Description
Classifying each variable as nominal, ordinal, interval or ratio so that the summaries, tests and models applied to UX data are appropriate to what the numbers mean.
Stevens' typology distinguishes nominal (categories, e.g. device type), ordinal (ordered categories, e.g. single Likert items), interval (equal intervals without a true zero, e.g. multi-item scale scores treated as such) and ratio (true zero, e.g. task time, error counts). The level constrains permissible statistics: modes and chi-square for nominal; medians, rank tests and ordinal models for ordinal; means and linear models for interval and ratio. In SPSS the measurement level is declared per variable and affects procedures and charts. Rigour comes from declaring levels deliberately, recognising contested cases (averaging single Likert items), and choosing models that respect the data, such as ordinal regression for single items, rather than defaulting to means everywhere.
Attributes
- Discipline
- Statistics & Modelling
- Loss type
- Vestigial
- Stage
- Evaluation
- Data
- Analytical / expert
- Effort
- Hours
Steps
- List every analysis variable with its storage type (numeric, string, date) and conceptual measurement level.
- Declare levels in SPSS (VARIABLE LEVEL) or as pandas categorical or ordered categorical types.
- Treat single rating items as ordinal by default; treat composite scales of 4 or more items as approximately interval if reliability is adequate.
- Treat times and counts as ratio, checking skew and considering log transforms or count models.
- Choose summaries and tests consistent with each level and document exceptions with justification.
- Check that dummy coding is used for nominal predictors in regression rather than arbitrary numeric codes.
Sources
- SPSS Syntax Editor
- Python integration (SPSS Python plugin)
- pyreadstat for Python ↔ SPSS bridge
- Stevens, S. S. (1946) On the theory of scales of measurement. Science
- Field, A. (2024) Discovering Statistics Using IBM SPSS Statistics, 6th ed.
- The Inquiry Series, C6 — SPSS for UX Researchers — The Right Way
- Myatt, G., Making Sense of Data: A Practical Guide to Exploratory Data Analysis and Data Mining (2007)
Also known as
- Levels of measurement
- Stevens' scales of measurement
- Nominal, ordinal, interval, ratio
Categories
- Analytics & Behaviour
- Statistics & Modelling