Effective sample size in weighted-z meta-analysis

[Wenbin Mei]

Hi plink team,

Thank you for providing such as nice tool! In https://www.cog-genomics.org/plink/1.9/postproc you mentioned "if the numbers of cases and controls are unequal, effective sample size should be 4 / (1/<# of cases> + 1/<# of controls>).". But some papers suggest Neff = 2 / (1/<# of cases> + 1/<# of controls>) (https://www.nature.com/articles/nprot.2014.071)

Shall we use 4 / (1/<# of cases> + 1/<# of controls>) or 2 / (1/<# of cases> + 1/<# of controls>) to calculate Neff?

Thanks

[Chris Chang]

Work out what happens when the number of cases is equal to the number of controls.

--
You received this message because you are subscribed to the Google Groups "plink2-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to plink2-users...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/plink2-users/7ef9ddf9-ea33-4a72-b524-f1f493e38098n%40googlegroups.com.