Ratio Level of Measurement

The most precise of the four levels of measurement is called the ratio level. The ratio scale is unique because, unlike the interval scale, it contains an absolute rather than arbitrary zero. In other words, a score of 0 on a ratio-level measure indicates a complete absence of the trait or construct being measured. Whereas at a temperature of 0° on a Fahrenheit scale, some molecules are still moving (because the zero of the scale is arbitrary), a reading of 0° on a Kelvin temperature scale signifies a total lack of molecular movement. Therefore, the Fahrenheit scale is an example of interval measurement, whereas the Kelvin scale is considered ratio-level measurement.

The reason the absolute zero makes the ratio scale unique is that it allows us to make meaningful fractions. On the Kelvin scale, 20° is twice as “hot” (has twice the amount of moving molecules) as 10°, and 150° is three times as “hot” as 50°. Using the Fahrenheit scale, we cannot make such a comparison; 20° is not twice as hot as 10° because the zero point on the scale does not indicate a complete absence of molecular movement and because temperatures below 0° are possible.

Whereas ratio-level data are fairly common in physical sciences (zero molecular movement, zero light, zero gravity), they are rarer in behavioral and social science fields. Even if someone scores a zero for whatever reason on an IQ test, the tester would not declare that the test taker has no intelligence. If a student receives 0 points on a vocabulary quiz, the teacher still cannot claim the student has no vocabulary.

An example of a case in which ratio data could be used in the behavioral sciences is if a researcher is using as a variable the amount of practicum hours students have logged during a semester. The researcher would give a score of 0 to any student who has seen no clients and logged no hours. Likewise, the researcher could accurately say that a student who has logged 50 hours of client contact during practicum has had twice the amount of client contact as a student who has logged 25 hours.

Although ratio-level measurement is not common in the behavioral and social sciences, its advantages [Page 826]make it a desirable scale to use. Because it is the most precise level of measurement, and because it contains all of the qualities of the three “lower” levels of measurement, it provides the richest information about the traits it measures.

Figure 1 Progressing Levels of Measurement

Ratio data let us know that Person A, who earns an annual income of $1 million, makes 100 times the amount of money as Person B, who earns an annual income of $10,000. We can tell more from this comparison than we could if we knew only that Person B makes $990,000 less per year than Person A (interval-level measurement), that Person A has a larger income than Person B (ordinal-level measurement), or that the incomes of Person A and Person B fall under different socioeconomic categories (nominal-level measurement). This information would give us a more accurate conception of the construct being measured.

...