Analysing Non-Linear Data - This page is in the process of development

Analysing a linear (straight-line) relationship is relatively straight-forward using the techniques of linear regression and calculations of slope and intercept (see analysing a straight line).
However, the method required to analyse a non-linear data relationship depends on the particular form of that relationship.
It is important to be aware of the most common methods available so that it is possible to assess the best choice of method for a given problem. In some cases it is necessary to use a combination of more than one method.

Common methods used to analyse non-linear data:

This section provides short video clips, tutorials and examples - browse the different methods to find the one that applies to your particular problem.  

1. Linearise the data using a theoretical relationship for the data

2. Perform a numerical analysis using a theoretical relationship for the data 

3. Perform a linear analysis on a selected section of the data 

  • Avoids curvature and non-linearity due to second order effects

4. Case studies

  • Common examples, e.g. Arrhenius equation, Beer’s law
     
  • Complex analytical problems

Uncertainties / errors in non-linear regression:

This section develops the handling of experimental errors, which can become distorted in non-linear analysis.

  • Calculation of uncertainties in linear and non-linear regression
  • Weighted least squares regression using Solver in Excel

QVA Tutorials (questions + video answers):

These short quizzes (5 - 10 questions each, with video feedback) provide a quick way of checking and increasing your knowledge and real understanding of the different forms of analysis.