[R] summary(svyglm) Pr (> | t |) ?

[John Fox]

Dear Diana,

> -----Original Message-----
> From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-project.org]
> On Behalf Of Diana Marcela Martinez Ruiz
> Sent: August-02-12 5:00 PM
> To: jfox at mcmaster.ca; tlumley at uw.edu; Ayuda en R
> Subject: Re: [R] summary(svyglm) Pr (> | t |) ?
> 
> 
> Thanks for the info, but with the anova () is obtained  p_valor of Wald
test for

 The standard anova() provides sequential ("type I") tests; by default,
Anova() in car provides partial ("type II") tests.

> the variable level of education, I want to see the nievel p_valor in view
of this
> variable as a dummy, so I wonder if in summary (mod. logis) the Pr (> | t
|)
> lets me know if each level is significant in the model

This is no different for svyglm() than for any model in R with a linear
predictor; the individual t-tests are for (partial) differences between each
other level and the reference level for the set of contrasts (dummy
regressors); if this is what you're interested in (and of course the
reference level isn't arbitrarily chosen), then you can use the t-tests for
the coefficients.

Best,
 John
> 
> 
> Thanks
> 
> > From: jfox at mcmaster.ca
> > To: dianammr06 at hotmail.com
> > CC: r-help at r-project.org; tlumley at uw.edu
> > Subject: RE: [R] summary(svyglm)  Pr (> | t |) ?
> > Date: Thu, 2 Aug 2012 16:50:37 -0400
> >
> > Dear Diana,
> >
> > The Anova() function in the car package will produce a Wald test for
> > each term in a model fit by svyglm().
> >
> > I hope this helps,
> >  John
> >
> > > -----Original Message-----
> > > From: r-help-bounces at r-project.org
> > > [mailto:r-help-bounces at r-project.org]
> > > On Behalf Of Diana Marcela Martinez Ruiz
> > > Sent: August-02-12 4:14 PM
> > > To: Ayuda en R; tlumley at uw.edu
> > > Subject: [R] summary(svyglm) Pr (> | t |) ?
> > >
> > >
> > > Hello
> > >
> > > I want to know if the summary of the logistic model with survey  Pr
> > > (> | t
> > |) to
> > > test if the coefficient of the model is significant,  ie is the
> > > p_valor
> > wald test
> > > for the model coefficients,  for I am interested to know if the
> > > three
> > levels of
> > > the variable educational level are significant to the model
> > > (significance
> > of
> > > handling 0.2), I present below the results of my model
> > >
> > >  summary(mod.logis)
> > >
> > > Call:
> > > svyglm(formula = APES_DICOT ~ Nivel_Educativo + Ocupacion_principal +
> > >     Afiliacion_salud + Tiene_cuidador + Presencia_enfer_cronica +
> > >     Consumo_tabaco + Consumo_alcohol + Presencia_Dolor +
> > > Frec_dolor_cronico +
> > >     Punt_min_mental + ABC_fis + ABC_instr + Anergia +
> > > Actividad_fisica_ultimo_aqo +
> > >     Perdida_peso_no_intensionada + Lenta_velocidad_marcha +
> > > Disminucion_fuerza_agarre +
> > >     Hospitalizacion_ultimo_aqo + Urgencias_ultimo_aqo + Miedo_caer,
> > >     design = Muestra.comp, family = quasibinomial())
> > >
> > > Survey design:
> > > svydesign(id = ~b, strata = ~d, nest = TRUE, weights = ~c, data =
am18,
> > >     fpc = ~e)
> > >
> > > Coefficients:
> > >                                         Estimate Std. Error t value
> > Pr(>|t|)
> > > (Intercept)                             -3.10084    1.85632  -1.670
> > 0.1104
> > > Nivel_EducativoNinguno/No recuerda       1.52556    0.88735   1.719
> > 0.1010
> > > Nivel_EducativoPrimaria                  1.16256    0.64404   1.805
> > 0.0861 .
> > > Nivel_EducativoSecundaria                0.55771    1.04909   0.532
> > 0.6009
> > >
> > > Thanks
> > > 	[[alternative HTML version deleted]]
> >
> >
> 
> 	[[alternative HTML version deleted]]
> 
> ______________________________________________
> R-help at 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.