What is Meta-Analysis?
Combining results from previous studies for further analysis in a new context. Can find similarities or differences within the group. Usually weighted by subject size. Can be used in any field, like a generic non-neuro example of reading through pharmacology publications to determine if reported side effects are truly related to a treatment or just coincidence or hypochondria.
Recent Nature publication
Meta-analysis and the science of research synthesis
Benefits
Provides greater power: With a single subject, it’s hard to know if an effect is true or just a coincidence. Multiple-subject studies, usually with tens of subjects, allow you average away the noise and get a stronger result. A meta-analysis including data derived from hundreds or thousands of subjects will also have a great effect.
Provides greater generalizability: there are known differences from age, sex, handedness, native language… There could be additional unknown external variables that cannot be explicitly controlled for. The more subjects you include, the more you average away these influences. (Hopefully! Unless you’re adding more data with the same biases as the original.)
Pitfalls
Study selection can introduce bias. Objective inclusion criteria and comprehensive searching are required. (Also, see more in Turkeltaub 2012 section.)
