1. Likert scale data is a set of ordered categories, not truly quantitative.
2. There is debate over whether parametric statistical procedures can be used with Likert scale data.
3. Researchers should proceed with caution and consider non-parametric alternatives or strict criteria for using numerical procedures with Likert scale data.
The article "Can Likert Scale Data ever be Continuous?" by The Analysis Factor provides a comprehensive overview of the debate surrounding the use of Likert scale data in parametric statistical procedures. However, the article has some potential biases and limitations that need to be addressed.
One-sided reporting is evident in the article, as it primarily presents arguments in favor of using Likert scale data in parametric tests. While it acknowledges that some researchers argue against this practice, it does not provide an equal representation of both sides of the debate. This bias may lead readers to believe that using Likert scale data in parametric tests is more widely accepted than it actually is.
The article also makes unsupported claims, such as stating that "tests that assume real numerical data still tell you a lot about what’s going on with this variable." This claim lacks evidence and may mislead readers into thinking that using parametric tests on Likert scale data is always valid.
Additionally, the article misses some points of consideration, such as the impact of sample size on the validity of using parametric tests on Likert scale data. Small sample sizes can lead to inaccurate results when using these tests, which should be noted as a potential limitation.
The article also lacks evidence for some claims made, such as stating that Lubke & Muthen (2004) found true parameter values in factor analysis with Likert item data under certain conditions. The article does not provide details about these conditions or how they were met, leaving readers without sufficient information to evaluate this claim.
Unexplored counterarguments are also present in the article. For example, while it acknowledges that some researchers argue against using parametric tests on Likert scale data due to unequal intervals between categories, it does not address how this issue could affect results or why non-parametric tests may be more appropriate.
Promotional content is evident in the recommendations section of the article, where The Analysis Factor promotes its own training on logistic regression and categorical data analysis. This promotional content may undermine the credibility of the article and lead readers to question its objectivity.
Overall, while the article provides a useful overview of the debate surrounding the use of Likert scale data in parametric tests, it has some potential biases and limitations that need to be addressed. Readers should approach the article with caution and seek additional sources to gain a more comprehensive understanding of this topic.