EDUKATE SINGAPORE · GOLD STANDARD PRACTICAL CASE STUDY 218
100% satisfied among nine respondents is not 100% of all families
Nine of 100 eligible invited families completed a fictional school satisfaction survey, and all nine were positive. The reported satisfaction among respondents is 9/9, or 100%. The response share is only 9/100, or 9%; it does not establish the opinions of the other 91 families.
All families, responses and schools in this article are hypothetical. The featured family photograph is contextual and does not portray survey participants.
The counts readers need
- 100 eligible families invited; nine complete questionnaires returned.
- 9/9 respondents reported satisfaction: 100% among recorded respondents.
- 9/100 eligible families responded: a 9% simple completion share under these assumptions.
- 91 eligible families have no recorded satisfaction answer in this survey.
- No valid conclusion about all 100 families’ satisfaction follows from the nine answers.
Chapter navigation
- 1. Nine happy families become a headline for a hundred
- 2. Two denominators tell two different stories
- 3. The American Association for Public Opinion Research provides the reference
- 4. A low response share creates uncertainty, not a known negative verdict
- 5. Why families who respond may differ from those who do not
- 6. The question and invitation channel also shape the answer
- 7. Why a conventional margin of error is not available automatically
- 8. A useful management note can retain the encouraging feedback
- 9. A better follow-up survey improves the opportunity to respond
- 10. Classroom exercise: write the denominator under every percentage
- 11. The closure decision should separate satisfaction from survey quality
1. Nine happy families become a headline for a hundred
A fictional school sends a feedback questionnaire to exactly 100 eligible families. Nine families return complete responses. All nine select ‘satisfied’ on the question about recent learning support. The office newsletter announces ‘100% of parents are satisfied’ and displays a full green circle without a respondent count.
The underlying answer count is true: nine favourable answers out of nine completed questionnaires equals 100% among respondents. The newsletter changes the population being described. It presents a result from nine self-selected responses as if all 100 invited families had expressed the same view.
Nothing in this fictional case establishes that the remaining 91 families are unhappy. They might be equally satisfied, busy, unaware of the survey or reluctant to respond. The key finding is that their views were not measured by the completed questionnaires.
The school, families, survey, questions and figures are invented. The related family photograph is editorial context rather than a depiction of any respondent or a real school’s consultation.
2. Two denominators tell two different stories
The satisfaction count among completed respondents is nine satisfied divided by nine completed surveys, giving 100%. The simple completion proportion from the verified list of 100 eligible invited families is nine completed divided by 100 invited, giving 9%. These are different measures with different denominators.
The 100% says every recorded respondent selected that answer. It does not say every invited family answered. The 9% says how many of those eligible invitations produced a completed questionnaire under the simplified case assumptions. It does not mean only nine percent of families are satisfied.
A responsible result can report both measurements together: ‘Nine of 100 invited families responded (9%), and all nine respondents reported satisfaction (9/9).’ Readers can evaluate the small response base immediately without having to search the footnotes.
Survey response rates in professional research use defined outcome and eligibility categories; an organisation should follow an appropriate standard when computing formal survey rates. The simple case calculation is transparent because all 100 families are stipulated eligible and the response count is known.
3. The American Association for Public Opinion Research provides the reference
The American Association for Public Opinion Research response-rates overview explains why response, cooperation and completion measures are not interchangeable. It sets out the value of standard definitions and stresses that a rate on its own does not establish survey accuracy.
AAPOR’s Standard Definitions resource addresses the final disposition of sample cases and the calculation of survey outcome rates. Its Best Practices for Survey Research asks researchers to explain who was studied, sample construction, questions, mode, weighting and relevant uncertainties.
These resources concern survey quality; they do not certify the fictional school. They show why simply presenting a favourable percentage is insufficient for readers who want to understand what was asked, who responded and how the school reached its claim.
A responsible educational questionnaire need not be a national research project, but it still benefits from accurate denominators, transparent wording and respect for the people whose views are missing.
4. A low response share creates uncertainty, not a known negative verdict
The 91 unobserved family opinions matter because they are not documented. It is possible that every one is satisfied, in which case all 100 could be satisfied. It is also possible that none of the 91 shares the nine respondents’ satisfaction, in which case only nine of the hundred are known to be satisfied under the invented conditions.
