Description
A matrix with columns as successive time periods and rows as components, roles, settings or activities, used to see when changes occurred and to spot gradual drifts.
The time-ordered matrix is the general form of the event listing. Its columns are time periods in sequence, and its rows depend on what is being studied: components of an innovation, roles, settings, event types or activity types. In the book's example the rows were the official components of a work-experience programme plus 'other aspects' that emerged in fieldwork, and the columns were the planning period and three school years. Each cell holds a short description of a change reported and confirmed by at least one other person; a blank cell is meaningful because it shows stability. Reading across rows reveals 'drifts', such as stiffer demands, more control and increased routinisation. The analyst then returns to the notes to explain them, and can condense the matrix into a summary table that names each drift with a gerund ('tightening up') and counts mentions. The book advises choosing time periods fine enough to keep sequences apart, looking for timing, sequence and stability, and expecting about two or three hours to assemble once notes are coded. In UX this suits tracking how a feature, workaround or team practice evolves across releases or months of use.
Attributes
- Discipline
- Qualitative Analysis & Synthesis
- Loss type
- Latent
- Stage
- Post-launch, Discovery
- Data
- Qualitative
- Effort
- Hours
Steps
- Decide the rows: product components, user roles, settings or activity types.
- Choose time periods as columns, fine enough to separate the sequences that matter.
- Set a decision rule for entry, for example a change counts if one other person confirms it and nobody contradicts it.
- Enter brief descriptions of changes; leave cells blank where nothing changed.
- Read across rows for drifts, name them, and return to the notes for explanations.
- Condense into a summary table of drifts with counts and significance if the full matrix is too busy.
Sources
- Miles, M. & Huberman, A. M., Qualitative Data Analysis: An Expanded Sourcebook, 2nd ed. (1994)
Also known as
- Time-ordered display
- Change-over-time matrix
Categories
- User Research
- Qualitative Analysis & Synthesis