Anchored MaxDiff Analysis

Anchored MaxDiff is a simple yet effective model to enhance a MaxDiff exercise from a relative model (where one option's utility is relative to the others) - to an absolute model. Anchored MaxDiff experiments supplement standard Max-Diff questions with additional questions designed to work out the absolute importance of the attributes.

Traditional MaxDiff experiment identifies the relative importance of the attributes, an anchored MaxDiff experiment permits conclusions about whether specific attributes are actually important or not important at all.

It may be that feature A is twice as important as feature B, which we would find out from MaxDiff. It may also be that neither A nor B is very important at all. We would not find this out from traditional MaxDiff. With anchoring - this allows researchers to add in an additional "Anchoring" question that determines if any of the features are fundamentally important.

There are two types of anchoring models we can use;
• Binary Direct
• Dual Response

Both these models allow for anchoring add in an additional step/question into the max diff exercise in the form of a multiple-choice question

The binary-direct anchoring model is by far the easiest and most convenient model for the researcher as well as the user.

In the Binary-Direct model, respondents are shown all the available choices and asked for all the "Must-Haves" - this allows respondents to draw a fundamental line between the "must-haves" and "nice to have" - this creates the binary anchor.

With the dual-response model, respondents are asked to identify some, all or none of the options - using a simple uni-choice question.

Anchoring is turned off by default in your QuestionPro survey. You can go to the MaxDiff question and turn the toggle on to allow anchoring in your max-diff survey.

When anchoring is enabled - depending upon the anchoring model used (binary direct or dual response) - the raw data will contain the output of that question type in an easy to understand model for a researcher.