Metrics for Analysis, Selection of

Levels of Measurement

When numbers are used as labels and each label is equal to another label, this represents the nominal level of measurement. Taken from the Latin word for name, the nominal level of measurement assigns numbers in lieu of category names. One can number objects, like basketball numbers for players, or number groups of objects. For example, one could number all individuals who attended college “5” and all individuals who only attended high school “4.” Notably, the 5 has no meaning other than the fact that it is a different number than 4 in this illustration.

The ordinal level of measurement uses numbers to rank or order individual objects according to a rule. The student with the highest grade on an exam could be assigned “1” and the student with the second highest score could be labeled “2” and so on. Notice, however, there is no information on [Page 1024]the distance between each rank, so the top student might have earned 96% correct and the second score may have earned 83% correct (and third 82%) but that information is not captured when merely ranking students by grades.

The third level of measurement is the interval level of measurement, which retains the qualities of the first two levels of measurement with additional consideration given to the distance between values. Interval scales do not have an absolute zero but the distances between scores is the same amount. One could say there is equality of intervals. Temperature is the most commonly cited example of an interval scale; 48° is 2° warmer than 46° but zero does not represent the absence of temperature. Zero on a thermometer is merely another measure of temperature in degrees.

...