Repertory Grid-Derived Rating Scales

Description

Using repertory grid interviews to elicit users' own bipolar descriptions of a domain, then consolidating them into a short set of rating scales for later testing.

Imposed rating scales can use words that users do not understand or would not choose. This method builds scales from users' own language. A small group takes part in repertory grid interviews using a shared set of elements, in the chapter eight game audio clips from altered-state and normal gameplay. Using the triad technique (which two of three are alike and how the third differs) or simple differentiation, each participant produces bipolar constructs and rates every element on each construct from 1 to 5. The grids are concatenated and analysed, for example with focus analysis or clustering, to find constructs used interchangeably. Similar constructs are then grouped into categories, each represented by one bipolar scale; the chapter derived seven for game audio: valence (nonviolent to violent), attention (atmospheric to urgent), distance, fear (nice to eerie), noise, speed and distortion. Ratings can be compared across element groups with modal analysis, plotted on a radar chart. The result is a quick, domain-specific instrument grounded in user vocabulary, which still needs further validation.

Attributes

Loss type
Latent
Data
Mixed
Effort
Weeks

Steps

  1. Choose a domain and a common set of representative elements (clips, screens, products).
  2. Run one-to-one repertory grid interviews, using triads or differentiation to elicit bipolar constructs, with a practice construct first.
  3. Have participants rate every element on each of their constructs on a 1 to 5 scale.
  4. Combine grids and use focus analysis or clustering to find related constructs.
  5. Group related constructs into categories and choose one bipolar label per category as a scale.
  6. Test the draft scales by checking they separate element groups that should differ, then refine and validate.

Sources

  • Garcia-Ruiz, M. (ed.), Games User Research: A Case Study Approach (2016)

Also known as

  • User-derived semantic differential scales
  • Construct-based scale development
  • Game audio rating scale

Categories

  • Analytics & Behaviour
  • Survey & Psychometrics