Levels of Measurement
Statistical investigations usually require in-depth data collection and analysis. For this reason, a researcher ought to possess a clear grasp of the fundamental levels of measurement. As a rule, a research variable (especially in quantitative experiments) has four levels of measurement including nominal, ordinal, interval, or ratio (Newcomer, Hatry, & Wholey, 2015). Psychologist Stanley Stevens defines a level of measurement as a cataloging process that describes information based on the values assigned to the individual variables. When viewed from this lens, a level of measurement is a differentiation method that shows information variations.
The Four Levels of Measurement
Nominal Level. This first level of measurement is used primarily for classification. One would be right to assert that it provides the least amount of details regarding a variable. It is often represented by alpha-numeric symbols or letters (Meaningful Evidence, 2015). A survey that explores lifestyle trends for three gender classes including male, female, and transgender appears quite exemplary. The investigator in such a scenario can assign letters M, F, and T – respectively – to the three variables. These letters would be considered the nominal level of measurement.
Ordinal Level. Like the nominal level of measurement, this one classifies the information (Newcomer, Hatry, & Wholey, 2015). A classroom performance rating system appears quite exemplary. Typically, a teacher assigns scores alongside student names during academic performance grading. The student names represent the nominal level of measurement while their scores mark the ordinal level of measurement (Norman, 2010). For the sake of this argument, consider the ranks 1-n (where n is defined by the total number of students): note that each student in each rank (nominal level) will have a unique score (ordinal level) (Meaningful Evidence, 2015).
Interval Level. This third level assumes two functions: classification of the measurements and specification of the quantitative/statistical gaps between each point. Note that this process follows the scale of low to high (Fife-Schaw, 2006). At this point, the reader is urged to reflect on the preceding classroom test grading example. The interval would be the score gap between two students: e.g. the interval between student 1 (score 95) and student 2 (score 80) would be the same as the interval between student 35 (score 21) and student 20 (score 16).
Ratio Level. This is the final, and the most detailed level of measurement. It observes the difference between two intervals then measures them on a scale containing the absolute value ‘zero (0)’. The primary objective of the ratio level of measurement is to link this interval to an absolute value (Norman, 2010). The classroom example still makes sense in this regard: score 90 will be said to be twice as much as score 45 since the measurement will be done on a scale that links both to an absolute value (zero).
The Distinction between Categorical and Continuous Levels of Measurements
By default, categorical levels of measurement (mainly ordinal level) offers a predetermined number of categories while continuous levels of measurement (mainly ratio) have an immeasurable number of values between two values. Time is a good example of a continuous level of measurement since there is infinite number of values between any two points: day-to-day, hour-hour, and century-to-century et cetera (Norman, 2010). On the other hand, categorical levels not logically ordered compared to continual levels of measurements. Gender, marital status, and material type are examples of data that fit into this category (Schnake & Dumler, 2003). Such a factor differs from the logically ordered continuous levels of measurement. This level of measurement assumes a value (or values) within a measurement scale. Temperature, time, distance, and weight are examples of data that fit into this category.
The Application of the Four Levels of Measurement in Public Administration
By now, it is obvious that the levels of measurement are common in statistical investigations. The United States Census Bureau is an example of a public administration organization that utilizes these levels to gain insight into social facts and trends. It categorizes the census data into the nominal (men, women), ordinal (age, income), interval (gaps), and ratio (men-to-women) levels (Ortman, Velkoff, & Hogan, 2017). Note that the items in brackets are among the various types of data that the bureau measures using the four levels.
Fife-Schaw, C. (2006). Levels of measurement. Research methods in psychology, 3, 50-63.
Meaningful Evidence (2015). 6 Steps to Effective Program Evaluation Planning. Retrieved from https://www.youtube.com/watch?v=jhQRXKsJzb0
Newcomer, K. E., Hatry, H. P., & Wholey, J. S. (2015). Handbook of practical program evaluation. Hoboken, NJ: John Wiley & Sons.
Norman, G. (2010). Likert scales, levels of measurement and the “laws” of statistics. Advances in health sciences education, 15(5), 625-632.
Ortman, J., Velkoff, V., & Hogan, H. (2017). An aging nation: the older population in the United States: population estimates and projection. Available from https://www.census.gov/en.html
Schnake, M. E., & Dumler, M. P. (2003). Levels of measurement and analysis issues in organizational citizenship behaviour research. Journal of occupational and organizational psychology, 76(3), 283-301.