Abstract
The use of mathematical transformations to reduce non-linear functions to linear problems, which can be tackled with analytical linear regression, is commonplace in the chemistry curriculum. The linearization procedure, however, assumes an incorrect statistical model for real experimental data; leading to biased estimates of regression parameters and should therefore not be used in formal data analysis. This fact is overlooked in many chemistry degrees, students do not yet have the mathematical knowledge to appreciate why linearization leads to bias when it is introduced. I hope that this commentary will start a discussion around the place of linearization in the chemistry curriculum, and more broadly around how mathematical and statistical training is currently provided to chemistry students.
| Original language | English |
|---|---|
| Pages (from-to) | 4174-4176 |
| Number of pages | 3 |
| Journal | Journal of Chemical Education |
| Volume | 100 |
| Issue number | 11 |
| Early online date | 9 Oct 2023 |
| DOIs |
|
| Publication status | Published - 14 Nov 2023 |
Bibliographical note
Funding Information:The author thanks Benjamin J. Morgan, Samuel W. Coles, Thomas Holm Rod, Gabriel Krenzer, and Kasper Tolborg for the insightful discussion that led to this work. Additionally, the author would like to thank those that engaged in discussion on Twitter, in particular Carl Poree and Fiona Dickinson, when the problem of linearization in Arrhenius modelling was initially raised.
Publisher Copyright:
© 2023 The Author. Published by American Chemical Society and Division of Chemical Education, Inc.