Skip to main content

Levels of Measurement — Previous Year Questions

📊 Topic Overview​

MetricValue
Papers Appeared12 of 27 (44.4%)
Priority🟠 HIGH
Section DistributionSec A: ~2
Most Repeated"Describe the four levels of measurement and indicate the data for which they are used."

🎯 Must-Know Concepts​

  1. Four scales: nominal, ordinal, interval, ratio — with characteristics and examples
  2. Each higher scale has all properties of lower scales plus additional properties
  3. Which statistical tests are appropriate for each scale
  4. Real psychological examples for each level

🔥 Most Repeated Question​

Common phrasings:

  • "Describe the four levels of measurement and indicate the data for which they are used."
  • "Describe the levels of measurement with suitable examples."

Answer (Section B — ~250 words):​

Levels of Measurement (S.S. Stevens, 1946) classify the properties of measurement scales. Four levels exist, each permitting different mathematical operations and statistical analyses.

1. Nominal Scale (Categorical) The lowest level. Numbers or labels are used only to name or categorise — they have no quantitative meaning.

Properties: Classification only. No order, no equal intervals, no absolute zero.

Examples: Gender (1 = Male, 2 = Female), religion, blood group, diagnostic categories (depression, anxiety, phobia).

Permitted statistics: Mode, frequency count, chi-square test, phi coefficient.

2. Ordinal Scale (Ranked) Numbers indicate rank order — greater than or less than — but intervals between ranks are not necessarily equal.

Properties: Classification + order. No equal intervals, no absolute zero.

Examples: Class rank, Likert ratings (1 = strongly disagree to 5 = strongly agree), socioeconomic status (low, middle, high), rank order of preferences.

Permitted statistics: Median, percentile, Spearman's rho, Kendall's tau, Mann-Whitney U.

3. Interval Scale Equal intervals between values; arbitrary zero point (zero does not mean absence of the attribute).

Properties: Classification + order + equal intervals. No absolute zero.

Examples: Temperature in Celsius or Fahrenheit, IQ scores, year (AD/BC), attitude scales (when intervals are assumed equal).

Permitted statistics: Mean, standard deviation, Pearson's r, t-test, ANOVA.

4. Ratio Scale Highest level. Has all interval properties plus an absolute zero (zero means complete absence of the attribute).

Properties: Classification + order + equal intervals + absolute zero.

Examples: Height, weight, reaction time, number of errors, frequency of behaviour, income.

Permitted statistics: All statistical operations including geometric mean, coefficient of variation.

ScaleOrderEqual IntervalsAbsolute ZeroExamples
Nominal✗✗✗Gender, religion
Ordinal✓✗✗Rank, Likert scale
Interval✓✓✗IQ, temperature (°C)
Ratio✓✓✓Height, weight, RT

🧠 Memory Aid​

"NOIR": Nominal → Ordinal → Interval → Ratio (increasing sophistication)

📚 Related Content​

✅ Self-Assessment​

  1. Which scale has an absolute zero? Give an example.
  2. Why can you not compute a meaningful mean from ordinal data?
  3. What scale is IQ measured on? Why not ratio?
  4. What statistical tests are appropriate for nominal data?
  5. What is the key difference between interval and ratio scales?

💡 Exam Tips​

  • The comparison table with all four scales and their properties is essential for Section B.
  • Always include 2 psychology-specific examples per scale.
  • State which statistical tests are appropriate for each level — this differentiates a good answer.
  • IQ is interval, not ratio (zero IQ does not mean absence of intelligence) — a common exam point.