Meta-analysis

A meta-analysis is a statistical procedure to combine a number of existing studies. Through such a procedure, effects which are hard or impossible to discern in the original studies because of a too small sample size[?] can be made visible, as the meta-analysis is (in the ideal case) equivalent to a single study with the combined size of all original studies. A weakness of the method is that problems with any of the studies will affect the result of the meta-analysis, so a good meta-analysis of bad studies will still result in bad data.

See also:

• Simpson's paradox
• Selection bias[?]