While technically individual matching can better improve efficiency ( sample size required  to detect given effect with given power), it may be difficult to select sufficient controls in some circumstances. Frequency matching is simpler to do and analyse and less likely to lead to overmatching.

in any case matching should only be used for strong confounders eg age,sex,ethnicity.

see Breslow and  Day vols 1,2.

This is so clear and helpful. Thank you very much James.

In a matched analysis the differences in exposure status among the matched sets are being used. In a pair-matched study the pairs with the same exposure status do not contribute, which is a pity. In a frequency-matched study this will not occur so quickly.

Dear friends

Individual matching is a method of controlling a priori a confounding factor when setting up groups to compare. This method consists in associating each case to one or more controls having the same characteristics for the factor or factors to neutralize (a male case, a male control ... a female case, a female control ...).

These factors are neutralized since their distribution is identical in the two groups to be compared and are no longer taken into account in the analysis that may be more difficult (one case for two or more controls).

Individual matching should be distinguished from frequency matching which is a stratification a priori on a factor. In the frequency matching, on the occasion of a case-control study, controls are selected with relative frequency equal to that of the cases for the different modalities of a factor on which stratification was performed. The matching frequency between cases and controls on sex means that there is the same proportion of cases and controls by sex : the proportion of men among the cases is identical to that of controls ( for example 40% of male subjects both among cases than controls and 60% of female subjects both among cases than controls).

Frequency matching on a factor must lead to a stratified analysis on the same factor at the time of analysis.

Thank you all very much for your valuable explanations.

Thank you very much for all

thank you