[R-br] DESDOBRAMENTO DE UM FATORIAL TRIPLO

[Tiago Souza Marçal]

Boa tarde pessoal,
Estou tentando desdobrar a interação tripla do exemplo abaixo, entretanto, as somas de quadrado e os graus de liberdade estão dando 0.
Ex:
anos<-c(rep(1,144),rep(2,144))
ambientes<-c(rep(1,48),rep(2,48),rep(3,48),rep(1,48),rep(2,48),rep(3,48))
genotipos<-c(rep(c(rep(1,6),rep(2,6),rep(3,6),rep(4,6),rep(5,6),rep(6,6),rep(7,6),rep(8,6)),6))
blocos<-c(rep(c(1,2,3,4,5,6),48))
resp<-c(195,196,192,208,180,213,216,194,198,209,238,236,307,341,305,316,335,340,435,403,499,422,487,483,218,192,222,221,218,219,227,422,429,448,463,461,134,178,188,285,294,295,707,803,828,714,848,808,141,153,168,270,146,250,244,134,142,253,261,234,342,359,349,359,369,330,448,465,431,449,459,440,243,163,250,244,210,231,246,440,444,454,462,464,141,154,172,269,249,259,759,840,838,741,845,806,569,460,433,373,480,384,490,191,406,500,424,314,842,703,881,804,912,929,739,749,629,630,848,721,931,849,988,880,893,993,290,204,380,311,215,295,190,109,239,137,200,146,200,218,132,249,235,202,294,324,386,382,319,395,625,620,719,619,627,620,626,540,668,662,639,575,882,984,994,796,802,709,525,515,540,409,589,467,180,256,260,198,292,109,365,372,364,274,396,302,929,783,894,866,772,872,130,394,399,203,387,213,211,396,192,208,234,238,303,442,302,316,332,449,436,608,492,429,487,489,218,393,225,220,214,217,226,424,426,541,462,467,139,279,389,288,296,298,701,801,823,913,849,805,195,390,199,239,217,180,306,380,488,372,309,375,502,534,514,726,682,459,273,324,278,296,302,293,304,315,316,337,291,312,273,355,286,283,331,403,305,296,283,318,309,481,283,255,306,313,274,484)
dados<-data.frame(anos,ambientes,genotipos,blocos,resp)

#Estou usando os seguintes comandos:
anos<-factor(anos)
ambientes<-factor(ambientes)
genotipos<-factor(genotipos)
blocos<-factor(blocos)
av<-aov(resp~anos*ambientes/genotipos+(anos:ambientes:blocos)) 
des.tab<-sapply(paste("anos:ambientes", levels(anos:ambientes), sep=""), simplify=F, grep, x=names(coef(av)[av$assign==4]))
des.tab1<-summary(av, split=list("anos:ambientes:genotipos"=des.tab))

Agradeço desde já por qualquer ajuda.
Att.
Tiago. 

 		 	   		  
-------------- Próxima Parte ----------
Um anexo em HTML foi limpo...
URL: <http://listas.inf.ufpr.br/pipermail/r-br/attachments/20130108/4c44148d/attachment.html>