Did you know? A class report can contain two percentages that both look like 65%, yet one says how many pupils answered and the other says how many of those respondents supported an idea. If a school newsletter swaps those meanings, it may turn a small collection of opinions into a misleading statement about the whole class.
The response rate is the percentage of invited, eligible people who completed the survey; the approval or support rate is the percentage of a clearly stated group who selected a supportive answer. For a simple voluntary classroom poll, count completed responses against invitations for the first rate, then count ‘yes’ answers against the declared response denominator for the second. Never write ‘65% of all pupils agreed’ unless that is actually the fraction you measured.
Every number and school project in this guide is a fictional teaching example, not a real poll, election, student record or claimed outcome. The simplified invitation count is for a known closed class list; professional surveys require more detailed rules for eligibility, partial responses and sampling.
One fictional class, three very different percentages
A school committee sends a voluntary, single-question proposal survey to 120 eligible pupils. Exactly 78 submit usable, complete responses; the remaining 42 do not respond. Among the 78 respondents, 51 choose ‘support’, 19 choose ‘do not support’, and 8 choose ‘not sure’. Check the count first: 51 + 19 + 8 = 78. Assume each pupil can submit only one response and the invitation list and answer categories do not change.
The response rate is 78 ÷ 120 × 100% = 65%. This tells us how much of the invited eligible group responded. It says nothing yet about which proposal answer they gave. A respondent can support, oppose or be unsure and still count as a respondent.
The support rate among respondents, including the ‘not sure’ group in the denominator, is 51 ÷ 78 × 100% ≈ 65.4%. This is quite close to 65%, which makes it easy to confuse the two. However, the first numerator was 78; the second numerator was 51. The first denominator was 120; the second was 78. The two statements have different meanings even when the results look similar.
A third legitimate calculation is 51 ÷ 120 × 100% = 42.5%: confirmed support among all those invited. It is not an estimate that only 42.5% of the full class would support the proposal if everyone answered; the missing 42 pupils’ opinions were not observed. It means precisely that 51 of the 120 invitees were recorded as saying yes. Keep the distinction between a known count and an inference about unobserved views.
What happens if someone excludes ‘not sure’?
If a report asks what share of pupils taking a yes-or-no position voted yes, the denominator changes again: 51 + 19 = 70. The result is 51 ÷ 70 × 100% ≈ 72.9%. That can be a meaningful statistic, but only with a label such as ‘72.9% of pupils who chose support or oppose’. It is not the share of everyone who replied, and certainly not the share of all 120 invited pupils.
A headline saying ‘nearly three quarters of pupils support the idea’ would hide the exclusion of eight undecided replies and 42 nonresponses. A more defensible report would say: ‘Of 120 invited pupils, 78 replied (65%); 51 of those 78 respondents (about 65.4%) supported the idea.’ That sentence conveys both reach and sentiment without inventing a majority among the people who did not answer.
The denominator is the real reading skill
A fraction is not defined by its numerator alone. In the phrase ‘percent of whom?’, the word after of chooses the denominator. ‘Of those invited’ uses 120. ‘Of those who replied’ uses 78. ‘Of those expressing support or opposition’ uses 70. The same 51 supportive answers can give three different percentages because the questions are different. The calculations are not contradictory.
Pew Research Center’s discussion of survey response rates and nonresponse makes an important broader point: response rate by itself does not fully tell us how representative the responses are; differences between respondents and nonrespondents can affect conclusions. A classroom poll should therefore report the actual group and avoid presenting a voluntary sample as everyone’s opinion.
A small diagnostic that catches the common mistake
Ask the learner: ‘If 78 of 120 pupils complete a survey, and 51 support an idea, what percentage is 78 ÷ 120?’ An answer of ‘65% approval’ reveals a label error even though the division is right. Then ask what 51 ÷ 78 measures. If the pupil says ’51 out of 120′, the problem is the denominator, not mental arithmetic.
One further diagnostic: show a sentence ‘73% agree’ and ask whether it can be evaluated without knowing who was counted. A careful pupil should say no. They need to know whether neutral answers, nonrespondents, duplicates or ineligible respondents were removed, and how many people were in each group.
