> ## Documentation Index
> Fetch the complete documentation index at: https://docs.voxworks.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Single choice

> The Single choice question type: one answer from a list. Covers answer options and keys, Read out and Volunteered only options, Keep position, Accept / Often heard as / Reject, the options lead-in, shuffling, answer mode, confirmation, don't-know and refusals, and what is recorded.

## What it's for

A **Single choice** question takes exactly one answer from a list — *"Which of these best describes where you live? Would you say: a capital city, a regional town, or a rural area?"* It's the default type for a new question and the workhorse of most surveys.

* For a scale (1 to 5, or *very satisfied* to *very dissatisfied*), use [Rating](/surveys/question-types/rating) — it works the same way.
* To ask several items against the same list, turn on **Battery** — see [Battery (grid)](/surveys/question-types/battery).
* There's no multiple-choice type for voice. Ask a yes/no Single choice per item, or a [Long text](/surveys/question-types/long-text) question and code the answers afterwards.

***

## Settings

Alongside the settings every question shares ([Question settings](/surveys/question-settings)), a Single choice question has:

| Setting                                      | What it does                                                                                                                                                                                                    | Default              |
| -------------------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | -------------------- |
| **Answer mode**                              | **Read out** — the options are read to the respondent after the prompt. **Unprompted** — no options are read; the respondent answers in their own words and the interviewer matches what they say to an option. | **Read out**         |
| **Battery**                                  | Turns the question into a battery: the options become a scale, asked against a list of items. *"Off: one answer, chosen from the options below."*                                                               | Off                  |
| **Shuffle answer order**                     | *"A new random order on every call."* Off: *"Options are read in the authored order, the same on every call."* Options with **Keep position** stay where they are.                                              | Off                  |
| **Options lead-in**                          | What's said between the question and the first option. *"Spoken between the question and the first option. Leave blank for the default."*                                                                       | *"The options are:"* |
| **Answer options**                           | The list of answers — see below. At least two.                                                                                                                                                                  | **Yes**, **No**      |
| **Read every answer back**                   | Reads every answer back for a yes. Off, only possible mishearings are checked. See [Answer confirmation](/surveys/answer-confirmation).                                                                         | Off                  |
| **Screen out if the question cannot be run** | Screens the respondent out if the question's own setup fails on the call, instead of moving on. Only worth turning on for a screener.                                                                           | Off                  |

Questions imported from another platform can show extra answer modes (**Interviewer recorded (imported)**, **System (imported)**) or an **Answer order** picker with **Reverse** or **Cyclic rotation**. These are kept as they came in; new questions offer only the settings above.

### The options lead-in

The lead-in turns the prompt and the list into one natural sentence. The interviewer adds the space after it for you.

| Prompt                              | Lead-in          | The respondent hears                                                            |
| ----------------------------------- | ---------------- | ------------------------------------------------------------------------------- |
| *"What is your age?"*               | *Are you:*       | "What is your age? Are you: 18 to 24, 25 to 34, …"                              |
| *"How would you rate the service?"* | *Would you say:* | "How would you rate the service? Would you say: excellent, good, fair or poor?" |
| *"Which party would you vote for?"* | *(blank)*        | "Which party would you vote for? The options are: …"                            |

Good lead-ins are short: *"Would you say:"*, *"Is it:"*, *"Are you:"*, *"Would that be:"*. Vary them between neighbouring questions so the interview doesn't fall into a rut.

***

## Answer options

Each option row has:

| Part                                | What it does                                                                                                                                                                                                                                    |
| ----------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **Option text**                     | The answer as the interviewer reads it — and as it's recorded.                                                                                                                                                                                  |
| **Key**                             | How routing rules refer to the option. It follows the option text (lower case, underscores) until you edit it. Lower case letters, numbers and underscores only; each key must be unique on the question. Once rules use a key, leave it alone. |
| **Read out** / **Volunteered only** | **Read out** options are read in the list. **Volunteered only** options are never read, but are recorded if the respondent says them.                                                                                                           |
| **Keep position**                   | *"Holds its authored position in every presented order."* Use it for an option like *None of these* that must stay last when the rest are shuffled. Not available on Volunteered only options.                                                  |
| **Voice**                           | Opens the option's matching lists — see below. The chip shows how many lines it holds.                                                                                                                                                          |

Use the arrows to reorder options, **+ Option** to add one, and the **×** to remove one. An option that a routing rule reads can't be removed until the rule is changed — the save is refused with a message naming the rule.

<Warning>A question can hold at most **ten** options, Read out and Volunteered only together. When you compile, any beyond the tenth are dropped (with a compile warning). For a long list, ask a yes/no gate question first and follow it with an open [Long text](/surveys/question-types/long-text) question.</Warning>

### Volunteered only options

Volunteered only options let the interviewer record answers you don't want to prompt:

* **Don't know** and **Prefer not to say** — when these exist as Volunteered only options, a respondent who says them has that answer recorded instead of being nudged and skipped.
* **Common "other" answers** — code the answers you expect as their own Volunteered only options.
* **A catch-all Other** — a Volunteered only option whose **Accept** line reads something like *"any answer that isn't one of the other options — record it as other without asking further"*.

