Statistical measurements require variables, but all variables are not the same. Some variables like weight or speed or dollars spent can be precisely measured. Opinions, though, are a different matter. Patients can rate their level of pain on a scale of one to ten, or movie-goers can rate how well they enjoyed a film they just saw. These types of indicators are ordinal measurements. They are not precise the way physical or economic measures can be, but ordinal measures can nevertheless provide valuable information for researchers.

The different statistical variables include categorical, interval, ratio and ordinal variables. Categorical variables refer to types without order. Birds, mammals, reptiles and fish are types that can be named but have no mathematical order in relation to one another. Interval variables are variables that relate equally along a common scale; for example, temperature changes, where the difference between 50 and 60 degrees is the same as the difference between 60 and 70 degrees -- 10 degrees.

Ratio variables begin with zero representing equality between two things, and proceed to factors representing relative difference. Comparing the population of China to the United States, a ratio variable might take the United States as the zero-base with 311 million people, which gives China, with 1.3 billion people, a ratio value of 4.29. China has 4.29 as many people as the United States. Ordinal variables measure qualities; for example, a survey might say, “With your current governor, you are: (1) very unsatisfied, (2) unsatisfied, (3) have no opinion, (4) satisfied or (5) very satisfied.”

Ordinal measurement is designed to infer conclusions, while other methods are used to describe conclusions. Descriptive conclusions organize measurable facts in a way that they can be summarized. If a statistical analysis of average per capita income in a town changes over three years, that change can be stated quantitatively. No inference, however, can be drawn about why the average changed. What you see is what you get: numbers. Inferential conclusions attempt to see beyond the actual numbers to some qualitative conclusion, for example, "Most customers of Frosty Boy Ice Cream are satisfied."

Ordinal measurement is normally used for surveys and questionnaires. Statistical analysis is applied to the responses once they are collected to place the people who took the survey into the various categories. The data is then compared to draw inferences and conclusions about the whole surveyed population with regard to the specific variables. The advantage of using ordinal measurement is ease of collation and categorization. If you ask a survey question without providing the variables, the answers are likely to be so diverse they cannot be converted to statistics.

The same characteristics of ordinal measurement that create its advantages also create certain disadvantages. The responses are often so narrow in relation to the question that they create or magnify bias that is not factored into the survey. For example, on the question about satisfaction with the governor, people might be satisfied with his job performance but upset about a recent sex scandal. The survey question might lead respondents to state their dissatisfaction about the scandal, in spite of satisfaction with his job performance -- but the statistical conclusion will not differentiate.