I am trying to pool the results of multcomp::glht - however, when tidy() does not return terms but contrasts, these are dropped, yielding meaningless results (averaged across contrasts). Should contrast be treated equivalently to term, parameter etc in pool.fitlist?
library(multcomp)
#> Loading required package: mvtnorm
#> Loading required package: survival
#> Loading required package: TH.data
#> Loading required package: MASS
#>
#> Attaching package: 'TH.data'
#> The following object is masked from 'package:MASS':
#>
#> geyser
library(mice)
#>
#> Attaching package: 'mice'
#> The following object is masked from 'package:stats':
#>
#> filter
#> The following objects are masked from 'package:base':
#>
#> cbind, rbind
lmod <- lm(mpg ~ am + cyl, mtcars)
lin_comp <- glht(lmod, linfct = c("am = 0", "cyl = 0"))
tidy(lin_comp)
#> # A tibble: 2 × 6
#> contrast null.value estimate std.error statistic adj.p.value
#> <chr> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 am 0 2.57 1.29 1.99 0.100
#> 2 cyl 0 -2.50 0.361 -6.93 0.000000254
lin_comp_mira <- rep(list(lin_comp), 10) |> as.mira()
pool(lin_comp_mira)
#> [OMITTING LOTS OF WARNINGS RE UNSUPPORTED tidy() ARGUMENTS]
#> Class: mipo m = 10
#> m estimate ubar b t dfcom df riv lambda
#> 1 20 0.03303853 0.8989917 6.759091 7.996037 Inf 24.11837 7.894451 0.8875704
#> fmi
#> 1 0.8958621
I am trying to pool the results of multcomp::glht - however, when
tidy()
does not return terms but contrasts, these are dropped, yielding meaningless results (averaged across contrasts). Shouldcontrast
be treated equivalently toterm
,parameter
etc inpool.fitlist
?Created on 2022-08-31 by the reprex package (v2.0.1)