# Missing survey answers and eligible denominators

Separate skipped questions, not-applicable cases and valid answers before interpreting an item’s answer percentage.

If nothing was unclear or difficult, say so. Skip questions about steps you did not experience; use not applicable where appropriate.

## 1. What support interaction would you use when answering this question?

Answer: ____________________

## 2. What would you do if you had not had that interaction?

Answer: ____________________

## 3. Which part of the question makes an answer difficult to give?

Answer: ____________________

## 4. What wording would make the skip instruction easier to understand?

Answer: ____________________

## Author notes / illustrative keys where applicable

1. Establish the experience behind the answer without asking for private ticket details.
2. Check whether not applicable is distinguishable from a skipped relevant question.
3. Separate unclear recall from missing experience.
4. Test a respondent-visible instruction rather than forcing a neutral rating.

Interpretation caution: Do not recode missing or not-applicable answers as neutral or dissatisfied. Item eligibility can be unknown and must then be reported as unknown.


## Illustrative finding and follow-up

Observation: Synthetic cohort: 50 people finished the survey; 12 experienced the support interaction, eight answered the item and four eligible people skipped it.

Action to test: Report eight of 12 eligible people answering the item, with four eligible skips and 38 out-of-scope cases retained separately.

Follow-up: Repeat the pilot after clarifying the eligibility prompt and check whether people can distinguish not applicable from choosing to skip.


## Collection protocol

When and whom to ask: Pilot after a recent support interaction with exposed and unexposed volunteers; allow a skip at every prompt.

Decision owner: The questionnaire reviewer owns exposure wording; the analyst records the item disposition rule.

Follow-up: Inspect item-level reasons for missingness separately from overall survey participation.

Response handling: Do not collect account credentials, ticket contents or private conversation extracts. Pilot participation and explanation of a skip remain optional.


## Worked measurement study

All example records are synthetic; manual worksheet, no automatic import.

### Inspect the synthetic example

Disposition | People
--- | ---
Valid item answers | 8
Eligible item skips | 4
Not exposed to this interaction | 38

### Derivation

Known eligible item denominator = 8+4=12. Answer fraction = 8÷12×100 ≈66.7%. Using all 50 gives 16%, which answers a different question.

Item note: eight valid answers, four eligible skips and 38 not-exposed cases; 66.7% of known eligible people answered. No missing answer is assigned an attitude.

Calculator inputs and convention: {"mode": "ratio", "labels": ["Valid item answers", "Known eligible people"], "values": [8, 12], "unit": "% eligible item answers", "integer": true, "subset": true}

### Evidence boundary

Observed / Exposure and three disposition counts / Supports the known-eligible item denominator.
Unknown / Attitudes of people who skipped / Do not fill their answers with an assumed rating.
Next check / Pilot the exposure and skip instructions / Retain unknown eligibility when it cannot be resolved.

### Method references

AAPOR survey best practices: https://aapor.org/standards-and-ethics/best-practices/ — Pretesting and transparent sample context.