Variables in which higher numbers indicate higher levels of that which is measured, but the differences between adjacent numbers are not necessarily equal. Thus, what matters is the order, but not the values between variables. On ordinal measurement scales. they imply ordering of categories while the quantitative distance between levels is unknown, and the meaning of different levels may not be the same for different individuals. The actual scale of measurement for most ordinal variables is the Likert scale, and typically used in questionnaires. In fact, ordinal variables or scale is part of classification of a four-category classification of measurement scales devised by the psychologist Stanley Smith Stevens (1906-1976) in 1946. The other variables were: categorical (nominal), interval and ratio. Categorical variable applies to mutually exclusive, but ordered, variables. An interval variable is similar to an ordinal scale, but the intervals between the values are equally spaced. Sometimes there are variables that take a middle position between ordinal and interval, such as a five-point Likert scale, with values given as, for example, strongly agree, neutral, disagree, and strongly disagree. A ratio variable, like interval variables, have equal intervals between values. However, a ratio variable has a clear and meaningful zero point. An example of a ratio variable is heart rate, but really does find application in behavioral research. Finally. a variable expressed at one level at another (lower) one, but the reverse is not true. For example BMI is typically measured on the interval scale, but can be transformed to ordinal categories (>: obese, 25-29.9: overweight, and <25: underweight. From here, nominal categories can be derived: overweight and not overweight.
See Measurement efficacy, Measurement theory, Rating scale, Variable