Nominal data are data that can be classified into distinct groups. For example, statistical analyses investigating the differences between men and women make heavy use of nominal data, because "men" and "women" are two distinct groups. Statisticians often use Chi-Square statistics in their analyses of nominal data.
Ordinal data is similar to nominal data in that the data can fit into distinct groups. However, ordinal data differs from nominal data in one important way: the groups can be ordered in a clear way. Researchers of well-being (happiness) often use ordinal data when using self-report surveys. Subjects answer questions about life with the answers "satisfied," "neutral," "unsatisfied" and so on. While these are distinct groups, they also have a natural order, i.e. unsatisfied to satisfied. Many forms of categorical data analysis allow researchers to retain the ordering of classifications during the analysis procedure.
Interval data take the form of numbers. All possible numbers can be used for interval data, but the number zero has no meaning for interval data. For example, in attraction studies, researchers can rate the faces of subjects according to how good-looking they are. These numbers of these ratings have no objective meaning, but researchers can use the numbers as a method of comparing one face to another. In this sense, the size of the number has meaning, but still cannot be interpreted in normal language. Many statistical techniques, such as ANOVA and regression use interval data as input.
Ratio data is very similar to interval data. Ratio data, like interval data, is in the form of numbers, but the number zero plays an important role. This objective definition of the zero point allows researchers to interpret ration data. An example is salary. A person's salary is ratio data, and a salary of zero implies that the subject has no income. The types of statistical analyses that examine ratio data are plentiful, with common methods including general linear models and logistic regression.