Data Transformations for Inference with Linear Regression: Clarifications and Recommendations
DOI
https://doi.org/10.7275/2w3n-0f07
Abstract
Data transformations have been promoted as a popular and easy-to-implement remedy to address the assumption of normally distributed errors (in the population) in linear regression. However, the application of data transformations introduces non-ignorable complexities which should be fully appreciated before their implementation. This paper adds to existing Practical Research and Assessment Evaluation (PARE) publications on data transformations by providing a broad overview underlying the use of data transformations for the specific purpose of statistical inference and interpreting meaningful effect sizes. Data transformations not only potentially change the scale of the transformed variable; they also alter the fundamental relationships among variables while simultaneously changing the distribution of the errors. Given these repercussions, we clarify the nature of certain data transformations and strongly recommend the use of data transformations when they can enhance the interpretation of effect sizes. Accessed 5,515 times on https://pareonline.net from October 11, 2017 to December 31, 2019. For downloads from January 1, 2020 forward, please click on the PlumX Metrics link to the right.
Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial-No Derivative Works 4.0 International License.
Recommended Citation
Pek, J.; Wong, O.; and Wong, A. C.
(2019)
"Data Transformations for Inference with Linear Regression: Clarifications and Recommendations,"
Practical Assessment, Research, and Evaluation: Vol. 22, Article 9.
DOI: https://doi.org/10.7275/2w3n-0f07
Available at:
https://scholarworks.umass.edu/pare/vol22/iss1/9
Comments
https://doi.org/10.7275/2w3n-0f07