Matching (statistics)

Please help to eliminate the shortcomings of this article, and please take part you in the discussion! (Insert item)

Matching is a method that is applied in the design of observational studies. In case -control studies, cases and controls can be chosen so that they match with respect to certain criteria ( eg, gender, socioeconomic status, age group ). This is at the individual level ( for each case is a matching in which / the selected criteria selected control ) or as a group matching ( frequency matching) possible. When groups matching the composition of all the controls according to the composition of cases is selected. For example, if the group of cases to 80% are women, the control group is put together with a similar percentage of women.

In individually matched case -control studies, which are analyzed using logistic regression, a special form of this method ( conditional logistic regression ) should be applied. When it is matched individually to a number of factors, there is a risk that, for cases of no checks are found that match according to the matching criteria.