From a teacher’s worked model to independent reporting
Stage 1: make a count ledger before any percentage
Draw four boxes: invited = 120, completed = 78, support = 51, oppose = 19, not sure = 8. Check the answer boxes sum to completed responses. Never force the totals by changing real responses; if they do not add up, investigate data entry or missing answers and describe any exclusions.
Stage 2: match each question to its population
For ‘What share of invitees replied?’, use 78/120. For ‘What share of respondents supported?’, use 51/78. For ‘What share of yes-or-no respondents chose yes?’, use 51/70. Ask the learner to write the denominator in words before writing any fraction.
Stage 3: write a report whose words match the computation
Require two clauses: ‘X people were invited and Y replied’; ‘Among those who replied, Z selected support.’ Add the corresponding percentages with ‘about’ when rounded. Test the sentence by having a second pupil reconstruct both fractions from the prose alone. If they cannot, the labels remain ambiguous.
Stage 4: a different dataset, solved independently
For a new fictional student-council poll, 30 pupils were invited; 24 replied, including 15 yes, 7 no and 2 not sure. The response rate is 24/30 = 80%. Support among all respondents is 15/24 = 62.5%. Confirmed support among all invited is 15/30 = 50%. A yes-only share among those choosing yes or no is 15/22 ≈ 68.2%. Ask the pupil to state all four denominators, not just calculate four numbers.
How much can the unanswered invitations change the story?
In the original 120-pupil example, 42 did not respond. If none of those 42 would have supported the proposal, recorded support could remain 51/120 = 42.5% of the total invitees. If all 42 would have supported it and their views were later collected without anyone else changing an answer, support could become (51 + 42)/120 = 77.5% of invitees. This is an illustrative logical range for the missing responses, not a confidence interval or a prediction.
Even the 65.4% respondent approval statistic cannot by itself tell us which extreme is likely. Nonrespondents might have different views, and the survey’s wording might influence answers. Do not claim the estimate is nationally representative or that a low response rate automatically proves bias. The honest next step is to examine the invitation method, missingness, question wording and intended decision.
Parents, teachers and pupils can use these safeguards
- Keep the purpose visible: decide whether the poll is for classroom discussion, planning or formal evaluation, and do not claim a stronger purpose than the design supports.
- Use anonymous aggregates where possible: avoid publishing students’ names, identifiers or raw individual opinions; respect consent and school safeguarding practices.
- Show all major counts: invited, responded, each answer category and explicitly excluded records.
- Explain unanswered items: a returned form with an unanswered question is not necessarily the same as a fully completed survey.
- Report rounding honestly: 51/78 is approximately 65.4%, not exactly 65%; choose an appropriate precision and stick to it.
- Invite independent checks: ask a pupil unfamiliar with the project to reproduce each percentage from the stated counts.
Progress is visible when a learner stops saying ‘the survey was 65%’ and begins naming both numerator and denominator. Teachers can compare a first unlabelled report with a second version containing exact group descriptions. Parents can ask which claim is directly observed and which would require more evidence. The goal is careful civic numeracy, not a preferred outcome.
Frequently asked questions
Are response rate and approval rate ever the same number?
Yes, accidentally or by the particular counts. Equality of numerical percentages does not make their definitions identical. Check the numerators and denominators.
Should ‘not sure’ always be excluded?
No. It depends on the stated question. For the share of all respondents who support an option, include ‘not sure’ in the respondent denominator. For the share of declared yes-or-no positions, it may be excluded, but the report must say so.
Can we report a percentage if only a few students answered?
You can report an accurate proportion, accompanied by the actual number of responses and a limitation. Percentages can look impressively precise even when based on very small counts.
Does a 100% response rate guarantee an unbiased survey?
No. Poor wording, pressure to answer, ambiguous choices or recording errors can still mislead. Good reporting and good data collection are separate skills.
The core aim: know who your percentage is about
The Core Aim guide to pilot-testing survey wording addresses what respondents think a question means. A related guide to leading questions shows why the wording also matters. Continue through the Core Aim hub and Education Hub. The Clementi Mathematics reference illustrates diagnosis, explanation, progressive practice and independent checking.
A trustworthy survey report does not begin with a large-looking percentage. It begins with an answerable question, a clearly defined group, and a count anyone can verify. When students can explain who was invited, who responded and who agreed, they have learnt a form of reasoning that carries far beyond school surveys.