|
The Kim and Burkart response (1) to a letter (2) discussing the visualization of statistical concepts includes the short statement, “one often tends to use the origin point (0,0) in the data. However, whether that is best practice or not is entirely arguable”. Let us get this straight. It is never, ever defensible to invent data and add it to actual observations in fitting or data analysis. I am sure none of your correspondents would countenance it for a second in any other context; why should they do so here?
The argument that a zeroed instrument is expected to provide a point at (0,0) is specious and misleading. What the experimenter expects the experiment to do is not—ever—an acceptable substitute for the observation itself. An actual observation at zero concentration on a zeroed instrument is practically never zero except by chance or censoring (e.g., by the instrument or analyst reporting zero if the response is below some arbitrary threshold—in which class I include the detection limit).
If you want a data point at concentration zero, get a data point for concentration zero. Do not invent it, and do not pretend that inventing it is even remotely good practice, even—or especially—in front of your students.
Literature Cited
- Kim, Myung-Hoon; Burkart, Maureen. J. Chem. Educ. 2008, 85, 635.
- de Levie, Robert. J. Chem. Educ. 2008, 85, 635.
See the authors' reply. |
