The Types of Scales in Statistics

Odds are you last thought about statistical scales in college.

Unless you have a career as a researcher, you are likely only concerned with scales used in statistics at the college level. Coursework at this level deals with research methods and data analysis. If you have ever taken a survey, though, you have revisited the subject, even if you never gave it a second thought. Researchers collect data using different scales, each of which allows for different types of analysis.

Nominal

Nominal data, according to Professor Jason T. Newsom of Portland State University, mean nothing from a quantitative standpoint. When you use the nominal scale, you seek to classify or categorize data. Examples of nominal variables include sex or gender, political party affiliation and hair color. When researchers assess such data, they often assign numbers to a particular response, but these numbers mean nothing. They are used because computer programs require numbers for analysis rather than words. For instance, a researcher may code all male responses as "0" and all female responses as "1." As Professor Emeritus David W. Stockburger of Missouri State University notes, calculating statistics, such as mean or standard deviation from nominal data, proves meaningless.

Ordinal

Ordinal scales possess the same qualities as nominal data except, as Dr. Newsom notes, they "have an evaluative connotation." Newsom uses job satisfaction as an example of an ordinal measure. If you assess job satisfaction on a scale from 1 to 10, you know that a score of 10 is better than a score of 9, a score of 9 is better than a score of 8, and so on. You do not know by how much 10 is better than 9 or if the distance between the two is the same as the distance between 9 and 8.

Interval

Celsius or Fahrenheit temperature is interval data.

Interval measures provide the same information as variables on the ordinal scale, but the distance between values is fixed. With interval variables, the distance between 10 and 9, the distance between 9 and 8, and so forth is known as well as equal. Newsom uses Celsius or Fahrenheit temperature as examples of interval scale data. The difference between 10 degrees and 20 degrees is the same as the difference between 90 and 100 degrees.

Ratio

Weight is measured on the ratio scale.

Ratio level data share the same qualities as data on the interval scale, except ratio variables have an absolute zero. As Newsom points out, unlike Celsius or Fahrenheit temperature, measuring temperature on the Kelvin scale is at the ratio level. There is no such thing as below-zero on the Kelvin scale. Weight is another common example of data on a ratio scale. Weight has an absolute zero. Something or somebody cannot have a weight of less than zero.

References

• Missouri State University: Measurement
• Portland State University: Types of scales & levels of measurement