Regression Point Displacement design

Explanations > Social Research > Design > Regression Point Displacement design

Description | Example | Discussion | See also

Description

Use this when measures already exist for a significant number of pre-existing subjects or groups.

Take a single group and apply the treatment with pre- and post-testing as usual. Then compare the result with the measures available for

Example

A standard treatment for addiction is known to result in a small reduction in a craving index, that is reported by city districts. A new treatment is tried and is shown on the before-and-after graph, where it is dramatically moved from the normal regression line, showing it has made a significant difference.

Discussion

This method is particularly useful when doing experiments with large coherent groups, such as communities or companies. It is necessary, however that significant other measures are available from comparable groups (that form the control) to allow for a good comparison.

Before-and-after measures are useful for showing change. The general principle, however, of comparing one item against a critical mass can be applied elsewhere.

Note that the single measure does have to be constituted from a sufficiently large sample to significant. Results from a single person may be surprising, but is unlikely to be significant.