Multilevel models Probl

m0g <- glmer(unempl~1+(1|country), data=alldat,family=binomial, control = glmerControl(optimizer = "bobyqa") )
## Warning in checkConv(attr(opt, "derivs"), opt$par, ctrl = control$checkConv, :
## Model failed to converge with max|grad| = 0.0359952 (tol = 0.002, component 1)
print(icc(m0g, by_group = TRUE))
## # ICC by Group
## 
## Group   |   ICC
## ---------------
## country | 0.046
####does not work with year in random effect
fm1 <- glmer(unempl~ country +factor(year) +sex+ +age +countrybirth +education +yearsar +compo + (1|country), data=alldat, family=binomial, control = glmerControl(optimizer = "bobyqa", calc.derivs = FALSE))
## Warning in commonArgs(par, fn, control, environment()): maxfun < 10 *
## length(par)^2 is not recommended.
htmlreg(list("Multilevel 0"=m0g, "Fixed-Effect 1"=fm1))
Statistical models
  Multilevel 0 Fixed-Effect 1
(Intercept) -2.46*** -2.36***
  (0.00) (0.19)
countryBG   0.81***
    (0.01)
countryCY   1.02***
    (0.01)
countryCZ   0.78***
    (0.01)
countryDE   0.06***
    (0.01)
countryDK   0.24***
    (0.02)
countryEE   0.95***
    (0.01)
countryEL   1.79***
    (0.01)
countryES   1.77***
    (0.01)
countryFI   0.80***
    (0.01)
countryFR   1.11***
    (0.01)
countryHU   0.93***
    (0.01)
countryIE   1.18***
    (0.01)
countryIT   1.00***
    (0.01)
countryLT   1.33***
    (0.01)
countryLU   0.29***
    (0.02)
countryLV   1.35***
    (0.01)
countryNL   0.24***
    (0.01)
countryPT   1.05***
    (0.01)
countryUK   0.49***
    (0.01)
factor(year)2009   0.35***
    (0.00)
factor(year)2010   0.49***
    (0.00)
factor(year)2011   0.57***
    (0.00)
factor(year)2012   0.76***
    (0.00)
factor(year)2013   0.83***
    (0.00)
factor(year)2014   0.78***
    (0.00)
factor(year)2015   0.70***
    (0.00)
factor(year)2016   0.62***
    (0.00)
factor(year)2017   0.51***
    (0.01)
factor(year)2018   0.40***
    (0.01)
factor(year)2019   0.32***
    (0.01)
sexfemale   0.15***
    (0.00)
age26-34   -0.57***
    (0.00)
age35-44   -0.98***
    (0.00)
age45-54   -1.26***
    (0.00)
age55-65   -1.37***
    (0.00)
countrybirthEU/EFTA   0.22
    (0.19)
countrybirthNew_Member   0.27
    (0.19)
countrybirthEurope_nEU/nEFTA   0.44*
    (0.19)
countrybirthNon_Europe   0.51**
    (0.19)
educationISCED3/4   -0.51***
    (0.00)
educationISCED5+   -1.02***
    (0.00)
yearsar5-9years   -0.09***
    (0.01)
yearsar>10years   -0.00
    (0.01)
yearsarnative   -0.04
    (0.19)
compoadult_ch   0.22***
    (0.01)
compocouple   -0.44***
    (0.00)
compocouple_ch   -0.46***
    (0.00)
compoconf_adults   -0.02***
    (0.00)
compoconf_adults_ch   -0.06***
    (0.00)
AIC 13689325.13 8154600.32
BIC 13689355.09 8155339.65
Log Likelihood -6844660.56 -4077249.16
Num. obs. 23758491 14600295
Num. groups: country 20 20
Var: country (Intercept) 0.16 0.00
***p < 0.001; **p < 0.01; *p < 0.05 
summary(m0g)

Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) [glmerMod] Family: binomial ( logit ) Formula: unempl ~ 1 + (1 | country) Data: alldat Control: glmerControl(optimizer = “bobyqa”)

 AIC      BIC   logLik deviance df.resid

13689325 13689355 -6844661 13689321 23758489

Scaled residuals: Min 1Q Median 3Q Max -0.4351 -0.3289 -0.3047 -0.2393 4.8478

Random effects: Groups Name Variance Std.Dev. country (Intercept) 0.1604 0.4005
Number of obs: 23758491, groups: country, 20

Fixed effects: Estimate Std. Error z value Pr(>|z|)
(Intercept) -2.456526 0.002017 -1218 <2e-16 *** — Signif. codes: 0 ‘’ 0.001 ’’ 0.01 ’’ 0.05 ‘.’ 0.1 ’ ’ 1 optimizer (bobyqa) convergence code: 0 (OK) Model failed to converge with max|grad| = 0.0359952 (tol = 0.002, component 1)

summary(fm1)

