Levels of Measurement

Posted May 3, 2017

There are numerous ways to describe and analyze your data, depending upon their level of measurement. The level of measurement of your variable describes the nature of the information that the variable provides. There are two main types of variables: categorical and continuous.

Categorical variables are those that have discrete categories or levels. Categorical variables can be further defined as nominal, dichotomous, or ordinal. Nominal variables describe categories that do not have a specific order to them. These include ethnicity or gender. To remember what type of data nominal variables describe, think nominal = name. Dichotomous variables are categorical variables with two levels. These could include yes/no, high/low, or male/female. To remember this, think di = two. Ordinal variables have two are more categories that can be ordered or ranked. For example, a variable with response data that ranges from I strongly disagree to I strongly agree would be considered ordinal. Keep in mind that researchers may sometimes treat ordinal variables as continuous if they have more than five categories. To remember this variable type, think ordinal = order.

Continuous variables are measured numerically, and have an infinite number of possible values. For example, an age variable measured continuously could have a value of 23.487 years old—if you wanted to get that specific! A continuous variable is considered ratio if it has a meaningful zero point (i.e., as in age or distance). A continuous variable is considered interval if it can be measured along a continuum that has fixed values between two points, but does not have a meaningful zero-point (e.g., temperature measured in Fahrenheit or Celsius).

The level of measurement of your variables influences what analyses you can conduct. The table below presents some example combinations of levels of measurement, and the suggested analysis to conduct.


Independent Variable Level Dependent Variable Level Analysis
Dichotomous Continuous Independent Samples t-Test, Linear Regression
Nominal or Ordinal Continuous ANOVA
Continuous Continuous Linear Regression, Pearson’s Correlation
Continuous or Categorical Dichotomous Binary Logistic Regression
Continuous or Categorical Ordinal Ordinal Logistic Regression
Categorical Categorical Chi Square

Pin It on Pinterest

Share This