Likert Scale Definition
Named after its inventor, Rensis Likert, the Likert scale is a tool used by researchers in questionnaires to determine responses by rating some items. It is commonly used interchangeably with the rating scale though there are other types of rating scales. This is because it is the generally used approach in survey research.
A Little More on What is the Likert Scale
With this psychometric scale, respondents are asked to rate items on a level of agreement, as follow:
• Strongly agree
• Strongly disagree.
The scale uses different alternatives on themes like agreement, frequency, quality and importance. For example
• Agreement: strongly agree to strongly disagree
• Frequency: often to never
• Quality: very good to very bad
• Likelihood: definitely to never
• Importance: very important to unimportant.
These items are called the Likert Scale Response Anchors
Steps to Developing a Likert Scale
- Define the focus – the focus of whatever topic to be measured should be uncomplicated. For example, ‘Customer Service’ or ‘This Website’.
- Generate the Likert Scale items – The items to be put on the scale should ratable. For example, polite/rude could be rated as “very polite”, “polite”, “not polite” or “very impolite.” Politeness could also be rated on a scale of 1 to 10, where 1 is not polite at all and 10 is extremely polite.
- Rate the Likert Scale items – If one want to be sure his focus is good, a previous rating should be carried out on items to be put on the scale in order to weed out the items that are mostly seen as unfavorable.
- Administer the Likert Scale test.
The Likert scale data is usually classified as an ordinal variables, as it is usually difficult to find the mean and average from items such as “agree”, “disagree”, and “neutral” because you don’t know the distance between the data items. This is because the distance between variables must be constant.
Statistics one can use are:
- The mode: the most common response.
- The median: the “middle” response when all items are placed in order.
- The rangeand interquartile range: to show variability.
- A bar chartor frequency table: to show a table of results.
References for Likert Scale
Academic Research on Likert scale
Likert scales and data analyses, Allen, I. E., & Seaman, C. A. (2007).Quality progress, 40(7), 64-65. The article discusses the use of the Likert scales and the classification of data gotten from surveys into four levels; nominal, ordinal, interval and ratio data. It also points out that analysis of ordinal data in relation to the Likert scale is not straight forward or transparent like the others it was compared to.
Likert scales, levels of measurement and the “laws” of statistics, Norman, G. (2010). Advances in health sciences education, 15(5), 625-632. The paper is an analysis of the argument for and against the criticism of statistical methods, especially parametric statistics. This is because the use of various parametric methods such as analysis of variance, regression, correlation are faulted. It goes ahead to show that parametric statistics are robust with respect to violations of these assumptions and can be utilized without any concerns. This means that issues like (a) the sample size is too small, (b) the data may not be normally distributed, or (c) The data are from Likert scales are of no basis.
Analyzing likert data, Boone, H. N., & Boone, D. A. (2012). Journal of extension, 50(2), 1-5. This article is based on a discussion of the differences between Likert-type and Likert scale data and how their analysis needs to be carefully and correctly carried out to prevent mistakes from occurring. This information is mostly directed towards Extension professionals as once the difference is understood, the decision on appropriate statistical methods will become obvious and easy to make .
Likert items and scales of measurement, Brown, J. D. (2011). Statistics, 15(1), 10-14. This article discusses and examines the various scales of measurement in a bid to explain the Likert scale. It also makes it clear that there is a difference between Likert scale and Likert items.
Analyzing and interpreting data from Likert-type scales, Sullivan, G. M., & Artino Jr, A. R. (2013). Journal of graduate medical education, 5(4), 541-542. The article examines the various methods of analyzing and interpreting data from Likert-type scales in relation to medical education and medical education research.
What issues affect Likert–scale questionnaire formats, Brown, J. D. (2000). Shiken: JALT Testing & Evaluation SIG Newsletter, 4(1). The author discusses issues that affect Likert-scale questionnaire formats and evidence that Likert scales can be as effectively analyzed as interval scales. An examination of the myths surrounding the Likert scale, the counter argument and antidotes for them is also carried out.
Statistics analysis and fuzzy comprehensive evaluation of Likert scale [J], Lai-bin, Q. I., & Kete, L. (2006). Shandong Science, 2, 005. This paper discusses the use of conventional statistic analytical method combined with fuzzy comprehensive evaluation to analyze Likert scale data. The feasibility and validity of this method is proven by its application on an applied example.
The effect of nonlinear transformations on a Likert scale, Baggaley, A. R., & Hull, A. L. (1983). Evaluation & the health professions, 6(4), 483-491. Due to the normality of the distribution being clouded as a result of not understanding the structure of the distances between the scale points, there is a need to develop an alternative. The alternative was in form an empirical approach using responses to a clinical performance evaluation instrument that used the four-point behaviorally-anchored scale. The journal discusses the use of combinations of nonlinear transformations on the Likert scale and its effects. It goes ahead to suggest that parametric statistics can be applied to behaviorally-anchored rating scales and that the nature of the distance needs to be understood.
An empirical study on the transformation of Likert–scale data to numerical scores, Wu, C. H. (2007). Applied Mathematical Sciences, 1(58), 2851-2862. This paper presents an estimation method for transforming Likert-scale data into numerical scores that follow the assumption of normality and its effects based on the need for a data being analyzed to follow some particular assumptions. The example of the AVOVA assumption is used here, though such assumptions are not usually observed by data collected through Likert scales. Also the paper addresses the decision on whether or not Likert-scale data should be transformed to scores that are more compliant to statistical assumptions. This is based on the scaling procedure proposed by E. J. Snell.
Can Likert scale data ever be continuous, Grace-Martin, K. (2008). Article Alley. This article considers the argument regarding the validity of the use of the Likert scale data in parametric statistical procedures that require interval data.
Fuzzy vs. Likert scale in statistics, Gil, M. Á.,& González-Rodríguez, G. (2012).(pp. 407-420). Springer, Berlin, Heidelberg. The journal using examples, gives a guideline on designing questionnaires allowing free fuzzy-numbered response format. This is because the fuzzy numbers are very expressive, describing the usual answers in this context in a friendly way. It also looks at some of the techniques that can be used in analyzing the responses.
Likert data: what to use, parametric or non-parametric?, Murray, J. (2013). International Journal of Business and Social Science, 4(11). Determining the type of statistical test, parametric or non-parametric, to be conducted on Likert scale data and its effect on the conclusions is the aim of this paper. Analysis conducted on actual scale data showed some positive relationships.