You are working with the text-only light edition of "H.Lohninger: Teach/Me Data Analysis, Springer-Verlag, Berlin-New York-Tokyo, 1999. ISBN 3-540-14743-8". Click here for further information.

Table of Contents Bivariate Data Correlation Confounded Variables	Index
See also: chance correlations

Confounded Variables

Two variables are said to be confounded if the influence of any of the two variables on the outcome of an experiment cannot be separated.
 

An example may clarify this: a research worker at a government lab in a small European country determined the concentration of harmful substances in the air on a busy road in the capital of that country. The measurements have been planned to last for a whole year, in order to get information for all seasons. The experiments started in April, sampling and analyzing the air twice a day. However, in September the sampling device broke and had to be replaced by a new one which was slightly different from the original one.

After having completed the experiments by end of March of the next year, the researcher started to analyse the data. The results clearly indicated different concentrations of the harmful substances for the cold and the warm seasons. However, since the sampling device had to be exchanged in autumn, the researcher did not know whether the difference was due to the different conditions during winter and summer, or due to the replaced sampling device.

In this case there are two factors (or variables) which are said to be confounded with respect to the outcome of the experiment: average daily temperatures, and the effectiveness of the sampling device.
 

Last Update: 2006-Jän-17