Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | Terry Haines <terrence.haines@monash.edu> |
To | statalist@hsphsun2.harvard.edu |
Subject | st: Which ICC to report from iccvar |
Date | Fri, 24 Jan 2014 14:58:04 +1100 |
I have a stepped wedge, cluster randomised trial of a group education program. Study participants (id) each had multiple assessments (ax) over the study year and are nested within school (school). The iccvar I ran after using xtmixed returned multiple ICC values. I would like to know which one I should report (or both) when reporting the result of this trial as normally only one ICC value is reported? iccvar output is below. . iccvar Intraclass Correlation Estimates Harmonic Mean of Level 2 Units per Level 3 Unit = 2.367 Harmonic Mean of Level 3 Units per Level 4 Unit = 10.375 ICC Std. Err. [95% Conf. Interval] school 0.00000 . . 1.00000 id 0.57872 2.57990 0.00000 1.00000 ax 0.28393 2.93062 0.00000 1.00000 Many thanks, Terry Haines * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/faqs/resources/statalist-faq/ * http://www.ats.ucla.edu/stat/stata/