Summary: Repertory Grids as a UX Research Tool

Speakers: Paul Matthews (@paulusm) and Henry Osadzinski (@HenryOz)

Repertory grids use the idea of personal construct psychology, and was designed to work in the clinical psychology setting. However, it became popular with market researchers and education for career planning. It has recently come to be used in UX for inspiration and evaluation.

Advantages

Paul and Henry outlined the main advantages of using repertory grids within a UX contents. These included:

• Repertory grids give structure to interviews and create a tangible record of the interview ready for analysis.
• They are fun – interviews using a repertory grid tend to be easier, clearer and teach users something about themselves.
• You get an honest opinion and a deeper sense of how the interviewee thinks about the issue under consideration, rather than a yes/no response.

What Goes into a Repertory Grid

Henry explained that there are three core parts to a repertory:

• Elements – These are the topics or materials for the grid.
• Constructs/contrasts – Opposing concepts about the elements.
• Ratings – The scale between constructs and contrasts.

Example:

Activity

This is an activity used with a user on a one-to-one basis, working with them to fill in the grid.

Each participant had a copy of the grid and were asked to think of they five most favourite or commonly used mobile/tablet apps. They wrote these on pieces of paper and shuffled them together.

Participants were then asked to pick three and consider the ways in which two of the apps are similar and one is different. From this, they were asked to come up with a name for this concept and add this to the grid.

The names of the apps were written across the top of the grid under the ‘elements’ heading. Each was rated against each of the constructs and contrasts.

Participants then discussed their grids in pairs to understand how the grids could be used to stimulate discussion within an interview.

Feedback

Paul and Henry invited participants to reflect on the activity and the issues this raised. Comments included:

• Coming up with constructs was difficult, as many favourite apps are quite similar.
• Constructs might need breaking down/revising – Henry noted that if two constructs are similar, you can use a technique called laddering to get the interviewee to breakdown these constructs.
• Constructs/contrasts can reveal personal values and how interviewees see the world.
• Sometimes constructs were not applicable to all elements – Henry emphasised that you need to avoid this if possible, or get interviewees to contribute new constructs that fill the gap (without being leading!)
• Should elements should be discrete items or part of a larger design? – Henry advised that competitive items may work best with this technique
• The exercise highlighted crossovers and connections between elements – Henry suggested asking interviewees to suggest additional constructs to explain how they are thinking when these arise.

Using a Repertory Grid in a Quantitative Way

Paul recommended OpenRepGrid to work with the data generated by repertory grids and allow you to create visualisations of individual or groups of grids to identify patterns that may be useful.

As the repertory grid was designed as an individual, subjective activity, this doesn’t necessarily allow you to compare groups of people. There is some debate about whether you can/should aggregate results.

Paul concluded by suggesting some techniques that could be used to account for variations in constructs between individuals to get some useful information, including looking at the patterns of ratings to identify constructs that may be similar.