Did you know that adding two judges’ marks can give the judge with the bigger score sheet more influence, even when the organisers intended both judges to count equally? In a school debate, one judge might score out of 20 and another out of 50. A hurried organiser adds the marks and thinks the team with the larger total must have performed better. That conclusion depends on a hidden weighting choice.
A maximum mark is the largest score available on a particular rubric, and a normalised percentage expresses an earned score relative to that maximum. If judges are meant to contribute equal weight and their rubrics are genuinely comparable, convert each judge’s mark to a percentage, then average those percentages. If the judges deliberately carry different weights, apply the published weights instead. Never choose the method after looking at which team wins.
Why raw marks can answer the wrong question
Imagine one judge uses 20 points and another uses 50. Simply adding marks gives the second judge a greater potential effect on the final score. One raw point on a 20-point scale represents 5% of that scale; one on a 50-point scale represents 2%. Comparing raw totals might be appropriate if the competition intentionally gives 20 and 50 points to differently weighted criteria. It is not automatically appropriate when the organisers say the two judges count equally.
Before computing, ask three questions: Are the judges scoring the same performance? Are the categories and standards comparable? What weighting did the rules specify? A calculation cannot repair an inconsistent rubric, a missing judge, or a biased scoring procedure. The numbers are evidence to examine, not a substitute for agreed judging principles.
Worked example: the rankings reverse
The names, teams and marks below are fictional examples, not real pupil scores. Mira’s debate team receives 18 out of 20 from Judge A and 35 out of 50 from Judge B. Noor’s team receives 12 out of 20 from Judge A and 45 out of 50 from Judge B. Suppose the judging agreement says the two judges have equal weight and their rubrics measure comparable qualities.
Convert Mira’s scores: 18 ÷ 20 × 100 = 90% and 35 ÷ 50 × 100 = 70%. Equal-weight average = (90 + 70) ÷ 2 = 80%. Convert Noor’s scores: 12 ÷ 20 × 100 = 60% and 45 ÷ 50 × 100 = 90%. Equal-weight average = (60 + 90) ÷ 2 = 75%. Under the stated equal-weight method, Mira’s team scores higher by 5 percentage points.
Now inspect the misleading shortcut. Mira’s raw total is 18 + 35 = 53 out of 70, or about 75.7%. Noor’s raw total is 12 + 45 = 57 out of 70, or about 81.4%. That method ranks Noor ahead. Neither calculation has an arithmetic mistake; the methods implement different weights. A fair judging report must state which weighting was authorised and why.
Under the raw-total method, Judge A’s maximum 20 points make up 20/70, about 28.6%, of all possible points, while Judge B’s 50 points make up about 71.4%. Under equal weighting, each judge contributes 50% after conversion. There is no universal rule that each judge must have equal influence. There is a rule of honest reporting: announce the intended influence and use it consistently.
A small diagnostic to find the misunderstanding
Ask a student whether 18/20 is greater or smaller than 35/50. If they answer 35/50 because 35 is larger than 18, they are comparing numerators without considering the denominator. Ask them to imagine a score of 9/10 compared with 35/50. Each conversion gives 90% versus 70%; the magnitude of the written number is not a reliable guide without the maximum.
Next ask why taking an average of 18 and 35 is not a fair equal-judge score. The learner should explain that the two marks are expressed on different scales. Then give a sheet where both judges score out of 20. In that special case, averaging raw marks and averaging converted percentages would preserve ranking because both are on the same scale.
A clear procedure for educators and student organisers
- Read the published competition rules and identify whether judges or scoring criteria have stated weights.
- Check that every reported mark includes its maximum, and flag missing or invalid entries rather than silently assuming zero.
- Confirm that judges were given compatible criteria, anchor descriptions and a consistent understanding of what each level means.
- For equal judge influence, calculate each judge’s percentage as earned ÷ maximum × 100.
- Average these percentages using the number of valid equally weighted judges; report missing scores separately and follow the competition’s rule for them.
- For declared unequal weights, multiply each converted percentage by its specified weight and sum the contributions.
- Recalculate one team by hand, check decimal and rounding rules, and report the method beside the outcome.
