Creating a MaxDiff

Introduction

MaxDiff (Best-Worst Scaling) is a method used to estimate preference or importance scores for a set of items, such as brands, product features, or claims. Instead of asking respondents to rank a large list, MaxDiff breaks it down into smaller tasks. In each task, respondents choose the best and worst items from a subset, typically 4-6 items at a time. This approach makes the process more manageable for respondents and produces more accurate, reliable results.

MaxDiff also outperforms traditional rating scales by providing clearer distinctions between preferences. Since respondents select only "best" and "worst" options — without relying on a response scale — it eliminates common biases found in cross-cultural research.

MaxDiff Example

Creating a MaxDiff exercise

  1. Click any Add button.
    The survey authoring interface with the add buttons highlighted.
  2. Select MaxDiff.
    MaxDiff is highlighted in the add menu.
  3. Enter question text.
  4. Enter list items. Enter options one by one or paste in a list from a program such as Microsoft Word or Excel. There must be at least six items.
    A MaxDiff in authoring is shown with steps 3 and 4 above highlighted.

Specifying response options

A new list is automatically created for your question, or you can select an existing list from the list selection dropdown. Enter options one by one or paste them from programs like Microsoft Word or Excel.

Labels

Modify MaxDiff labels by clicking the Exercise Settings menu (Gear icon Settings - Gear) at the top right of the exercise. There are three different labels available for a MaxDiff.

  1. Best and worst labels: These label the scale on which a respondent considers and chooses the items in the list.
  2. Items label: This label gives context to the items the respondents are choosing among. For example, you can use the label “Cookies” over a list of cookies flavors.
  3. Show exercise progress (number of tasks) & Task label: This counter shows the respondent the progress they have made through the exercise. The counter can be turned off and the label field is not required.
The MaxDiff labels described above are highlighted.

Advanced settings

On the Advanced tab, you’ll find additional settings that aren’t necessary for preparing a basic MaxDiff exercise, but are available if you’d like to further refine it.

Exercise design

Discover has a design recommender that generates a default design, meeting the criteria for proper MaxDiff exercise designs that achieve a high degree of precision at the individual level. However, if you would like to view how the design has been configured or modify the design, you can do so here. To learn more about what goes into a design, please read our MaxDiff design settings article.

MaxDiff design settings can be configured on the second tab of the exercise settings.

Anchored MaxDiff

In a MaxDiff exercise, respondents choose items from groups of typically 3 to 6 items as either best or worst. These choices are relative judgments among the items the researcher has included in the study. Anchored MaxDiff establishes an anchor point, allowing researchers to determine if a respondent views the items as important/not important or buy/no buy in a more absolute sense. Read the Anchored MaxDiff article to learn more.

Anchored Question

Related Content

HELP ARTICLE

This 10 minute video provides an overview of MaxDiff.

TECHNICAL PAPER

Best-worst scaling gives you better information with fewer respondents—it works better than traditional rating scales.