Misuse of Linear Regression Technique in Analytical Chemistry?

Abstract

Nowadays, frontiers among different sciences are revealed as diffuse, and as a consequence, research must necessarily be faced from an interdisciplinary approach. Similarly, teaching certain topics in Chemistry requires the consideration of developments in other sciences (Mathematics, Biology, Physics, etc.). For instance, the estimation of the parameters of calibration lines in Analytical Chemistry (via linear regression) exemplifies this mandatory interaction since the comprehension of the regression tools should also consider how estimation is regarded from a statistical perspective. This work focuses on how to overcome the contradictions that undergraduate Chemistry students may encounter between the chemists' and statisticians' perspectives, when they are lectured on least-squares linear regression. A mixed analytical chemistry−statistics approach is proposed to present a method to get over the discordant issues between both scientific viewpoints.