Scales of measurement

What it means to “measure” something has long been a topic of both scientific and philosophical debate. The concept of measurement is fundamental to the field of psychology because we need reliable measurements of psychological constructs in order to trust any statistical results pertaining to those constructs. Despite the importance of measurement, this topic is often glossed over in many psychological applications—researchers often begin by assuming that they have measured their construct of interest, without necessarily providing any concrete evidence that such measurements are reliable or valid. Of course, this is a serious problem for interpreting results of psychological studies because statistical methods cannot overcome issues pertaining to poor measurement. More specifically, most statistical methods abide by the “garbage in, garbage out” principle, so you should expect to obtain invalid results if your input variables are measured inadequately. In this chapter, we will not cover all of the specifics regarding psychological measurement— entire books and courses have been devoted to this topic. Instead, I will provide a brief overview of the “Theory of Scales of Measurement” that was proposed by Stevens (1946). In this influential paper, Stevens defined measurement as “the assignment of numerals to objects or events according to rules” (p. 677), and this broad definition still seems to be embraced by many applied psychological studies. In his paper, Stevens presents four different scales (or levels) of measurement that can characterize different types of measures that are used in psychological and other social science studies. It should be noted that Steven’s approach to measurement has been widely criticized by researchers who specialize in measurement and statistics (e.g., see Michell, 1986). However, it is important to understand Steven’s ideas, which are an implicit part of applied psychology.