4. Analyze results or data

You can easily tabulate the values, parameters and results using office software (Excel, for example) and try to draw a relationship curve.

If you model this relationship by adding a graph, you'll find what's known as a "trend curve" - in other words, the curve connecting the theoretical points closest to the experimental points, according to a chosen equation model. The process is so simple that it's sometimes tempting to choose a very complicated equation model, which will result in an ideal, sometimes (too) perfect model!

Over-ideal correlation coefficients often mask the fact that we no longer have enough degrees of freedom to assess whether the chosen model is the right one. Most of the time, it's better to want to explain only 95 or 98% of variations with 3 parameters or 3 components than to explain 100% with 10 terms.