Volume 14, Issue 2 (March 1986)
A Better Way to Present Results from a Least-Squares Fit to Experimental Data: An Example from Microhardness Testing
Fitting a straight line to data using the least-squares method is very often the method of choice when presenting experimental results. When there are a number of sets of data, each with its fitted line, it can be difficult at first to appreciate the meaning of the results. Also, parameters calculated from least-squares fits are often quoted without estimates of accuracy, even when the original data contain considerable scatter.
A method is described here that aids the assimilation of a number of least-squares fits and takes full account of estimates of accuracy.