Those possibilities are deliberately extreme teaching scenarios, not estimates. The nine positive replies establish nine positive responses, but the questionnaire by itself cannot determine which picture is nearer to reality. The interval between nine and 100 is not a statistical confidence interval; it is a logical illustration of what remains unconstrained without more information.
A manager who insists the actual figure must be around 100% may be ignoring selection effects. A critic who claims the school must have a 91% dissatisfaction problem makes the opposite unsupported leap. Neither side has evidence of the missing opinions.
The Gold Standard represents unknown information as unknown and plans a responsible method of learning more. It does not turn absent records into praise or blame.
5. Why families who respond may differ from those who do not
Families with a strong positive experience may be more motivated to leave feedback. Families who are concerned or disappointed may also be more likely to answer, depending on the subject, trust and survey channel. Some recipients never see the invitation, while others may not have time, language access or a convenient device.
The risk is nonresponse bias: the respondents’ distribution of views may differ systematically from that of the broader target population. A low response rate can be a warning that motivates further checks; it does not by itself quantify the size or direction of bias.
AAPOR cautions that response rate alone is not a reliable arbiter of survey accuracy. Better work combines the participation record with information about the sampling method, known population characteristics and possible differences between respondents and nonrespondents.
The school’s next task is therefore not to guess why the 91 did not answer. It is to improve a respectful and accessible feedback process and report what is actually known.
6. The question and invitation channel also shape the answer
Suppose the survey asks ‘How satisfied are you with our excellent extra support?’ That leading wording would create a different interpretive problem from a neutral question with balanced options. The school needs the actual question text, response categories and instructions to judge what the nine favourable selections mean.
The mode matters too. A message sent to a family chat group may miss caregivers who are not in that group. A form that requires sign-in to a particular service may exclude people with limited technical access. Providing translations or reasonable non-digital options may help where appropriate, but any collection should protect privacy and avoid pressuring families.
Timing can influence who answers. A survey issued during examinations or a holiday may get very different responses from one offered after an ordinary school week. Comparisons between periods should disclose changes in question wording, contact mode and respondent eligibility.
The goal is not to manipulate a higher satisfaction result. It is to reduce avoidable missing voices and present useful feedback without confusing an invitation campaign with a representative assessment.
7. Why a conventional margin of error is not available automatically
A presenter may try to repair the headline by attaching ‘±3%’ beneath the 100% satisfaction figure. The fictional survey does not contain the design details required to justify such a conventional sampling-error margin. It has nine self-selected respondents from a defined invitation list, not a demonstrated probability sample with a known selection mechanism.
Even a correctly calculated margin of sampling error would not automatically account for nonresponse bias, leading questions, recording errors or coverage gaps. A sophisticated-looking interval cannot rescue an uncontrolled recruitment process.
The school’s first responsibility is to describe the sample honestly. ‘Nine responses from 100 invited eligible families’ is more meaningful than an invented confidence interval. For a more defensible population estimate, the team should choose an appropriate design and, where warranted, obtain survey-methods expertise.
This is distinct from Case 209’s confusion between standard deviation and confidence interval. Here the central obstacle is whether the observed respondents provide evidence about families whose opinions were not recorded.
8. A useful management note can retain the encouraging feedback
Staff need not hide the fact that nine families expressed satisfaction. Those respondents provided real feedback under the fictional assumptions, and their comments may suggest aspects of teaching support that deserve recognition or improvement.
The management note should say: ‘The pilot feedback form received nine complete answers from 100 eligible invited families. All nine respondents marked satisfied. The low response share means the result should not be reported as the satisfaction rate of all families.’
If the nine answers include useful open-text suggestions, the team can summarise themes carefully without identifying families unnecessarily. It must not portray a small set of comments as a balanced cross-section of everyone.
An accurate report can celebrate positives while acknowledging that evidence is too limited for a school-wide satisfaction claim. Openness protects the school from building improvement plans around an unsupported sense of certainty.
