Estimation is realized by MLE, estimators are asymptotically normal Try this reg<-vglm(...) p.value<-1-pnorm(abs(coef(reg)/sqrt(diag(vcov(reg)))))
Justin BEM BP 1917 Yaoundé Tél (237) 76043774 ________________________________ De : suuz <suuz_b...@hotmail.com> À : r-help@r-project.org Envoyé le : Mardi 23 Août 2011 15h04 Objet : [R] P values for vglm(zibinomial) function in VGAM Hi , I know this question has been asked twice in the past but to my knowldege, it still hasn't been solved. I am doing a zero inflated binomial model using the VGAM package, I need to obtain p values for my Tvalues in the vglm output. code is as follows > mod2=vglm(dmat~Season+Diel+Tidal.phase+Tidal.cycle,zibinomial, data=mp1) > summary(mod2) Call: vglm(formula = dmat ~ Season + Diel + Tidal.phase + Tidal.cycle, family = zibinomial, data = mp1) Pearson Residuals: Min 1Q Median 3Q Max logit(phi) -3.6496 0.273679 0.285619 0.296763 1.0974 logit(mu) -6.3631 -0.029027 -0.020786 -0.011719 80.6051 Coefficients: Value Std. Error t value (Intercept):1 2.365835 0.029142 81.18334 (Intercept):2 -3.182376 0.050054 -63.57944 SeasonSpring -0.080840 0.054201 -1.49147 SeasonSummer 0.204781 0.049936 4.10083 SeasonWinter 0.385078 0.043874 8.77692 DielE -0.079190 0.079859 -0.99163 DielM 0.071607 0.074620 0.95963 DielN 0.132377 0.036419 3.63488 Tidal.phaseNT -0.252715 0.054053 -4.67536 Tidal.phaseST 0.145777 0.045554 3.20010 Tidal.cycleF 0.114808 0.044897 2.55713 Tidal.cycleH -0.074224 0.048063 -1.54428 Tidal.cycleL -0.049681 0.047717 -1.04116 Number of linear predictors: 2 Names of linear predictors: logit(phi), logit(mu) Dispersion Parameter for zibinomial family: 1 Log-likelihood: -7831.693 on 30569 degrees of freedom Number of Iterations: 13 I have tried the suggestions pt(coef(summary(mod2)), 30569, lower.tail = TRUE, log.p = FALSE) Value Std. Error t value (Intercept):1 0.9910021735 0.5116242 1.000000e+00 (Intercept):2 0.0007310901 0.5199600 0.000000e+00 SeasonSpring 0.4677849543 0.5216125 6.792427e-02 SeasonSummer 0.5811276264 0.5199133 9.999794e-01 SeasonWinter 0.6499089525 0.5174974 1.000000e+00 DielE 0.4684409796 0.5318250 1.606939e-01 DielM 0.5285424555 0.5297410 8.313752e-01 DielN 0.5526565030 0.5145256 9.998607e-01 Tidal.phaseNT 0.4002451101 0.5215532 1.473475e-06 Tidal.phaseST 0.5579507181 0.5181669 9.993124e-01 Tidal.cycleF 0.5457011061 0.5179053 9.947207e-01 Tidal.cycleH 0.4704164848 0.5191670 6.126507e-02 Tidal.cycleL 0.4801883607 0.5190291 1.489050e-01 pt(coef(mod2), 30569) (Intercept):1 (Intercept):2 SeasonSpring SeasonSummer SeasonWinter 0.9910021735 0.0007310901 0.4677849543 0.5811276264 0.6499089525 DielE DielM DielN Tidal.phaseNT Tidal.phaseST 0.4684409796 0.5285424555 0.5526565030 0.4002451101 0.5579507181 Tidal.cycleF Tidal.cycleH Tidal.cycleL 0.5457011061 0.4704164848 0.4801883607 But these seem to be giving strange results, those which are clearly more different to the basline levels (Autumn, D, bl and E) are coming up as least significant. Perhaps it is my interpretation. I couldnt follow 2*min(pt(coef(summary(mod2)), 30569), 1-pt(coef(summary(mod2)), 30569)) ?? Does anyone know what I might be doing wrong or how to go about the last code? I have read as much as I can find on VGAM and zib but have ran out of ideas. I apologise if this appears to be a beginners question. many thanks in advance -- View this message in context: http://r.789695.n4.nabble.com/P-values-for-vglm-zibinomial-function-in-VGAM-tp3762858p3762858.html Sent from the R help mailing list archive at Nabble.com. ______________________________________________ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code. [[alternative HTML version deleted]]
______________________________________________ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.