Likert scales are widely used in surveys and questionnaires to assess respondents’ attitudes, opinions, or perceptions. However, there has been ongoing debate regarding whether Likert scales should be classified as ordinal data or interval data. This tutorial aims to explore the nature of Likert scales and provide insights into their appropriate treatment in statistical analysis.

**Treating Likert Scales as Ordinal Data**

From a strict perspective, Likert scales are considered ordinal data. The response options in a Likert scale represent an ordered set of categories without a fixed numerical value or equal intervals. The categories typically range from “strongly disagree” to “strongly agree” or similar constructs. The key feature of ordinal data is that it reflects a ranking or hierarchy, but the precise differences between adjacent categories are not defined.

The problem of ordinal data is that in general we do not calculate mean for ordinal data. This is because mean assumes equal intervals between values, which may not hold true for ordinal variables. With that, it makes it impossible to use t-test or ANOVA if Likert scales are considered as ordinal data.

**Treating Likert Scales as Interval Data**

In practice, Likert scales are often treated as if they were interval data, despite their ordinal nature. This treatment is based on certain assumptions.

Firstly, it assumes that the intervals between the response options are equal and carry a consistent quantitative meaning. Secondly, it assumes that respondents interpret and use the scale in a way that reflects interval-level measurements. This means that respondents perceive equal intervals between adjacent response options and consider the numerical values associated with the scale meaningfully.

**Implications of Treating Likert Scales as Interval Data**

Treating Likert scales as interval data allows for a broader range of statistical analyses and calculations. Researchers can utilize parametric statistical techniques, such as t-tests, ANOVA, and correlation analysis, which are typically reserved for interval or ratio data. This approach assumes that the numerical values assigned to the response options reflect meaningful and equidistant intervals.

**Critiques and Considerations**

It is important to acknowledge that the treatment of Likert scales as interval data is not universally accepted. Critics argue that Likert scales lack true numerical properties and may not satisfy the assumptions of interval-level measurement. Furthermore, the interpretation of Likert scales as interval data relies on respondents’ understanding and utilization of the scale in a manner consistent with interval measurements.

**Conclusion**

Likert scales are inherently ordinal data, reflecting an ordered set of response options without fixed numerical values or equal intervals. However, in practice, they are often treated as if they were interval data under certain assumptions. Researchers should carefully consider the nature of their data, research objectives, and the appropriateness of treating Likert scales as interval data when conducting statistical analysis.

## Reference

Can Likert Scales be Treated as Interval Scales?—A Simulation Study