9. A better follow-up survey improves the opportunity to respond
The school can review invitation delivery, reminder schedule, accessibility, neutral question wording, respondent privacy and whether its roster is up to date. It can provide reasonable opportunities to respond without repeated pressure or public identification of nonrespondents.
Where appropriate, a small number of follow-up contacts may improve understanding of nonresponse. Any follow-up must be lawful, proportionate and respectful of families’ preferences. The school should not misrepresent feedback as mandatory or use grades or student treatment to compel a response.
After new data arrive, the team should keep survey rounds distinct and disclose whether question text and recruitment changed. A later survey with more responses may produce a more useful picture, but a higher response percentage alone still cannot guarantee unbiased results.
The Gold Standard requires a process that can support the particular decision: improving communications, helping learners and learning whether families’ experiences differ. It is not simply a contest to obtain the happiest headline.
10. Classroom exercise: write the denominator under every percentage
Give pupils a chart with 100 small squares for the invited families and colour nine squares green for the satisfied respondents. Leave the other 91 squares uncoloured rather than red. Ask learners what 9/9, 9/100 and 91/100 mean.
They should find that all completed respondents were satisfied, only nine percent of the eligible invited population completed the questionnaire and 91 percent of that population has no recorded satisfaction answer in this survey. No student should infer that an uncoloured square is an unhappy family.
For an advanced extension, invent two hypothetical nonrespondent groups with different distributions of satisfaction. Show how the school-wide proportion could change while the observed 9/9 remains constant. That demonstrates why a respondent statistic does not necessarily transfer to the population.
Finish with a one-sentence newsletter correction. A successful answer includes both the favourable response count and the low participation base. This practice applies to customer reviews, workplace polls and community consultations as much as school surveys.
11. The closure decision should separate satisfaction from survey quality
The fictional manager should withdraw the unqualified ‘100% parent satisfaction’ headline and publish the actual counts. The survey administration can be reported as nine completed responses from the 100 eligible invitations, with the result limited to those nine respondents.
Before acting as though the school has measured everyone’s satisfaction, staff should improve the evidence-gathering design and report the unanswered population plainly. No remedial action should be based on an invented assumption that the 91 missing responses are favourable or unfavourable.
The source questionnaire, invitation list, response logs, question text and calculation should remain available to authorised reviewers, with personal data protected. A later reader must be able to reconstruct how each percentage was obtained.
The Gold Standard is not an impressive circle chart. It is a correct statement of whose views have actually been heard and a fair plan to hear more.
Survey-reporting checklist
- State the target population and verified number invited.
- Show respondents and completed questionnaires as counts, not only percentages.
- Preserve question wording, mode and recruitment details.
- Investigate nonresponse and possible selection effects without assuming their direction.
- Avoid conventional margin-of-error claims when design assumptions do not support them.
- Publish a correction when respondent satisfaction is wrongly labelled population satisfaction.
Frequently asked questions
Does nine satisfied responses mean every parent is satisfied?
No. It means every respondent in this nine-response set selected satisfied. The other 91 eligible invited families did not provide a recorded answer.
Is the response share 9%?
In this simplified scenario, yes: nine complete responses from 100 stipulated eligible invited families. Formal response rates should use appropriate disposition definitions.
Does a 9% response share mean the survey is definitely biased?
No. It creates a reason to assess potential nonresponse bias; its existence, direction and size require further evidence.
Can the school attach a three-point margin of error?
Not without a suitable sampling design and warranted analysis. A conventional margin does not automatically address self-selection or nonresponse.
What can the school truthfully publish now?
It can say that nine of 100 eligible invited families responded, and all nine respondents reported satisfaction.
Authoritative references and connected cases
- AAPOR · Response Rates Overview
- AAPOR · Standard Definitions for Survey Outcomes
- AAPOR · Best Practices for Survey Research
- The Gold Standard of Statistics
- The Gold Standard of Education
- Gold Standard Case 217 · P-value misinterpreted as truth probability
- Gold Standard Case 209 · Standard deviation versus confidence interval
- Gold Standard · Core Requirements and Evidence
- Gold Standard · Practical Casebook
AAPOR provides methodological guidance, not an audit of the invented school. Every numerical survey example is fictional and stipulated for teaching. The Gold Standard principle is to report what the evidence actually covers.