```Date: Tue, 27 Sep 2005 16:51:04 +0200 Reply-To: fvillamayor-research@ferrergrupo.com Sender: "SPSSX(r) Discussion" From: Frederic Villamayor Subject: Help on MIXED Content-Type: text/plain; charset="ISO-8859-1" Dear co-listers: I need some help on the proper syntax to run a MIXED procedure on my data. The experimental design is as follows: We've got three volunteers. From each volunteer a sample of skin is taken The samples are divided in different aliquots (replicates) and subjected to two possible treatments. The concentration of a certain substance (PNC) is repeatedly meassured along 12 different times on three replicates per treatment on volunteer 1, three replicates for treatment 1 and 2 replicates for treatment 2 on volunteer 2, and only one replicate per treatment on volunteer 3. It is clear enough? Maybe this table helps: VOL TREAT REPL TIME 1 1 1 1-2-3-4-5-6-7-8-9-10-11-12 2 1-2-3-4-5-6-7-8-9-10-11-12 3 1-2-3-4-5-6-7-8-9-10-11-12 2 1 1-2-3-4-5-6-7-8-9-10-11-12 2 1-2-3-4-5-6-7-8-9-10-11-12 3 1-2-3-4-5-6-7-8-9-10-11-12 2 1 1 1-2-3-4-5-6-7-8-9-10-11-12 2 1-2-3-4-5-6-7-8-9-10-11-12 3 1-2-3-4-5-6-7-8-9-10-11-12 2 1 1-2-3-4-5-6-7-8-9-10-11-12 2 1-2-3-4-5-6-7-8-9-10-11-12 3 1-2-3-4-5-6-7-8-9-10-11-12 3 1 1 1-2-3-4-5-6-7-8-9-10-11-12 2 1-2-3-4-5-6-7-8-9-10-11-12 3 1-2-3-4-5-6-7-8-9-10-11-12 2 1 1-2-3-4-5-6-7-8-9-10-11-12 2 1-2-3-4-5-6-7-8-9-10-11-12 3 1-2-3-4-5-6-7-8-9-10-11-12 What I'm trying to test if that the treatment influences on the concentration of PNC. After repeated assays and errors, I've got this syntax, reflecting (I expect so...) that the experimenta subjects are the 3 volunteers, than the replicates are a random factor nested for each combination of treatment and volunteer, that repeated meassures are taken for each combination ot time, replicate and treatment, and thatthe fixed effects on the model are treatment, time and its interaction. Furthermore, I comparetreatments on each timepointadjusting for multipla comparisons. MIXED pnc BY vol treat repl time /FIXED = treat time time*treat /RANDOM repl(treat*vol)) | SUBJECT(vol) COVTYPE(AR1) /REPEATED = time*repl*treat | SUBJECT(vol) COVTYPE(AR1) /EMMEANS = TABLES(time*treat) COMPARE(treat) ADJUST(BONFER) . The results are convenient to me, but It would be splendid if I can read some comments from any of you, since I'm rather novice with this command and I'm not sure if I'm doing what I really want to do. Thanks in advance. Frederic %%%%%%%%%%%%%%%%%%%%%%% Frederic Villamayor Unitat de Bioestadística Àrea de Desenvolupament Preclínic CIDF Ferrer Grupo Juan de Sada, 32 08028 ? Barcelona Espanya E-mail: fvillamayor-research@ferrergrupo.com Tel: +34 935093236 Fax: +34 934112764 WWW: www.ferrergrupo.com %%%%%%%%%%%%%%%%%%%%%%% "Sanity is not statistical" 1984 (George Orwell) ```

Back to: Top of message | Previous page | Main SPSSX-L page