Mass Communication Research
Determining Levels of Measurement
Rules of Thumb
INDEX SYLLABUS SCHEDULE e-MEDIA COMM-STOP

Determining the level of variable measurement is an essential step for choosing the appropriate descriptive and inferential statistics. While some people find understanding levels of measurement difficult, the rules of thumb provided here may help.

Rules of thumb provide quick and easy ways to make decisions without going through rigorous procedures. The rules listed below will generally work, but they can lead to errors. Knowing these rules of thumb should not be considered an adequate substitute for learning the principles involved in distinguishing levels of measurement.

  1. Assume the following are measured at the nominal level:

    • Dichotomies (variables with only two values), e.g., gender, answers to yes-no questions.

    • Most variables with four or fewer values, e.g. year in college

    • Variables that take on more than four values but have no logical order, e.g., college major, type of residence with response categories such as dorm, fraternity house, university apartment, private apartment, private home.


  2. Assume the following are measured at the ordinal level:

    • Variables that specify a rank or standing, e.g. rankings of college football teams.

    • Variables made by collapsing ratio variables into a few broad categories,

      e.g. age in categories such as
      • 21-35
      • 36-65
      • over 65;

      or income in categories like
      • less than $30,000 per year
      • greater than $30,000 but less than $60,000
      • greater than $60,000 but less than $90,000
      • greater than $90,000 per year.


  3. Assume the following to be measured at the interval level:

    • Questionnaire items with five or more categories that have a clear logical order such as items that ask about levels of agreement with a statement using the categories strongly agree, agree, neutral, disagree, strongly disagree. (Note this scale could have seven categories by adding very strongly agree and very strongly disagree.)

    • Questionnaire items that ask respondents to rate things on specific numeric scales — usually 10-point scales, but there are other possibilities.

    • Questionnaire items that ask respondents to evaluate things on 100-point "feeling thermometers."


  4. Assume the following are measured at the ratio level:

    • Measures of variables that are used in the natural sciences such as:
      • Weight in units like pounds and kilograms
      • Height in units like inches or meters
      • Distance in units like miles or kilometers
      • Time in units like hours or minutes
      • Age in years

    • Money in units like dollars or pesos.

    • Measures with a clear order and meaningful zero point, e.g. Grade Point Average.

If you don't understand something in this Web note, please e-mail Dr. Sitton.

INDEX SYLLABUS SCHEDULE e-MEDIA COMM-STOP
©M. Mark Miller & Ronald W. Sitton 2009
Revised 092811 — http://www.uamont.edu/FacultyWeb/sitton/crz/mrea/lomrot.html