Boa tarde! Estou tentando fazer um trabalho e está realmente dificíl. Será que alguém poderia me ajudar?<br><br>Dois vetores, cada um com três elementos:<br><br><b>Curitiba <-c(24673, 505, 27)<br>Lapa <- c(682, 2, 0)</b><br clear="all">
<br>Quando obtenho o teste obtenho os seguintes resultados:<br><br><b>chisq.test(Curitiba, Lapa)</b><br>     <span style="color: rgb(51, 102, 255);">"> chisq.test(Curitiba, Lapa)</span><br style="color: rgb(51, 102, 255);">
<br style="color: rgb(51, 102, 255);"><span style="color: rgb(51, 102, 255);">        Pearson's Chi-squared test</span><br style="color: rgb(51, 102, 255);"><br style="color: rgb(51, 102, 255);"><span style="color: rgb(51, 102, 255);">data:  Curitiba and Lapa </span><br style="color: rgb(51, 102, 255);">
<span style="color: rgb(51, 102, 255);">X-squared = 6, df = 4, p-value = 0.1991</span><br style="color: rgb(51, 102, 255);"><br style="color: rgb(51, 102, 255);"><span style="color: rgb(51, 102, 255);">Warning message:</span><br style="color: rgb(51, 102, 255);">
<span style="color: rgb(51, 102, 255);">In chisq.test(Curitiba, Lapa) : Chi-squared approximation may be incorrect</span>"<br><br>Vi nos argumentos da chisq.test que para se fazer o teste, deve se usar Matriz ou vetores. Acontece que quando coloco os dois dados em formato de matriz, o resultado dá bem diferente:<br>
<br><b>m <- matrix(c(Curitiba, Lapa),2,3, byrow=TRUE)<br>chisq.test(m)</b><br>     "> <span style="color: rgb(51, 102, 255);">chisq.test(m)</span><br style="color: rgb(51, 102, 255);"><br style="color: rgb(51, 102, 255);">
<span style="color: rgb(51, 102, 255);">        Pearson's Chi-squared test</span><br style="color: rgb(51, 102, 255);"><br style="color: rgb(51, 102, 255);"><span style="color: rgb(51, 102, 255);">data:  m </span><br style="color: rgb(51, 102, 255);">
<span style="color: rgb(51, 102, 255);">X-squared = 10.9144, df = 2, p-value = 0.004266</span><br style="color: rgb(51, 102, 255);"><br style="color: rgb(51, 102, 255);"><span style="color: rgb(51, 102, 255);">Warning message:</span><br style="color: rgb(51, 102, 255);">
<span style="color: rgb(51, 102, 255);">In chisq.test(m) : Chi-squared approximation may be incorrect</span>"<br><br>A idéia é observar se existe algum tipo de associação entre Tipo de Gravidez e Munícipio. Os dados representam Um, dois, três ou mais filhos respectivamente.<br>
Algum dos meus testes está certo? Caso não estejam, alguém pode me informar onde eu estou errando? Essa é a primeira vez que trabalho com testes envolvendo a distribuição X²<br><br>-- <br>att,<br>Rafael Simões<br>(41) 9836 1010<br>
<a href="http://alienado.com.br" target="_blank">http://alienado.com.br</a><br><a href="http://twitter.com/cwbrafael" target="_blank">http://twitter.com/cwbrafael</a><br>