Covariance experiments

Explanations > Social Research > Design > Covariance experiments

Description | Discussion | See also

Description

Covariance experiments use more than one variable to reduce noise.

The principle is to measure the same variable in different contexts, such as before and after a single test and across tests.

This allows for 'adjustments' and 'removing the effect' of unwanted other variables. Although this is in principle a subtraction, in practice it uses a statistical method.

Discussion

When an experiment is performed, the results may vary significantly. However, this might be due to many additional variables as well as the main variable being changed. But by measuring a 'before and after', the variation caused by the experiment can be eliminated, as in the simple example below.

For more complex cases across groups a more statistical method is used. Analysis of covariance is sometimes called ANACOVA.