Matrix questions π’
Not as exciting as The Matrix, but matrix questions in Ethnio are pretty powerful, too. Mildly controversial from a survey science UX perspective, a matrix question asks participants to rate several items on the same scale, in one grid. Each row is an item (a feature, a statement, an attribute) and each column is a point on the scale. Participants answer every row without re-reading the scale each time, and you get comparable ratings across items in a single question.
Use a matrix when you want to compare things against each other, like rating ease of use, reliability, and support on the same scale. If you only have one item to rate, a single select or a scale question is simpler for participants. If you have more than six or seven rows, consider splitting into two matrix questions so the grid stays readable.
Adding a matrix question
Open your study and go to Editor βΊ Content. Click Add content and choose Matrix. The question lands in your content list with a default grid you can edit, and its settings open in the panel on the left.
Type your question in the text box at the top of the card. You can bold part of the question, add an emoji, or pipe in values with the pipe icon (see Piping below).
Rows and columns
Rows are the items you want rated. Columns are the scale.
- Click + Row to add an item. Type its label in the row field. Click the image icon next to a row to attach an image to that row, which is handy for rating designs, screens, or product shots.
- Click + Column to add a scale point. Type the label participants will see, like Very low or Strongly agree.
- Drag the handle on any row or column to reorder it. Click the Γ to remove one.
Keep column labels short. The grid gets narrower with every column you add, and long labels wrap on smaller screens. Five columns is a comfortable maximum for most desktop layouts; on mobile, Ethnio reflows the grid so width is not a concern (see Mobile below).
Matrix type: Default or Bipolar
Default is the classic grid. Every column has a label (Very low, Low, Neutral, High, Very high) and every row is rated against those labels.
Bipolar puts two opposing statements at either end of each row, with unlabeled points between them. Participants pick the point closest to their view. Each row has its own left and right statement, so one row might run from "I prefer short examples" to "I prefer very long examples" while the next runs from "Confusing" to "Clear." Columns in a bipolar matrix have no text labels, only positions. Bipolar works well for semantic differential questions, where you want to know where someone sits between two extremes rather than how much they agree with a single statement.
To switch, pick Default or Bipolar under Matrix type in the left panel. When you switch to Bipolar, each row gains a second text field for its right-hand statement.
Answer type: Buttons or Sliders
Buttons shows a radio button (or checkbox, if multiple answers are allowed) in each cell. This is the standard look and the easiest to answer on any device.
Sliders replaces the buttons in each row with a slider that snaps to each column position. Sliders feel more natural for bipolar matrices and for scales where the midpoint matters, but they take slightly more effort on mobile. Pick one under Answer type in the left panel.
Options
These toggles live in the left panel when a matrix question is selected.
| Option | What it does |
|---|---|
| Required | Participants must answer every row before moving on. |
| Allow multiple answers | Cells become checkboxes so a participant can pick more than one column per row. Use this for "select all that apply" grids, like which channels they use for each task. Leave it off for rating scales. |
| Mobile sees one row at a time | On phones, each row becomes its own step instead of a stacked list. Good for long grids or rows with images. |
| Show scale numbers in matrix | Displays the numeric value of each column next to its label, both in the editor and to participants (for example β2 Very low, β1 Low, 0 Neutral, +1 High, +2 Very high). |
| Auto-advance to next page | Moves participants to the next page as soon as they have answered every row. Great for intercepts. |
Column scale starts at
Every column carries a number behind the scenes, and this setting picks the number for the first column. The rest count up by one. With five columns and a start of β2, the scale runs β2, β1, 0, +1, +2, which puts Neutral at zero. With a start of 1, the same five columns run 1 through 5.
The numbers are what you see in Responses and exports, so pick the scheme that matches how you analyze. Turn on Show scale numbers in matrix if you also want participants to see them.
Piping
Click the pipe icon in the question text to insert a value from elsewhere in the study: the incentive amount, the interview duration, any embedded data field (like a UTM parameter), or the answer to an earlier question. For example, "How would you rate {Q2 label} on each of the following?" fills in whatever the participant chose in Q2.
Logic
Skip logic and Display logic are at the bottom of the left panel. Display logic controls whether the matrix shows at all, based on earlier answers or embedded data. Skip logic sends participants somewhere else based on how they answer. Both work the same way as on other question types; see Logic flow for the full picture of how your questions connect.
π± What participants see on mobile
On a phone, the grid reflows so participants never have to scroll sideways. Each row becomes a collapsible section with the row label as its header, and the columns stack vertically underneath as tappable options. One row is open at a time; tapping another row header expands it and collapses the previous one. Turn on Mobile sees one row at a time if you would rather show each row as its own step. Use Preview in the top right of the editor to check both the desktop and mobile layouts before you publish.
Sliders on mobile only work with numbers
If you show scale numbers in your matrix options, then sliders can render on mobile, but if you turn that option off, the sliders contain your text labels which are far too wide for mobile
Responses and exports
Each row is stored as its own answer, so a three-row matrix produces three values per participant. In Responses you can see the label chosen for each row along with its scale number. Exports include both, which makes it easy to average scores across participants or compare rows side by side.
Tips
- Use an odd number of columns (5 is the sweet spot) when you want a true midpoint. Use an even number when you want to force a lean one way or the other.
- Keep every row on the same scale. If one item needs a different scale, give it its own question.
- Write rows as short labels or short statements, not full questions. "Ease of use" or "The checkout was easy to complete" both work; "How easy was the checkout?" does not fit a grid.
- Put the negative end of the scale on the left and the positive on the right. Participants expect it, and it keeps your scale numbers climbing in the direction you would expect.
- Add a page break after a matrix if the next question depends on it, so participants finish the whole grid before moving on.