There's no "other, please specify" on voice. Don't follow an *Other* answer with a "please specify" question; if the detail matters, ask a Long text question instead.

### Accept, Often heard as and Reject

Click **Voice** on an option to open three lists, one line per entry:

| List               | Hint in the app                                            | Use it for                                                     | Example (option *Concerned*)         |
| ------------------ | ---------------------------------------------------------- | -------------------------------------------------------------- | ------------------------------------ |
| **Accept**         | *"What counts as this option."*                            | The ways people genuinely say this answer.                     | *worried*, *a bit anxious about it*  |
| **Often heard as** | *"What the recogniser produces instead. Never a keyterm."* | What the speech recogniser writes when it mishears the option. | *can send*, *concert*                |
| **Reject**         | *"What must not be taken as this option."*                 | A neighbouring answer that must not be matched here.           | *very concerned* (a separate option) |

* Put phrases people really say in **Accept**, not **Often heard as**. An answer matched through **Often heard as** is always checked with a "did you say…?", so a real phrase listed there costs every respondent who uses it an extra turn.
* Use **Reject** sparingly, for a genuine neighbour. Don't use it to separate degrees ("a bit" vs "very") that the options already separate.

The **Recode value** field (*"Reported value — blank uses the label"*) is also in this panel. The answer recorded on the call is the option's text. Speak to the Voxworks team for further custom configuration.

***

## What the respondent hears

**The ask.** The prompt, the lead-in, then the Read out options: *"How satisfied were you with your visit? Would you say: very satisfied, satisfied, dissatisfied or very dissatisfied?"* On an **Unprompted** question, only the prompt.

**Hearing the options again.** "What were the options?" gets only the options again — *"Sure — the options are: …"* — without the whole question. This doesn't use an attempt, and an answer given straight afterwards is recorded. On an Unprompted question the interviewer says *"I'm not able to read the responses out — just tell me in your own words."*

**A clear answer** is recorded, and the interviewer moves on — with a moving-on line if that's turned on.

**A possible mishearing** — an answer matched only through an **Often heard as** line, or because it sounds like an option — is checked first: *"Did you say neutral?"* A "yes", or repeating the option, records it. See [Answer confirmation](/surveys/answer-confirmation).

**An answer that isn't one of the options** gets a clarify line:

* unclear — *"Sorry, I didn't quite catch that — which one was it?"*;
* off the list, when the question has Volunteered only options — *"It would need to be one of the listed options — though I can also record Don't know if that fits better?"*;
* off the list otherwise — *"I can only take one of the options I read out — which would be closest for you?"*

**"I don't know"** gets one nudge — *"No worries — if you had to pick one, which would be closest?"* — and a second "don't know" skips the question, unless a Volunteered only *Don't know* option exists, in which case that answer is recorded.

**A refusal of this question** ("I'd rather not say") gets one gentle nudge — *"That's no problem at all — we can skip it. Unless one of those is close enough for you?"* — then the question is skipped. Set **Decline** to 0 to accept the first refusal (unless the question is **Required**).

**"What does that mean?"** gets a short, neutral definition drawn from the question and its **Context for the matcher**, without suggesting an answer.

**Silence** is met with a re-read in other words, then skipped when the **Ask** cap is spent.

**A question about the call, "I'm busy" or "stop"** is handed to your [FAQs](/surveys/faqs) or [Behaviours](/surveys/behaviours); after an FAQ answer, the question resumes in other words. See [Answers given early and interruptions](/surveys/riders-and-interruptions).

The number of each of these turns is set by the attempt caps — see [Attempts and step levels](/surveys/attempts-and-step-levels).

***

## Routing on the answer

Rules on the **Logic** tab read the answer by option **key**, never by the spoken words:

* **Jump to question**, **Show question** or **Hide question** — a later question, depending on the answer.
* **Screen out** or **Terminate interview** on a particular answer — the respondent hears that outcome's closing text and the farewell. If the screening answer was matched from a possible mishearing, it's confirmed before the respondent is screened out.
* **Include option** / **Exclude option** on a later question — for example, drop the party a respondent already chose from a "second preference" question.
* **Set variable** to keep the answer for a later prompt.

See [Routing rules](/surveys/routing-rules).

***

## What is recorded

| Variable                                 | Type | Example          | Notes                                                                     |
| ---------------------------------------- | ---- | ---------------- | ------------------------------------------------------------------------- |
| `{{survey.<key>.result}}`                | text | `Very satisfied` | The chosen option's text. Empty when skipped.                             |
| `{{survey.<key>.status}}`                | text | `captured`       | `captured`, `skipped`, `invalid`, `abandoned`, or empty if never reached. |
| `{{survey.<key>.termination_requested}}` | text | `false`          | `true` when this answer screened the respondent out.                      |

In **Results → Questions**, the question's card lists the most common answers with counts and percentages. In the CSV export, each response has one row for the question, with the option text in the `answer` column. See [Results](/surveys/results) and [Responses and export](/surveys/responses-and-export).

***

## Next Steps

* [Rating](/surveys/question-types/rating) — the same question, as a scale.
* [Battery (grid)](/surveys/question-types/battery) — several items against one list.
* [Answer confirmation](/surveys/answer-confirmation) — when answers are read back.