p values for betas about lme4qtl HOT 3 CLOSED

Hello,

Thank you for the quick reply.
That is almost it.
I am looking for the p value of the beta associated with AGE in m1.

For example, the p value associated with x below.

> x <- rnorm(100,0,1) + 1:100
> y <- rnorm(100,0,1) + 1:100
> summary(lm(y~x))

Call:
lm(formula = y ~ x)

Residuals:
    Min      1Q  Median      3Q     Max 
-3.4624 -1.0323 -0.0721  0.9766  3.0269 

Coefficients:
            Estimate Std. Error t value Pr(>|t|)    
(Intercept) 0.431093   0.286746   1.503    0.136    
x           0.994736   0.004932 201.703   **<2e-16** *****

from lme4qtl.

Hi,

Thank you for using the package. Having a reproducible example would help to answer your question.

Here I put an example code to test the effect of AGE.

library(lme4qtl)
data(dat40, package = "lme4qtl")
dat40 <- na.omit(dat40) # to deal with error in anova.merMod(m0, m1) : models were not all fitted to the same size of dataset
m0 <- relmatLmer(trait1 ~ (1|ID), dat40, relmat = list(ID = kin2))
m1 <- relmatLmer(trait1 ~ AGE + (1|ID), dat40, relmat = list(ID = kin2))
anova(m0, m1)

Output:

> anova(m0, m1)
refitting model(s) with ML (instead of REML)
Data: dat40
Models:
m0: trait1 ~ (1 | ID)
m1: trait1 ~ AGE + (1 | ID)
   npar    AIC    BIC  logLik deviance Chisq Df Pr(>Chisq)
m0    3 942.93 953.10 -468.47   936.93
m1    4 944.72 958.27 -468.36   936.72 0.219  1     0.6398

Is that what you are looking for?

Best,
Andrey

from lme4qtl.

Hi there, I have a similar situation and I had a follow-up question. For the anova(), I noted it refit the models from REML to ML. Does this have an impact on the calculation of the fixed effect AGE in terms of changing the test statistic? Would the p-value in this case be more/less conservative than an REML value assuming such a value could be calculated?

from lme4qtl.