Measurement in Research

1. Meaning of Measurement

• Measurement is a systematic process of assigning numbers, symbols, or values to objects, events, or attributes according to specific rules.
• It transforms abstract concepts (like attitude, motivation, adoption) into empirical indicators that can be studied scientifically.

2) Definition

• Kerlinger (1986): “Measurement is the assignment of numerals to objects or events according to rules.”
• Stevens (1946): “Measurement is the assignment of numerals to objects or events according to a rule.”

In simple words: Measurement means quantifying qualities and qualifying quantities in research.

3) Postulates of Measurement

Measurement rests on logical and mathematical postulates which ensure consistency and accuracy.

• Classification (a = b or a ≠ b); Any two things are either equal or not equal. E.g., Farmer A is literate and Farmer B is literate → a = b.
• Equality (If a = b and b = c, then a = c); Objects equal to the same object are equal to each other. E.g., If farmer A has 10 years’ experience, farmer B has 10 years’ experience, then farmer A = farmer C (if C also has 10 years).
• Transitivity of Inequality (If a > b and b > c, then a > c); Inequalities can be ordered logically. E.g., If Farmer A’s income > Farmer B’s income, and Farmer B’s income > Farmer C’s, then Farmer A > Farmer C.

These postulates guarantee scientific objectivity and logical consistency.

4) Levels of Measurement (S.S. Stevens, 1946)

Stevens proposed four levels (scales) of measurement, each with increasing precision and statistical applicability.

(i) Nominal Scale (Lowest level); Numbers used only as labels/categories.

• No order, no magnitude.
• Examples: Gender (1 = Male, 2 = Female), Religion, Occupation.
• Statistics allowed: Frequency, percentage, mode, chi-square test.

(ii) Ordinal Scale; Categories can be ranked/ordered, but intervals between ranks are not equal.

• Only relative position matters.
• Examples: Socio-economic status (Low, Medium, High), Attitude rank, Farmer adoption levels.
• Statistics allowed: Median, percentile, rank correlation (Spearman’s).

(iii) Interval Scale; Equal intervals between scale points, but no true zero.

• Allows comparison of differences, but not ratios.
• Examples: Temperature in °C or °F, IQ scores, Attitude scales.
• Statistics allowed: Mean, standard deviation, correlation, t-test, ANOVA.

(iv) Ratio Scale (Highest level); Has all the properties of interval scale + absolute zero.

• Ratios are meaningful (twice, half, etc.).
• Examples: Age, Income, Yield, Number of extension contacts.
• Statistics allowed: All parametric statistics (mean, SD, correlation, regression, ANOVA).

Summary Table

Key Exam Facts

• Measurement = assignment of numbers by rules.
• Postulates = Classification, Equality, Transitivity of Inequality.
• Levels of measurement = Nominal, Ordinal, Interval, Ratio (Stevens, 1946).
• Nominal = lowest precision, Ratio = highest precision.
• Most social science research uses ordinal and interval scales.