<p class="MsoNoSpacing">Prezados(as)</p><p class="MsoNoSpacing">Precisamos de ajuda no seguinte problema:</p>
<p class="MsoNoSpacing"> Estamos ajustando
um modelo de transição em dados binários longitudinais, usando a resposta imediatamente
anterior como covariavel. Os dados estão no arquivo "d_dogs.txt" em anexo.
O interesse é construir IC simultâneos para as probabilidades condicionais.</p>
<p class="MsoNoSpacing"> </p>
<p class="MsoNoSpacing"><span lang="EN-US"><font face="comic sans ms, sans-serif">d1=read.table("d_dogs.txt",
header=T, quote="\"")</font></span></p>
<p class="MsoNoSpacing"><span lang="EN-US"><font face="comic sans ms, sans-serif">attach(d1)</font></span></p>
<p class="MsoNoSpacing"><span lang="EN-US"><font face="comic sans ms, sans-serif">gmod1
<- glm(y ~ Trat+y_ant, data = d1, </font></span><span style="font-family:'comic sans ms',sans-serif">family = binomial())</span></p>
<p class="MsoNoSpacing"><span lang="EN-US"><font face="comic sans ms, sans-serif">summary(gmod1)</font></span></p>
<p class="MsoNoSpacing"><font face="comic sans ms, sans-serif">###########################################################</font></p>
<p class="MsoNoSpacing"><font face="comic sans ms, sans-serif">library(multcomp)</font></p>
<p class="MsoNoSpacing"><span lang="EN-US"><font face="comic sans ms, sans-serif">gmod1_ci
<- confint(glht(gmod1, linfct = mcp(Trat =
"Tukey",covariate_average = TRUE)))</font></span></p>
<p class="MsoNoSpacing"><span lang="EN-US"><font face="comic sans ms, sans-serif">gmod1_ci$confint
<- apply(gmod1_ci$confint, 2, binomial()$linkinv)</font></span></p>
<p class="MsoNoSpacing"><span lang="EN-US"><font face="comic sans ms, sans-serif">gmod1_ci$confint</font></span></p>
<p class="MsoNoSpacing"><span lang="EN-US"><font face="comic sans ms, sans-serif">apply(gmod1_ci$confint,
2, binomial()$linkinv)</font></span></p>
<p class="MsoNoSpacing"><font face="comic sans ms, sans-serif">plot(gmod1_ci, xlab = "Media da Prob marginal",</font></p>
<p class="MsoNoSpacing"><font face="comic sans ms, sans-serif"> xlim = c(0,
1))</font></p>
<p class="MsoNoSpacing"><font face="comic sans ms, sans-serif"> </font></p>
<p class="MsoNoSpacing"><font face="comic sans ms, sans-serif">A dúvida é a seguinte:</font></p>
<p class="MsoNoSpacing"><font face="comic sans ms, sans-serif"> </font></p>
<p class="MsoNormal" style="margin-bottom:0.0001pt"><span lang="EN-US" style="font-size:10.0pt;font-family:"Lucida Console";color:blue">> apply(gmod1_ci$confint, 2,
binomial()$linkinv)</span></p>
<p class="MsoNormal" style="margin-bottom:0.0001pt"><span lang="EN-US" style="font-size:10pt;font-family:'Lucida Console';background-color:rgb(225,226,229);background-repeat:initial initial"> </span><span style="font-size:10pt;font-family:'Lucida Console';background-color:rgb(225,226,229);background-repeat:initial initial">Estimate lwr upr</span></p>
<p class="MsoNormal" style="margin-bottom:0.0001pt"><span style="font-size:10pt;font-family:'Lucida Console';background-color:rgb(225,226,229);background-repeat:initial initial">B - A 0.5615396 0.5062051 0.6919771</span></p>
<p class="MsoNormal" style="margin-bottom:0.0001pt"><span style="font-size:10pt;font-family:'Lucida Console';background-color:rgb(225,226,229);background-repeat:initial initial">C - A 0.7310586 0.5000000 0.7310586</span></p>
<p class="MsoNormal" style="margin-bottom:0.0001pt"><span style="font-size:10pt;font-family:'Lucida Console';background-color:rgb(225,226,229);background-repeat:initial initial">D - A 0.5439900 0.5035155 0.6819754</span></p>
<p class="MsoNormal" style="margin-bottom:0.0001pt"><span style="font-size:10pt;font-family:'Lucida Console';background-color:rgb(225,226,229);background-repeat:initial initial">C - B 0.7310586 0.5000000 0.7310586</span></p>
<p class="MsoNormal" style="margin-bottom:0.0001pt"><span style="font-size:10pt;font-family:'Lucida Console';background-color:rgb(225,226,229);background-repeat:initial initial">D - B 0.5973698 0.5089817 0.7148991</span></p>
<p class="MsoNormal" style="margin-bottom:0.0001pt"><span style="font-size:10pt;font-family:'Lucida Console';background-color:rgb(225,226,229);background-repeat:initial initial">D - C 0.5000000 0.5000000 0.7310586</span></p>
<p class="MsoNoSpacing"> </p>
<p class="MsoNoSpacing">Estes intervalos estão incoerentes com os resultados do
gráfico:</p>
<p class="MsoNoSpacing"> </p>
<p class="MsoNoSpacing">plot(gmod1_ci, xlab = "Media da Prob marginal", xlim = c(0,
1))</p>
<p class="MsoNoSpacing"> </p>
<p class="MsoNoSpacing">Qual é o problema???????</p>
<p class="MsoNoSpacing">Desde já agradeço</p>
<p class="MsoNoSpacing"> </p><div><br></div>-- <br>Gilenio Borges Fernandes<br>Universidade Federal da Bahia<br>Instituto de Matemática<br>Departamento de Estatística<br>Av. Adhemar de Barros, s/n – Ondina.<br>40.170-110 - Salvador - BA, Brasil<br>
Tel.: (071)3283-6280/6336 Fax: (071)3283-6276<br>URL: <a href="http://lattes.cnpq.br/6764860618464860" target="_blank">http://lattes.cnpq.br/6764860618464860</a><br><br>