When extracting random effects from an lmer model using estimate_grouplevel(), the formatting of factor levels in the Parameter column is inconsistent. The first instance of a factor's random slope gets cleaned up nicely (e.g., treatment [1]), but all subsequent rows for that same parameter revert to the raw lme4 naming convention (e.g., treatment1).
library(lme4)
#> Loading required package: Matrix
library(easystats)
#> # Attaching packages: easystats 0.7.5
#> ✔ bayestestR 0.17.0 ✔ correlation 0.8.8
#> ✔ datawizard 1.3.0 ✔ effectsize 1.0.1
#> ✔ insight 1.4.6 ✔ modelbased 0.14.0
#> ✔ performance 0.16.0 ✔ parameters 0.28.3
#> ✔ report 0.6.3 ✔ see 0.13.0
# Prepare built-in data with a 0/1 categorical predictor
dat <- sleepstudy |>
dplyr::mutate(treatment = factor(ifelse(Days > 4, "1", "0")))
# Fit a model with a random slope for the factor
fit <- lmer(Reaction ~ treatment + (1 + treatment | Subject), data = dat)
# Extract random effects
# Notice Subject 308 gets "treatment [1]" while 309+ get "treatment1"
estimate_grouplevel(fit, type = "random")
#> Group | Level | Parameter | Coefficient | SE | 95% CI
#> ------------------------------------------------------------------------
#> Subject | 308 | (Intercept) | 21.70 | 10.38 | [ 1.36, 42.04]
#> Subject | 308 | treatment [1] | 42.41 | 14.18 | [ 14.61, 70.20]
#> Subject | 309 | (Intercept) | -55.24 | 10.38 | [-75.58, -34.90]
#> Subject | 309 | treatment1 | -44.64 | 14.18 | [-72.44, -16.85]
#> Subject | 310 | (Intercept) | -45.89 | 10.38 | [-66.23, -25.55]
#> Subject | 310 | treatment1 | -33.27 | 14.18 | [-61.06, -5.48]
#> Subject | 330 | (Intercept) | 11.64 | 10.38 | [ -8.70, 31.98]
#> Subject | 330 | treatment1 | -19.84 | 14.18 | [-47.64, 7.95]
#> Subject | 331 | (Intercept) | 16.66 | 10.38 | [ -3.68, 37.00]
#> Subject | 331 | treatment1 | -18.86 | 14.18 | [-46.66, 8.93]
#> Subject | 332 | (Intercept) | 4.69 | 10.38 | [-15.65, 25.03]
#> Subject | 332 | treatment1 | 7.72 | 14.18 | [-20.08, 35.51]
#> Subject | 333 | (Intercept) | 13.59 | 10.38 | [ -6.75, 33.93]
#> Subject | 333 | treatment1 | 4.51 | 14.18 | [-23.28, 32.31]
#> Subject | 334 | (Intercept) | -4.43 | 10.38 | [-24.77, 15.91]
#> Subject | 334 | treatment1 | 4.34 | 14.18 | [-23.45, 32.14]
#> Subject | 335 | (Intercept) | -22.23 | 10.38 | [-42.57, -1.89]
#> Subject | 335 | treatment1 | -51.99 | 14.18 | [-79.78, -24.19]
#> Subject | 337 | (Intercept) | 49.50 | 10.38 | [ 29.16, 69.84]
#> Subject | 337 | treatment1 | 45.91 | 14.18 | [ 18.11, 73.70]
#> Subject | 349 | (Intercept) | -20.23 | 10.38 | [-40.58, 0.11]
#> Subject | 349 | treatment1 | 1.49 | 14.18 | [-26.30, 29.29]
#> Subject | 350 | (Intercept) | -1.37 | 10.38 | [-21.72, 18.97]
#> Subject | 350 | treatment1 | 38.02 | 14.18 | [ 10.22, 65.81]
#> Subject | 351 | (Intercept) | -1.65 | 10.38 | [-21.99, 18.69]
#> Subject | 351 | treatment1 | -14.84 | 14.18 | [-42.63, 12.96]
#> Subject | 352 | (Intercept) | 28.95 | 10.38 | [ 8.61, 49.29]
#> Subject | 352 | treatment1 | 12.62 | 14.18 | [-15.17, 40.42]
#> Subject | 369 | (Intercept) | 5.29 | 10.38 | [-15.05, 25.63]
#> Subject | 369 | treatment1 | 3.26 | 14.18 | [-24.53, 31.06]
#> Subject | 370 | (Intercept) | -15.14 | 10.38 | [-35.48, 5.20]
#> Subject | 370 | treatment1 | 24.26 | 14.18 | [ -3.53, 52.06]
#> Subject | 371 | (Intercept) | 0.73 | 10.38 | [-19.61, 21.07]
#> Subject | 371 | treatment1 | -9.96 | 14.18 | [-37.75, 17.84]
#> Subject | 372 | (Intercept) | 13.44 | 10.38 | [ -6.90, 33.78]
#> Subject | 372 | treatment1 | 8.86 | 14.18 | [-18.94, 36.65]
Created on 2026-03-01 with reprex v2.1.1
Expected Behavior
The Parameter column should apply the factor [level] formatting consistently across all grouping levels rather than just the first one.
When extracting random effects from an
lmermodel usingestimate_grouplevel(), the formatting of factor levels in theParametercolumn is inconsistent. The first instance of a factor's random slope gets cleaned up nicely (e.g.,treatment [1]), but all subsequent rows for that same parameter revert to the rawlme4naming convention (e.g.,treatment1).Created on 2026-03-01 with reprex v2.1.1
Expected Behavior
The
Parametercolumn should apply thefactor [level]formatting consistently across all grouping levels rather than just the first one.