Survey Not Applicable vs Missing Answer

Did you know that a blank survey box is not automatically the same answer as ‘No’? It might mean the question did not apply, the student was unsure, the wording was confusing, or the student preferred not to answer. Mixing those possibilities turns tidy numbers into misleading stories.

‘Not applicable’ is an explicit response meaning the situation described by the question does not fit the respondent. It is different from ‘No’, ‘I don’t know’, and an unanswered item. A class survey can offer an appropriate route for inapplicable questions and record missing answers separately; whether to offer each option depends on what the survey is actually trying to measure.

The goal is not to collect as many checkmarks as possible. It is to let people answer truthfully, preserve the limits of the data and avoid claiming that a blank has a meaning nobody supplied.

A practical design check before collecting responses

Define the population and what each question is intended to measure.
Ask whether every respondent can logically answer the question.
If some cannot, consider a screening question or an explicit ‘Not applicable’ choice.
Distinguish ‘I don’t know’ from ‘this situation does not apply’ when both matter.
Allow the survey’s rules on skipping to be clear, and never retrospectively turn unanswered items into ‘No’.
Pilot the wording with a small set of volunteers and revise confusing response options before collecting real data.

Why response options change the measurement

Survey choices influence what respondents can tell you. The Pew Research Center’s questionnaire guidance stresses that closed-ended choices should cover reasonable responses without overlap. A Pew experiment on explicit ‘don’t know’ choices also illustrates that adding such an option can shift the answers people give. These findings are from specific survey contexts; they should not be used as a claim that one option design is universally best for every classroom task.

A good school project tells readers what answer choices were visible, whether the question could be skipped and who was eligible to answer. Otherwise two apparently similar percentages may describe different groups.

Worked example: a school library borrowing question

Invented classroom dataset: Ten students complete an anonymous, voluntary practice form. The question says, ‘Did you use the self-service borrowing station on your most recent library visit?’ A student who has never visited the library during the specified period cannot truthfully choose Yes or No about a most recent visit. The class therefore first asks whether there was a library visit in that period.

Suppose among the 10 fictional students, 6 visited and 4 did not. Of the 6 visitors, 3 say Yes, 2 say No and 1 leaves the borrowing-station question unanswered. The four non-visitors are recorded as Not applicable for that follow-up. These numbers are entirely illustrative; they are not a real survey or observed student outcome.

Report the counts without inventing answers

All respondents: 10.
Eligible for the follow-up (visited the library): 6.
Valid Yes/No answers to the follow-up: 3 + 2 = 5.
Yes among valid Yes/No answers: 3 ÷ 5 × 100 = 60%.
Yes among all eligible visitors: 3 ÷ 6 × 100 = 50%, with one eligible response missing.
Not applicable: 4 non-visitors. Missing follow-up answer: 1 eligible visitor.

The 60% and 50% figures answer different questions. The first describes the fraction of valid Yes/No answers that were Yes. The second counts Yes among all eligible visitors while explicitly retaining the missing answer in the denominator. Neither is automatically ‘the percentage of all students who used the station’. If reporting the entire group, that separate arithmetic would be 3 ÷ 10 = 30%, and the limits must be explained.

What not to write

Do not claim ‘40% answered No’ by converting the four inapplicable records into negative responses. Do not silently exclude the eligible unanswered record while calling the denominator all visitors. Both errors are tempting because they make a table easier to fill, but they change the meaning.

A simple four-way interpretation exercise

Ask pupils to classify these fictional statements: ‘I visited but did not use the station’ is No; ‘I did not visit the library during the period’ is Not applicable; ‘I visited but cannot remember whether I used it’ may be Not sure if that option exists; and ‘I saw the question but left it blank’ is unanswered. There is no evidence that the blank means any of the other three categories.

Discuss the important boundary between ‘not sure’ and ‘not applicable’. The first means uncertainty about an applicable experience; the second means the experience does not fall within the question. If both categories are permitted, define them in language pupils understand. If the survey uses a skip path, preserve the skip information in the raw record.

Progressive practice: design, test, analyse

Stage 1: repair an unfair question

Original: ‘How helpful was the online homework portal yesterday?’ with choices Very helpful, Helpful, Not helpful. A pupil who never opened it yesterday has no truthful option. Add a screening question such as ‘Did you use the portal yesterday?’ and ask the help rating only for Yes responses. If it is a single-item questionnaire, consider an explicit ‘Did not use it’ choice that is analysed separately.

Stage 2: compare two summary tables

Show a fictional table with 12 students, 8 eligible respondents, 6 substantive answers, 2 eligible missing answers and 4 inapplicable answers. Ask which denominators apply. The eligible response completion rate is 6 ÷ 8 = 75%. The 4 inapplicable answers do not belong in its denominator. Clarify what counted as a substantive answer.

Stage 3: build a fresh question independently

Ask learners to write a different question about a resource that some pupils have never used. They must explain who is eligible, what each choice means and how they would record a skipped item. Review the design before collecting any real pupil information.

Ethical and useful classroom safeguards

Do not collect names or personal details unless the task truly requires them and school rules allow it.
Do not pressure classmates to give an answer or disclose a private reason for skipping.
Explain whether a question is optional and what ‘not applicable’ means.
Report exact denominator definitions alongside percentages.
Never infer motives from missing data; describe uncertainty instead.
Have a teacher check questions about sensitive matters before using them.

What progress looks like

A beginner may initially merge all ‘other’ answers. After instruction, the learner can build a small table with separate columns for eligible, valid response, inapplicable and missing; calculate a justified percentage; and write a one-sentence limitation. The improvement is observable in the table and explanation, not in a claim that survey results must become more positive.

Frequently asked questions

Can ‘not applicable’ and ‘no’ be combined?

Only if the research question genuinely defines them as the same category, which is uncommon. Usually one describes a negative answer about an applicable situation and the other says the question does not fit.

Should a survey always offer ‘I don’t know’?

No single rule applies everywhere. An explicit uncertainty option may be appropriate and can affect responses. The researcher should justify the design and test whether it supports the intended interpretation.

What if the form software skips the follow-up automatically?

Record the screening response and make clear that the item was not presented to an ineligible respondent. A system skip is different from an eligible respondent seeing the question and leaving it blank.

The Core Aim of Education | How to Use ‘Not Applicable’ Without Hiding Missing Survey Answers