`conf.level` not working for merMod objects

[strengejacke]

Argument conf.level does not affect the confidence intervals in the output...

library(lme4)
#> Loading required package: Matrix
library(broom.mixed)

m <- glmer(cbind(incidence, size - incidence) ~ period + (1 | herd), cbpp, binomial, nAGQ = 0)

broom.mixed::tidy(m, conf.int = TRUE, conf.level = .95)
#> # A tibble: 5 x 9
#>   effect group term  estimate std.error statistic  p.value conf.low
#>   <chr>  <chr> <chr>    <dbl>     <dbl>     <dbl>    <dbl>    <dbl>
#> 1 fixed  <NA>  (Int~   -1.36      0.228     -5.98  2.26e-9    -1.81
#> 2 fixed  <NA>  peri~   -0.976     0.303     -3.22  1.29e-3    -1.57
#> 3 fixed  <NA>  peri~   -1.11      0.324     -3.43  5.94e-4    -1.75
#> 4 fixed  <NA>  peri~   -1.56      0.424     -3.67  2.38e-4    -2.39
#> 5 ran_p~ herd  sd__~    0.642    NA         NA    NA          NA   
#> # ... with 1 more variable: conf.high <dbl>

broom.mixed::tidy(m, conf.int = TRUE, conf.level = .5)
#> # A tibble: 5 x 9
#>   effect group term  estimate std.error statistic  p.value conf.low
#>   <chr>  <chr> <chr>    <dbl>     <dbl>     <dbl>    <dbl>    <dbl>
#> 1 fixed  <NA>  (Int~   -1.36      0.228     -5.98  2.26e-9    -1.81
#> 2 fixed  <NA>  peri~   -0.976     0.303     -3.22  1.29e-3    -1.57
#> 3 fixed  <NA>  peri~   -1.11      0.324     -3.43  5.94e-4    -1.75
#> 4 fixed  <NA>  peri~   -1.56      0.424     -3.67  2.38e-4    -2.39
#> 5 ran_p~ herd  sd__~    0.642    NA         NA    NA          NA   
#> # ... with 1 more variable: conf.high <dbl>

^{Created on 2018-10-04 by the reprex package (v0.2.1)}

[bbolker]

thanks for this and others, will try to get to them all soon ...

[bbolker]

should be fixed: e9da820

[IndrajeetPatil]

Do you plan to have the next release soon (with all the confidence interval issues raised by @strengejacke fixed)?

After its release on CRAN, I had immediately added broom.mixed as a dependency for two of my packages. But now the results from my package functions that rely on merMod tidiers look incorrect on CRAN and I would like to correct that asap.

[bbolker]

Yes. CRAN wants a new version by October 22 anyway to deal with some dependency issues. The most critical (IMO) of the issues raised by @strengejacke are addressed - all the ones that I would classify as actual bugs - but I was hoping to close a couple more issues (#34, #37) before sending it ...

[IndrajeetPatil]

Cool, sounds good.