Imbalance in MZ and DZ pairs

1 reply [Last post]
Sabha's picture
Offline
Joined: 05/18/2010

I am looking for advice on how to deal with huge imbalance in no of MZ and DZ pairs prior to estimating A,C and E by univariate Twin ACE model. My data set consists of 2.5 times more number of MZ pairs compare to DZ pairs.

Are there any methods available to normalize or remove the bias.

Thanks in advance.

neale's picture
Offline
Joined: 07/31/2009
Rule of 2/3

This is pretty standard in studies of volunteer samples (2:1 ratio of MZ:DZ). There is no issue of bias to worry about UNLESS the propensity to volunteer is related to the trait you are studying. If it is, some reduction in bias can come from making sure that you include discordant pairs in the analysis (one volunteered and one did not).

This is not the greatest paper on the topic, but it's perhaps better than nothing:


Neale, M.C. & Eaves, L.J. (1993) Estimating and controlling for the effects of volunteer bias with pairs of relatives. Behavior Genetics 23: 271-278.