Hello,

Could you please advise as to whether or not the following analysis is suitable?

I would like to compare groups of monozygotic twins (MZ concordant vs. MZ control ; MZ discordant affected twin vs. unaffected twin) using qdec. This is not as straightforward as group comparisons usually are because I cannot make the assumption of all data being independent as twin pairs are, obviously, related to one another.

To get around this issue I have started a qdec analysis and added a category 'pair' in which twin pairs share a category as follows (qdec.table.dat), in order to indicate that they have shared genetic and environmental influence:

subject no
subjid  pair    diagnosis       age     gender
1       a       MZ_Discord_Well ...     ...
2       a       MZ_Discord_Ill         
3       b       MZ_Discord_Ill         
4       b       MZ_Discord_Well        
5       c       MZ_Discord_Well        
6       c       MZ_Discord_Ill         
7       h       MZ_Discord_Well        
8       h       MZ_Discord_Ill         
9       l       MZ_Discord_Well        
10      l       MZ_Discord_Ill         
11      m       MZ_Discord_Well        
12      m       MZ_Discord_Ill         
13      n       MZ_Discord_Well        
14      n       MZ_Discord_Ill         
15      p       MZ_Discord_Well        
16      p       MZ_Discord_Ill         
17      s       MZ_Discord_Well        
18      s       MZ_Discord_Ill         
19      t       MZ_Discord_Well        
20      t       MZ_Discord_Ill         

I would appreciate if you could suggest a more robust way around this.

Many thanks,

Rysia