[R-sig-ME] P-value associated to explanatory from glmer binomial family
g@i@rrido @ending from gm@il@com
Fri May 25 10:13:34 CEST 2018
thanks so much for the clarification. After I run the LRT I get those
gm1 <- glmer(Mycoplasma~ type+ (1|donor.number), family=binomial)
Analysis of Variance Table
Df Sum Sq Mean Sq F value
type 1 3.2124 3.2124 3.2124
gm0<- glmer(Mycoplasma~ 1+ (1|donor.number), family=binomial)
gm0: Mycoplasma ~ 1 + (1 | donor.number)
gm1: Mycoplasma ~ type + (1 | donor.number)
Df AIC BIC logLik deviance Chisq Chi Df Pr(>Chisq)
gm0 2 46.492 49.869 -21.246 42.492
gm1 3 44.901 49.968 -19.451 38.901 3.5905 1 0.05811 .
The F-value associated to type, my only explanatory variable, is 3.124, as
anova(gm1) shows above
1. So which values should I take to calculate the P-value associated to the
variable type? 2 and 3 as shows anova(gm1, gm0)
Is like that then?
2. anova(gm1, gm0) give a P-value associated of 0.058, since I have only
one explanatory variable, is not this value the defining the significance
of this variable (the ine that makes the difference between the 2 models)
3. What is in this case the F-value and df provided by anova(gm1)?
Srry, I am a little confused with the results. Thanks!
2018-05-24 11:55 GMT+03:00 Thierry Onkelinx <thierry.onkelinx using inbo.be>:
> Dear Mario,
> Calculating the degrees of freedom of a mixed model is not straightforward.
> A workaround would be to use a likelihoodratio test between two nested
> models: one with and one without the variable. See the example below.
> gm1 <- glmer(cbind(incidence, size - incidence) ~ period + (1 | herd),
> data = cbpp, family = binomial)
> gm0 <- glmer(cbind(incidence, size - incidence) ~ (1 | herd),
> data = cbpp, family = binomial)
> anova(gm1, gm0)
> Best regards,
> ir. Thierry Onkelinx
> Statisticus / Statistician
> Vlaamse Overheid / Government of Flanders
> INSTITUUT VOOR NATUUR- EN BOSONDERZOEK / RESEARCH INSTITUTE FOR NATURE AND
> Team Biometrie & Kwaliteitszorg / Team Biometrics & Quality Assurance
> thierry.onkelinx using inbo.be
> Havenlaan 88
> <https://maps.google.com/?q=Havenlaan+88&entry=gmail&source=g> bus 73,
> 1000 Brussel
> To call in the statistician after the experiment is done may be no more
> than asking him to perform a post-mortem examination: he may be able to say
> what the experiment died of. ~ Sir Ronald Aylmer Fisher
> The plural of anecdote is not data. ~ Roger Brinner
> The combination of some data and an aching desire for an answer does not
> ensure that a reasonable answer can be extracted from a given body of data.
> ~ John Tukey
> 2018-05-23 13:14 GMT+02:00 Mario Garrido <gaiarrido using gmail.com>:
>> Dear lme4-users,
>> I am trying to get the P-value associated with a glmer model from the
>> binomial family.
>> My model is the following:
>> glmer(Infection.status~origin+ (1|donationID), family=binomial)->q7H
>> where Infection status is a dummy variable with two levels, infected and
>> I tried to get the P-value associated to the the explanatory variable
>> but I get only the F-value and the degrees of freedom
>> (aov <- anova(q7H))
>> Analysis of Variance Table
>> Df Sum Sq Mean Sq F value
>> origin 2 5.3061 2.6531 2.6531
>> I have 2 different questions
>> 1. Am I doing correctly or am I using an incorrect command?
>> 2. with the F-value I get and the df, should I go to test the significance
>> to a F or Chi-squared table? I guess I should go to the latest since I am
>> running a binomial test, right?
>> In case I have to go to an F table, how can I know the numerator and
>> denominator degrees of freedom?
>> Thanks in advance
>> [[alternative HTML version deleted]]
>> R-sig-mixed-models using r-project.org mailing list
[[alternative HTML version deleted]]
More information about the R-sig-mixed-models