r¿€skin rash data¿6)ØH66ZF)õ‹¿Ëô‹¿Øô‹¿|=$€ºGbî3øô‹¿)ØH66€19 Dec 2006 02:19ûûûûûresponsetreatmentclinicdoctorpatient%9.0g%9.0g%9.0g%9.0g%%9.0g%9.0g%9.0g%9.0gakenakeodelœò·€œò·!!n.0while ! !( !( ! "1  2 treatment protocolsn.0while ! !( !( ! "1  4 clinics per treatmentn.0while ! !( !( ! "1  3 doctors per clinicntn.0while ! !( !( ! "1  4 patients per doctortn.0while ! !( !( ! "1  c_dta0050601 Jun Jul Aug Sep Ocnote7Gˆô‹¿cpÄ·ã©ë·c²Ž'anova response t / d|t / p|d|t /D_dta0050601 Jun Jul Aug Sep Ocnote0Gˆô‹¿DpÄ·ã©ë·D²Ž'7®_dta0050601 Jun Jul Aug Sep Ocnote6Gˆô‹¿®pÄ·ã©ë·®²Ž'Or you could redo your anova. (Be careful to notice how I have numbered the levels of the different terms).§_dta0050601 Jun Jul Aug Sep Ocnote5Gˆô‹¿§pÄ·ã©ë·§²Ž'test c|t d|c|t / p|d|c|t ---- gives test on pooled doctor and center term with patient as error term¶_dta0050601 Jun Jul Aug Sep Ocnote4Gˆô‹¿¶pÄ·ã©ë·¶²Ž'test t / c|t d|c|t ---- gives test on treatment with pooling of doctor and center nested in treatment as error termv_dta0050601 Jun Jul Aug Sep Ocnote3Gˆô‹¿vpÄ·ã©ë·v²Ž'You could use -test- to get the results you desire:š_dta0050601 Jun Jul Aug Sep Ocnote2Gˆô‹¿špÄ·ã©ë·š²Ž'After examining the above anova results some people would want to drop the clinic term.m_dta0050601 Jun Jul Aug Sep Ocnote1Gˆô‹¿mpÄ·ã©ë·m²Ž'anova response t / c|t / d|c|t / p|d|c|t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`2```