Fundamentals of Statistics contains material of various lectures and courses of H. Lohninger on statistics, data analysis and here for more.


One phenomenon which can be particularly misleading is autocorrelation. The term serial correlation is sometimes used for autocorrelation. Autocorrelation means correlation between successive values in the data. It mainly occurs when data is measured over time, and the values are not independent of each other. Please note, that most inferential tests and modeling techniques fail if data is autocorrelated.

The figure below shows a data set exhibiting serial correlation. As you can see, serial correlation can be easily recognized just by looking at the data.

In contrast to the figure above, the following figure shows a signal which exhibits no autocorrelation.

In order to get a more objective measure of serial correlations, one should calculate the autocorrelation function.

Last Update: 2012-10-08