matching

Last reviewed 01/2018

Matching is a procedure used to ensure that samples being contrasted are similar with respect to characteristics associated with the factors being studied. Thus matching adjusts for confounding effects, and involves the use of constraints in the selection of comparison groups.

There are two types of matching:-

  • continuous variable, for example age, BP, where a specification of a cut off is required
  • categorical variable, for example sex or race.

Note that stringent matching criteria result in difficulty in matching - the more precisely two groups must be matched, the harder it is to find suitable members of the group.