Log in
Register|
ratnakar1811 ★ India, 2008-07-31 13:18 (6529 d 04:32 ago) Posting: # 2114 Views: 6,230 |
|
|
Dear forum members, I just have query regarding balancing the sequence at the time of statistical analysis. In case of studies which are mentioned as balanced sequence, but during the study due to various reason like dropout or discontinuation data is unbalanced. In such situation should we balance the sequence at the time of statistical analysis or if not then we will have to remove the subject from statistical analysis just to balance the sequence. But this data being balance or unbalanced does have effect of the sequence effect from ANOVA. Expert opinion on the same will be highly appreciated. Ratnakar Jadhav |
|
ElMaestro ★★★ Denmark, 2008-07-31 13:34 (6529 d 04:16 ago) @ ratnakar1811 Posting: # 2115 Views: 5,012 |
|
|
Dear Ratnakar Jadhav, First of all, imbalance is not an absolute problem. You can do a valid analysis on imbalanced data (proc mixed in SAS, etc). Second, you should not attempt to balance imbalanced data by removing something else. Third, Sequence effects are seen now and then. Often they are randomisation failures, and this forum has quite some searchable material on it. EM. |
|
MGR ★ India, 2008-08-05 09:16 (6524 d 08:34 ago) @ ratnakar1811 Posting: # 2145 Views: 5,050 |
|
|
Dear Ratnakar, As said by the ElMaestro, we can't delete the subject data to get an balanced study thats why we have to mention this in the protocol itself before the approval in the appropriate section stating that "Unbalanced Sequences will be considered for calculations when there are more dropouts from the same sequence" and mention the appropriate statistical methods for analysis like SAS/Mixed Procedure etc. Regards, Raghavendra  Strides Arcolab |
Ing. Helmut Schütz