Measurement

Abstract This chapter addresses measurement, the assignment of numbers to variables. Applying measurement scales to information creates a way to quantify phenomena, such as attitudes or behaviors, objectively. Four levels of measurement are nominal, ordinal, interval, and ratio. Nominal data represent names or labels of categories; even when numbers are assigned, they have no quantitative meaning (e.g., 1 = pianists and 2 = singers). Ordinal data can be ranked, but the distance separating one value from another on a scale is not necessarily consistent (e.g., first chair, second chair). Interval data communicate both order and quantity, and the intervals between data points are assumed to be equal along a scale. With interval data (e.g., many test scores), it is possible to perform mathematical operations and there are many suitable statistical analyses. Ratio-level data are assumed to be from a scale that has a true absolute zero. Reliability and validity are critical to measurement. “Reliability” refers to the consistency of measurement. “Validity” generally refers to the extent to which a measurement tool measures what it is supposed to. If a measurement tool is unreliable (i.e., inconsistent), we have no confidence that scores represent true values. If a measurement tool does not measure what it is supposed to (i.e., is inaccurate), it is invalid.