Which description best fits an ordinal variable?

Which description best fits an ordinal variable?

• A. Data that convey exact numerical differences between values.
• B. Data that categorize items with no inherent order.
• C. Data that indicate order of categories but not the magnitude of differences.
• D. Data measured on a ratio scale.

Answer: (C)

The key idea is that ordinal data show an order among categories but do not specify how large the difference is between categories. You can tell which category comes first, second, etc., but you can’t quantify the exact gap between them. An example is a rating scale like 1 to 5 (strongly disagree to strongly agree) or a ranking (first, second, third). Because the distances aren’t known or assumed to be equal, you should rely on order information rather than calculating means or standard deviations, and use nonparametric methods if you’re comparing groups. The description in the correct choice captures exactly that: it indicates order of categories without stating the magnitude of differences. By contrast, data that show exact numerical differences belong to interval or ratio scales; data with no inherent order are nominal; and a ratio scale involves true zero and meaningful ratios, which is a different level of measurement.

The key idea is that ordinal data show an order among categories but do not specify how large the difference is between categories. You can tell which category comes first, second, etc., but you can’t quantify the exact gap between them. An example is a rating scale like 1 to 5 (strongly disagree to strongly agree) or a ranking (first, second, third). Because the distances aren’t known or assumed to be equal, you should rely on order information rather than calculating means or standard deviations, and use nonparametric methods if you’re comparing groups.

The description in the correct choice captures exactly that: it indicates order of categories without stating the magnitude of differences. By contrast, data that show exact numerical differences belong to interval or ratio scales; data with no inherent order are nominal; and a ratio scale involves true zero and meaningful ratios, which is a different level of measurement.