Generalized linear mixed model fit by maximum likelihood (Laplace Approximation) [glmerMod] Family: binomial ( logit ) Formula: unempl ~ country + factor(year) + sex + +age + countrybirth +
education + yearsar + compo + (1 | country) Data: alldat Control: glmerControl(optimizer = “bobyqa”, calc.derivs = FALSE)

 AIC      BIC   logLik deviance df.resid

8154600 8155340 -4077249 8154498 14600244

Scaled residuals: Min 1Q Median 3Q Max -1.7040 -0.3443 -0.2529 -0.1761 13.6431

Random effects: Groups Name Variance Std.Dev. country (Intercept) 0 0
Number of obs: 14600295, groups: country, 20

Fixed effects: Estimate Std. Error z value Pr(>|z|)
(Intercept) -2.359969 0.193860 -12.174 < 2e-16 countryBG 0.809957 0.007535 107.500 < 2e-16 countryCY 1.021607 0.008875 115.108 < 2e-16 countryCZ 0.775839 0.007925 97.895 < 2e-16 countryDE 0.061263 0.006369 9.619 < 2e-16 countryDK 0.237136 0.015283 15.516 < 2e-16 countryEE 0.949694 0.011028 86.120 < 2e-16 countryEL 1.787535 0.005652 316.238 < 2e-16 countryES 1.769388 0.006338 279.181 < 2e-16 countryFI 0.797219 0.009078 87.816 < 2e-16 countryFR 1.111627 0.006761 164.429 < 2e-16 countryHU 0.926046 0.006078 152.364 < 2e-16 countryIE 1.181569 0.006097 193.802 < 2e-16 countryIT 1.002984 0.005511 182.003 < 2e-16 countryLT 1.334706 0.007569 176.332 < 2e-16 countryLU 0.293899 0.016562 17.745 < 2e-16 countryLV 1.345414 0.009013 149.273 < 2e-16 countryNL 0.242677 0.008008 30.305 < 2e-16 countryPT 1.046255 0.006213 168.388 < 2e-16 countryUK 0.493482 0.008133 60.679 < 2e-16 factor(year)2009 0.354847 0.004838 73.341 < 2e-16 factor(year)2010 0.489904 0.004758 102.958 < 2e-16 factor(year)2011 0.574763 0.004804 119.648 < 2e-16 factor(year)2012 0.758693 0.004666 162.607 < 2e-16 factor(year)2013 0.832287 0.004684 177.706 < 2e-16 factor(year)2014 0.778806 0.004750 163.969 < 2e-16 factor(year)2015 0.699263 0.004851 144.156 < 2e-16 factor(year)2016 0.617126 0.004921 125.405 < 2e-16 factor(year)2017 0.507343 0.005060 100.257 < 2e-16 factor(year)2018 0.401336 0.005155 77.858 < 2e-16 factor(year)2019 0.315577 0.005332 59.186 < 2e-16 sexfemale 0.145584 0.001912 76.126 < 2e-16 age26-34 -0.573652 0.003027 -189.492 < 2e-16 age35-44 -0.981822 0.003110 -315.748 < 2e-16 age45-54 -1.258426 0.002986 -421.452 < 2e-16 age55-65 -1.368932 0.003590 -381.347 < 2e-16 countrybirthEU/EFTA 0.218257 0.193740 1.127 0.25993
countrybirthNew_Member 0.271467 0.193733 1.401 0.16114
countrybirthEurope_nEU/nEFTA 0.439845 0.193701 2.271 0.02316 *
countrybirthNon_Europe 0.508784 0.193669 2.627 0.00861 ** educationISCED3/4 -0.511572 0.002196 -232.982 < 2e-16 educationISCED5+ -1.023143 0.002893 -353.619 < 2e-16 yearsar5-9years -0.085168 0.008246 -10.328 < 2e-16 yearsar>10years -0.001182 0.007130 -0.166 0.86836
yearsarnative -0.037703 0.193733 -0.195 0.84569
compoadult_ch 0.221820 0.006057 36.625 < 2e-16 compocouple -0.436728 0.003905 -111.846 < 2e-16 compocouple_ch -0.463997 0.003577 -129.727 < 2e-16 compoconf_adults -0.021587 0.003286 -6.569 5.06e-11 compoconf_adults_ch -0.056514 0.003756 -15.046 < 2e-16 — Signif. codes: 0 ‘’ 0.001 ’’ 0.01 ’’ 0.05 ‘.’ 0.1 ’ ’ 1

## 
## Correlation matrix not shown by default, as p = 50 > 12.
## Use print(x, correlation=TRUE)  or
##     vcov(x)        if you need it

optimizer (bobyqa) convergence code: 0 (OK) maxfun < 10 * length(par)^2 is not recommended.

Including Plots

You can also embed plots, for example:

Note that the echo = FALSE parameter was added to the code chunk to prevent printing of the R code that generated the plot.