Suppose the organisers instead decide that Judge A carries 40% and Judge B carries 60%. Mira’s weighted result becomes 0.40 × 90 + 0.60 × 70 = 78%. Noor’s becomes 0.40 × 60 + 0.60 × 90 = 78%. The example produces a tie under those weights. This is not an argument for picking that weighting; it is an illustration of why the decision belongs in the rules before adjudication.
What percentage conversion cannot fix
Suppose Judge A assessed clarity and structure while Judge B assessed factual accuracy, but their rubrics were developed with different purposes and definitions. Converting to percentages does not make the constructs equivalent. An 80% on ‘evidence quality’ is not necessarily interchangeable with an 80% on ‘delivery’. The organiser should first clarify the learning goal and define the rubric dimensions and their explicit contributions.
Similarly, judges may be more or less strict even when they share a scale. A conversion does not eliminate differences in expectations, training or observation. The strongest quality check is an agreed rubric, sample performances discussed before judging, and a transparent process for reviewing errors. Where a scoring sheet contains an obvious addition mistake, correct it under the established rules and retain an audit record.
Practice from guided calculation to independent explanation
Guided: two scales, same performance
A student receives 8/10 from one judge and 32/40 from another. Both are 80%. With equal weights the mean remains 80%. Ask why adding 8 + 32 = 40 is not a score out of 50 that reveals a different performance level; it also gives 80% here, but only because the two scale percentages happen to coincide.
Partner: invent a case where ranking changes
Pairs invent scores for two teams on different maxima. They must calculate an equal-weight average and a raw-total percentage, then explain whether the ranking changes. Partners exchange papers, check every fraction and give a one-sentence explanation of the weighting embedded in each method. One partner should argue for equal weight; the other should identify a legitimate reason why the rules might prescribe unequal weights.
Independent: make the correct decision before seeing marks
A hypothetical school contest states Judge A’s 30-point rubric contributes 25% and Judge B’s 40-point rubric contributes 75%. Team C receives 24/30 and 30/40. Convert: 24/30 = 80%, 30/40 = 75%. Weighted final = 0.25 × 80 + 0.75 × 75 = 76.25%. Check independently: 20 + 56.25 = 76.25. A plausible wrong method gives (24 + 30)/70 ≈ 77.14%, which answers a differently weighted question.
How to report progress without promising outcomes
Learner progress appears when a child labels every numerator and denominator, names the actual weighting, can recreate the result independently and is willing to question a rule that was not specified. Teachers can include one ‘choose the method’ question before any arithmetic. Parents can ask ‘What does each mark measure?’ rather than merely ‘Who won?’. If a judge’s notes explain a team’s next improvement, that feedback may be more useful educationally than a narrow ranking difference.
This is connected to comparing tests with unequal word-list difficulty and to holding a learning goal steady when response formats differ. The wider eduKate Education Hub encourages learning that transfers between contexts. The Secondary 1 Mathematics tutorial reference illustrates the same principle of understanding a calculation before rushing toward an answer.
Questions about scoring fairness
Can we always average percentages?
No. That is suitable only when equal judge weight is the stated rule and the scales are comparable enough for that use. Published criteria may require other weights or separate component thresholds. Follow the actual rule and explain it.
What if a judge has not submitted a score?
Do not silently record zero. A missing assessment is not the same as an earned zero. Flag it, check contest rules and follow the authorised handling procedure. If there is no agreed procedure, organisers should resolve it transparently before announcing results.
Should we round each judge’s percentage first?
Prefer retaining sufficient precision for intermediate steps, then round the final reported result according to the declared rule. Premature rounding can create artificial ties or change close rankings.
Does a higher result mean the team learned more?
Not necessarily. Scores reflect performance against a particular rubric at one occasion. Learning is better evidenced by explanation, feedback, improvement in comparable tasks and independent transfer, not one ranking alone.
The real educational result
The core lesson is that numbers always travel with assumptions. Dividing a mark by its maximum is easy. Knowing whether to do so, what to weight and what the outcome can actually justify is the intellectual work. Students who learn to ask these questions become fairer judges of data and of their own progress.