> ## 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. # Battery (grid) > Battery questions: several items asked one at a time against the same scale. Covers turning Battery on, items and their keys, longer forms for re-asks, the decline caps, item shuffling, what the respondent hears, and how each item's answer is recorded. ## What it's for A **battery** — a grid, on a paper questionnaire — asks the respondent to rate a list of items one after another on the same scale: > *"I'm going to read out some areas of life. For each one, please tell me whether you think artificial intelligence will have a positive, negative or no real impact. For each one, just say: positive, negative or no real impact. Let's start with your own life?"* … *"And education?"* … *"And finally, healthcare?"* A battery is not a separate type. It's a [Single choice](/surveys/question-types/single-choice) or [Rating](/surveys/question-types/rating) question with the **Battery** switch on: *"The options are the SCALE; each item below is asked against it in turn."* *** ## Settings Turning **Battery** on changes the inspector: | Setting | What it does | Default | | ---------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | ------------------------------ | | **Battery** | On for a battery. | Off | | **Shuffle item order** | *"The battery’s items are asked in a new random order on every call."* Off: *"the battery’s items are asked in the authored order."* The scale is always read in its authored order. | Off | | **Items** | The things being rated — see below. At least two, up to 200. | None | | **Declines in a row** | How many items in a row the respondent can decline before the rest of the battery is skipped. Minimum 1. | 3 | | **Declines in all** | How many items in total the respondent can decline before the rest of the battery is skipped. Minimum 1. | 4 | | **Scale** | The answer options, now labelled **Scale** — each point with its key, **Read out** / **Volunteered only** and **Voice** (**Accept**, **Often heard as**, **Reject**, **Recode value**). At least two, and no more than ten. | **Yes**, **No** (replace them) | | **Options lead-in** | Introduces the scale in the opening read. *"For each one, just say:"* works well. | *"The options are:"* | **Answer mode**, **Voice prompt**, **Alternative phrasings**, **Required**, the rest of the **Voice** panel and the **Behaviour** panel work as they do for any question — see [Question settings](/surveys/question-settings). A battery never reads a rating back and never screens a respondent out, so **Read every answer back** and **Screen out if the question cannot be run** are hidden, and **Keep position** isn't offered on scale points. ### Items *"Each item is asked against the scale below, one at a time. The key is what the item’s result is reported under."* Click **+ Item** to add one. Each item has: | Part | What it does | | ---------------------------------------------------------- | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | **Label** | *"What this item is called"* — a short noun phrase of two to five words, such as *Your own life* or *Public transport*. The first time an item is asked, the interviewer says its label. | | **Key** | The name the item's result is reported under. It follows the label (lower case, underscores) until you edit it. Lower case letters, numbers and underscores only. | | **Longer forms, used when the item has to be asked again** | *"The first ask always says the label. Each longer form has to fit the same sentence the label does — “the impact of artificial intelligence in your own life”."* One per line, up to 20. When an item has to be asked again, one is picked at random, so the respondent hears it put a little differently. | Write each longer form so it can replace the label in the same sentence — after *"And …"*, *"How about …"* or *"For …"*. Leave it empty and every re-ask says the label again. Readiness blocks a battery with fewer than two items — *"… is a battery with fewer than two items."* — and a battery on any type other than Single choice or Rating. *** ## What the respondent hears **The preamble.** The voice prompt and the scale, spoken once: *"…please tell me whether you approve or disapprove. For each one, just say: approve, disapprove, or no opinion."* An answer given during the preamble isn't recorded against any item. **The items**, one at a time, with lead-ins that change with position — *"Let's start with Chris Minns?"*, *"And Mark Speakman?"*, *"And finally, the Greens?"* The first ask of an item doesn't repeat the scale, to keep the pace up. **Re-asks.** When an item has to be asked again, the scale comes back with it — *"For public transport, would you say approve, disapprove or no opinion?"* — using a longer form if you've written one. On a scale of three points or fewer, a request to hear it again brings the scale straight back. **Hearing the scale again.** "What were the options?" gets *"Sure — the scale is: …"*. This doesn't use an attempt, and a rating given straight afterwards counts for the current item. **A possible mishearing** — a rating matched only from an **Often heard as** line or because it sounds like a scale point — gets a single check: *"Sorry, did you say favourable?"* Otherwise ratings are never read back. **Declining.** * *"Skip that one"* declines just the current item, and the battery moves to the next. * *"I'm done with this"* stops the whole battery. * After **Declines in a row** declines one after another (3 by default), or **Declines in all** declines in total (4 by default), the rest of the battery is skipped. Rating an item resets the in-a-row count. **Attempts per item.** Each item gets the question's attempt caps (ask, clarify, don't know and so on). When an item runs out of attempts, it's left unanswered and the battery moves on to the next item — it doesn't end the battery. FAQs, busy and stop work as they do on any question — see [Answers given early and interruptions](/surveys/riders-and-interruptions). *** ## Routing * **Exclude battery item** (in the advanced rule builder) drops an item so it's never asked — for example, an item that doesn't apply to this respondent. A dropped item is recorded as `NA`. See [Routing rules](/surveys/routing-rules). * A battery can't screen a respondent out. *** ## What is recorded | Variable | Type | Example | Notes | | ---------------------------------------- | ----------------- | --------------------------------------------------------------- | -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | `{{survey..result}}` | text (a JSON map) | `{"your_own_life":"Positive","education":"","healthcare":"NA"}` | One entry per item, keyed by **item key**. The scale point's text means the item was rated; `NA` means it was excluded by a rule and never asked; empty means it was asked and not answered, or never reached. | | `{{survey..status}}` | text | `captured` | `captured` if at least one item was rated, `skipped` if none was. `abandoned` or empty as for other questions. | | `{{survey..termination_requested}}` | text | `false` | Always `false` — a battery never screens. | After the call, each item is stored as its own answer, with the item key as its occurrence key — so in the CSV export each rated item appears as a separate row with the item key in the `occurrence_key` column and the scale point in the `answer` column. See [Responses and export](/surveys/responses-and-export). *** ## Tips * Keep item labels short and parallel — *Your own life*, *Your job*, *Your family* — and put the detail in longer forms. * Keep the scale short. Three to five points is easy to hold in mind across a run of items. * Put the scale in the preamble with a lead-in like *"For each one, just say:"*; don't repeat it in every item. * Turn on **Shuffle item order** to spread order effects across respondents. *** ## Next Steps * [Single choice](/surveys/question-types/single-choice) — scale points work like answer options. * [Rating](/surveys/question-types/rating) — authoring a good scale. * [Avoiding repetition](/surveys/repetition) — longer forms and other ways to vary re-asks.