What is Likert scale? Why are they popular in business and other social sciences?
Likert scale refers to a psychometric response scale that is used in survey instruments to obtain research subject’s degree of agreement with every statement in the questionnaire or survey instrument. According to Bertram (n.d.), Likert scales measure a single trait. That is, they are unidimensional. In each of the statements or items making the questionnaire, the respondents are required to state their level of agreement with each of the statements or items by way of an ordinal scale. Likert scales is named after its inventor, Dr. Rensis Likert, the US organizational behavior psychologist and a sociologist at the University of Michigan (Chowdhury, 2014).
These scales enable researchers to ascribe quantitative values to qualitative data, thus making it possible to carry out statistical data analysis (Chowdhury, 2014). The most commonly used Likert scales are 5-point, with scores ranging from 1 (“Strongly Disagree”) to 5 (“Strongly Agree”). However, some researchers used 7-point or 9-point scales. Likert scales are popular in business and other social sciences because of the benefits associated with their use. First, Likert scales are popular because they are capable of measuring the attitude of the respondents easily (Subedi, 2016). Secondly, using a Likert scale, researchers can easily make statements that measure the aspects of a given construct. Thirdly, this type of scale can be easily understood by the respondents hence making the researchers to effortlessly collect research subjects’ data.
Another benefit linked to the use of the Likert scale is its ability to measure many constructs (Revilla, Saris, & Krosnick, 2014). Moreover, it is easy and quicker to administer Likert scale items because the researcher needs to only explain to the respondents how to complete the survey instrument once. Because of this, Likert scales are less expensive to administer (Revilla et al., 2014).
What is the author’s issue with the way Likert scales are used?
In the article titled “Likert scales: how to (ab)use them,” Jamieson (2004) provided justifications for why he believes Likert scales are not excellent measures of constructs. First, the author pointed out that even though responses to Likert scales items are ordinal, there is no guarantee that the intervals between values are equal. Because of this, Jamieson (2004) argued that the descriptive statistics that are usually used to analyze data might not be appropriate for ordinal data. Additionally, the inferential statistical analysis conducted on Likert scale data may be inappropriate for addressing the hypotheses because the data is usually assumed to be interval rather than nominal.
Why do social sciences use Likert scales as interval scales? Do social sciences use parametric or non-parametric tests on Likert scale data? Is this okay?
In social sciences, Likert scales are generally treated as interval scales rather than ordinal scales to enable the researchers to carry out statistical analysis (especially those involving significance testing) that would otherwise be impossible if the scale was regarded as ordinal. Social sciences use parametric tests on Liker scale data instead of non-parametric tests because the data is viewed as interval rather than nominal. Using parametric tests is not okay because Likert scales are not measured on the interval level of measurement. That is, there is no guarantee that the intervals between values are equal. Consequently, social researchers should use parametric tests instead.
- Bertman, D. (n.d.). Likert scales. Retrieved from pages.cpsc.ucalgary.ca/~saul/wiki/uploads/CPSC681/topic-dane-likert.doc
- Chowdhury, M. S. H. (2014). Forest conservation in protected areas of Bangladesh: Policy and community development perspectives. Springer.
- Jamieson, S. (2004). Likert scales: how to (ab) use them. Medical Education, 38(12), 1217–1218.
- Revilla, M. A., Saris, W. E., & Krosnick, J. A. (2014). Choosing the number of categories in agree–disagree scales. Sociological Methods & Research, 43(1), 73–97.
- Subedi, B. P. (2016). Using Likert type data in social science research: Confusion, issues and challenges. International Journal of Contemporary Applied Sciences, 3(2), 36–49.