Bookmark and Share

Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: cendif

From   "Roger B. Newson" <>
Subject   Re: st: cendif
Date   Thu, 22 Nov 2012 15:53:41 +0000

If you want to compare 2 intervention groups separately with the control using -somersd- and -cendif-, then you need to use an -if- qualifier for each intervention-control pair. If your grouping variable is -grupo-, with values 0 for control, 1 for Intervention 1, and 2 for Intervention 2, then you can compare interventions 1 and 2 with the control by typing:

xi i.grupo, noomit
somersd grupo T6convulsion if inlist(grupo,0,1), transf(z) tdist cluster(medcod3); cendif T6convulsion if inlist(grupo,0,1), by(_Igrupo_0) transf(z) tdist cluster(medcod3); somersd grupo T6convulsion if inlist(grupo,0,2), transf(z) tdist cluster(medcod3); cendif T6convulsion if inlist(grupo,0,2), by(_Igrupo_0) transf(z) tdist cluster(medcod3);

This will give the 2 median intervention-control differences in outcome, with confidence limits, and also confidence linits, and P-values, for the Somers' D parameters of outcome with respect to each intervention compared to the control group.

I hope this helps. Let me know if you have any more queries.

Best wishes


Roger B Newson BSc MSc DPhil
Lecturer in Medical Statistics
Respiratory Epidemiology and Public Health Group
National Heart and Lung Institute
Imperial College London
Royal Brompton Campus
Room 33, Emmanuel Kaye Building
1B Manresa Road
London SW3 6LR
Tel: +44 (0)20 7352 8121 ext 3381
Fax: +44 (0)20 7351 8322
Web page:
Departmental Web page:

Opinions expressed are those of the author, not of the institution.

On 22/11/2012 13:55, alfonsa leiva wrote:
Dear stata user

In the between group analisis (control vs Intervention1/ control vs Intervention2/Intervention1 vs Intervention2)of a three arm cluster trial,

  I can get p- values by somersd:

xi, noomit: somersd   T6convulsion i.grupo, taua transf(z) tdist cluster(medcod3)

testparm _I*

. testparm _I*

  ( 1)  _Igrupo_1 = 0
  ( 2)  _Igrupo_2 = 0
        Constraint 2 dropped

        F(  1,    46) =    7.02
             Prob > F =    0.0110

Could i use also Hodges-lehman estimates by cendiff to report differences between groups?

xi,noomit: cendif   T6irritabilidad , transf(z) tdist cluster(medcod3) by(grupo)

  Percent    Pctl_Dif     Minimum     Maximum
         50           0           0           0

I apreciatte any comment

*   For searches and help try:

*   For searches and help try:

© Copyright 1996–2018 StataCorp LLC   |   Terms of use   |   Privacy   |   Contact us   |   Site index