justifications, see the National University of Singapore interdisciplinary science grading instructions. Its rules apply to its own context, not automatically to any school rubric.
The education aim behind the arithmetic
Fair grading is not achieved by making every cell contain a number. It is achieved by keeping the evidence, the decision and the scoring rule distinct. When a criterion has not been rated, say so; then complete the assessment by the authorised method. The learner benefits from a result that explains what they demonstrated and what they should practise next.