Putting Measurement to Work- social research and statistics

Measurement means classifying units of analysis by categories to represent a variable concept. Measuring concepts in quantitative research involves conceptualizing, or clarifying, what we mean by them, and then operationalizing them, or defining the specific ways we will make observations about the concepts in the real world. Measurement in qualitative research often involves making observations, coding those observations into categories, and then looking for relationships among the categories that might constitute usable variables.

LEVELS OF MEASUREMENT

Nominal (from the Latin, no men or name)level variables –

are variables whose categories have names or labels. Any variable worth the name is nominal. If your variableis “cause of death” and your categories are “suicide,” “homicide,” “natural causes,” “accident,” and “undetermined,” your variable is a nominal level variable.

Ordinal (from the Latin, ordinalis, or order or place in a series) level variables are variables whose categories have names or labels and whose categories can be rank-ordered in some sensible fashion. If one measured social class using categories like “upperclass,” “middleclass,” and “lower class,” one category (say “upper class”) could be seen to be “higher” than the others.

Interval level variables-

Are those whose categories have names, whose categories can be rank-ordered in some sensible way, and whose adjacent categories are a standard distance from one another.

EX. Because of the constant distance criterion, the categories of interval scale variables can be meaningfully added and subtracted. SAT scores constitute an interval scale variable, because the difference between 550 and 600 on the math aptitude test can be seen as the same as the difference between 500 and 550. As a result, it is meaningful to add a score of 600 to a score of 500 and say that the sum of the scores is 1100.

Ratio level variables

Are those whose categories have names, whose categories can be rank-ordered in some sensible way, whose adjacent categories are a standard distance from one another, and one of whose categories is an absolute zero point—a point at which there is a complete absence of the phenomenon in question.

Ex. Age, in years since birth, is often measured as a ratio level variable (with categories like one, two, three, or four years, and so on) because one of its possible categories is zero. Other variables such as income, weight, length, and area can also have an absolute zero point and are therefore ratio level variables.