ABSTRACT
According to Stevens (1951), “In its broadest sense measurement is the assignment of numerals to objects or events according to rules” (p. 1). The process of research begins with a problem about the relationship between two or more variables. Measurement is introduced when these variables are translated into labels (categories) or numbers. Stevens went on to describe a hierarchy of four scales or levels of measurement that he labeled nominal, ordinal, interval, and ratio. Since the 1950s, these four levels have been discussed in most statistics and research methods textbooks and used to describe the level of measurement necessary to compute certain statistics. In general, the mean, standard deviation, and parametric statistics such as the t test that use them are said to depend on having at least interval level measurement of the variables. However, we have found the distinction between the traditional ordinal and internal levels somewhat confusing to apply, and other researchers have argued that this traditional distinction is not necessary to use parametric statistics if the data are approximately normally distributed (Gaito, 1980, 1986; Velleman & Wilkinson, 1993).
