# Interval scale

Interval scale | |
---|---|

See also |

**Interval scale** is a scale of measurement for a variable in which the interval between observation is expressed in terms of a fixed standard of measurement (J.K.Sharma 2012, p.14).

J.K.Sharma says also that, an **interval scale** lets us perform serial arithmetical operations on the information collected from the respondents. While the nominal scale only allows us to qualitatively, exhaustive sets, the ordinal scale lets us to rank-order the preferences, and the interval scale allows us to calculate the mean and the standard deviation of the information on the variables. To put it differently, *the internal scale not only classifies individuals according to some categories and determines the order of these categories; it also measures the magnitude of the differences in the preferences between the individuals*(J.K.Sharma 2012, p.14).

## Characteristics of interval scale

An Interval Scale main qualities:

- An interval scale contains all the features of an ordinal scale, and, what is more, allows the researches to compare the variations between objects. The values in the interval scale show us how far apart the objects are with respect to a specific characteristic. The difference between the scale values 1 and 2 is the same as the difference between the scale 2 and 3.What is more, the difference between values 2 and 4 is twice the difference of that between 2 and 3(1 and 2, etc.)(J.P.Neeklankavil 2015, p.197).
- In interval scales, the subjective votes from respondents are translated into valuable and quantitative information. By designing an interval, a researcher can achieve a higher level of measurement than with ordinal scales. Sometimes the researcher has to impose equal intervals between descriptors. The researcher would assume that the variation between one descriptor and the next one on a scale using "strongly agree", "agree", "neutral", "disagree", "strongly disagree" as one unit. The values assigned to this set of responses run from 1 to 5.Respondents will treat the differences between adjacent response categories as equal(J.P.Neeklankavil 2015, p.198).

## Location of starting point

In an **interval scale**, the place of starting point(zero) is not fixed. The zero and the units of measurement units are arbitrary.
For example(J.P.Neeklankavil 2015, p.198):

- Temperature measurement for a day in a city ranged between 20°Fahrenheit(F) and 40°F. From these values, can we infer from these measurements that the high temperature of 40°F was twice as hot as the low temperature of 20°F? The answer is no. This can be proven by translating the temperature from Fahrenheit to centigrade(C). The low temperature of 20°F is equal to-7°C, and the high temperature of 40°F is equal to 4°C. The high temperature of 4°C is not as twice as hot as the low temperature of -7°C.

## References

- Ayyub B., McCuen R. (2011),
*Probability, Statistics, and Reliability for Engineers and Scientists* - Neeklankavil J. (2015),
*International Business Research*, M.E.Sharpe, p.197-198, New York - Protonotarios E., Baum B., Johnston A., Hunter G., Griggin L. (2014),
*An absolute interval scale of order for point patterns*, Royal Society - Sharma J. (2012),
*Business Statistics*, Dorling Kindersley(India)Pvt. Ldt, p.14 - Sreejsh S., Mohapatra S., Anusree M. (2013),
*Business Research Methods: An Applied Orientation*, Springer, New York

**Author:** Szymon Olejniczak