Add option to pull coef() instead of ranef() for random effects by strengejacke · Pull Request #1068 · easystats/parameters

strengejacke · 2025-03-02T18:29:49Z

strengejacke · 2025-03-02T21:47:16Z

data("fish", package = "insight")

m1 <- suppressWarnings(glmmTMB::glmmTMB(
  count ~ child + camper + (1 | ID),
  data = fish,
  family = poisson()
))

m2 <- suppressWarnings(glmmTMB::glmmTMB(
  count ~ child + camper + (1 + xb | persons) + (1 + zg | ID),
  ziformula = ~ child + livebait + (1 + zg + nofish | ID),
  dispformula = ~xb,
  data = fish,
  family = glmmTMB::truncated_poisson()
))

m3 <- suppressWarnings(glmmTMB::glmmTMB(
  count ~ child + camper,
  ziformula = ~ child + livebait + (1 | ID),
  data = fish,
  family = glmmTMB::truncated_poisson()
))

m4 <- suppressWarnings(glmmTMB::glmmTMB(
  count ~ child + camper + (1 + xb | persons),
  ziformula = ~ child + livebait,
  dispformula = ~xb,
  data = fish,
  family = glmmTMB::truncated_poisson()
))

m5 <- suppressWarnings(lme4::glmer(
  count ~ child + camper + (1 | ID),
  data = fish,
  family = poisson()
))

m6 <- suppressWarnings(lme4::lmer(
  Reaction ~ Days + (1 + Days | Subject),
  data = lme4::sleepstudy
))


parameters::model_parameters(m1, effects = "total")
#> Group | Level |   Parameter | Coefficient |   Component | Effects
#> -----------------------------------------------------------------
#> ID    |     1 | (Intercept) |        0.22 | conditional |   total
#> ID    |     2 | (Intercept) |        0.77 | conditional |   total
#> ID    |     3 | (Intercept) |        0.86 | conditional |   total
#> ID    |     4 | (Intercept) |        1.36 | conditional |   total

parameters::model_parameters(m2, effects = "total")
#> Group   | Level |   Parameter | Coefficient |     Component | Effects
#> ---------------------------------------------------------------------
#> persons |     1 | (Intercept) |       -1.20 |   conditional |   total
#> persons |     1 |          xb |        1.43 |   conditional |   total
#> persons |     2 | (Intercept) |       -0.88 |   conditional |   total
#> persons |     2 |          xb |        1.30 |   conditional |   total
#> persons |     3 | (Intercept) |       -0.88 |   conditional |   total
#> persons |     3 |          xb |        1.30 |   conditional |   total
#> persons |     4 | (Intercept) |        0.23 |   conditional |   total
#> persons |     4 |          xb |        0.87 |   conditional |   total
#> ID      |     1 | (Intercept) |        1.94 |   conditional |   total
#> ID      |     1 |          zg |        0.20 |   conditional |   total
#> ID      |     2 | (Intercept) |        2.06 |   conditional |   total
#> ID      |     2 |          zg |        0.15 |   conditional |   total
#> ID      |     3 | (Intercept) |        2.35 |   conditional |   total
#> ID      |     3 |          zg |        0.04 |   conditional |   total
#> ID      |     4 | (Intercept) |        1.52 |   conditional |   total
#> ID      |     4 |          zg |        0.36 |   conditional |   total
#> ID      |     1 | (Intercept) |        1.92 | zero_inflated |   total
#> ID      |     1 |          zg |       -0.97 | zero_inflated |   total
#> ID      |     1 |     nofish1 |    2.59e-18 | zero_inflated |   total
#> ID      |     2 | (Intercept) |        1.58 | zero_inflated |   total
#> ID      |     2 |          zg |       -0.67 | zero_inflated |   total
#> ID      |     2 |     nofish1 |    1.77e-18 | zero_inflated |   total
#> ID      |     3 | (Intercept) |        1.38 | zero_inflated |   total
#> ID      |     3 |          zg |       -0.50 | zero_inflated |   total
#> ID      |     3 |     nofish1 |    1.33e-18 | zero_inflated |   total
#> ID      |     4 | (Intercept) |        1.63 | zero_inflated |   total
#> ID      |     4 |          zg |       -0.71 | zero_inflated |   total
#> ID      |     4 |     nofish1 |    1.86e-18 | zero_inflated |   total

parameters::model_parameters(m3, effects = "total")
#> Group | Level |   Parameter | Coefficient |     Component | Effects
#> -------------------------------------------------------------------
#> ID    |     1 | (Intercept) |        0.30 | zero_inflated |   total
#> ID    |     2 | (Intercept) |        0.30 | zero_inflated |   total
#> ID    |     3 | (Intercept) |        0.30 | zero_inflated |   total
#> ID    |     4 | (Intercept) |        0.30 | zero_inflated |   total

parameters::model_parameters(m4, effects = "total")
#> Group   | Level |   Parameter | Coefficient |   Component | Effects
#> -------------------------------------------------------------------
#> persons |     1 | (Intercept) |       -1.23 | conditional |   total
#> persons |     1 |          xb |        1.35 | conditional |   total
#> persons |     2 | (Intercept) |       -0.87 | conditional |   total
#> persons |     2 |          xb |        1.22 | conditional |   total
#> persons |     3 | (Intercept) |       -1.11 | conditional |   total
#> persons |     3 |          xb |        1.30 | conditional |   total
#> persons |     4 | (Intercept) |        0.10 | conditional |   total
#> persons |     4 |          xb |        0.87 | conditional |   total

parameters::model_parameters(m5, effects = "total")
#> Group | Level |   Parameter | Coefficient | Effects
#> ---------------------------------------------------
#> ID    |     1 | (Intercept) |        0.22 |   total
#> ID    |     2 | (Intercept) |        0.77 |   total
#> ID    |     3 | (Intercept) |        0.86 |   total
#> ID    |     4 | (Intercept) |        1.36 |   total

parameters::model_parameters(m6, effects = "total")
#> Group   | Level |   Parameter | Coefficient | Effects
#> -----------------------------------------------------
#> Subject |   308 | (Intercept) |      253.66 |   total
#> Subject |   308 |        Days |       19.67 |   total
#> Subject |   309 | (Intercept) |      211.01 |   total
#> Subject |   309 |        Days |        1.85 |   total
#> Subject |   310 | (Intercept) |      212.44 |   total
#> Subject |   310 |        Days |        5.02 |   total
#> Subject |   330 | (Intercept) |      275.10 |   total
#> Subject |   330 |        Days |        5.65 |   total
#> Subject |   331 | (Intercept) |      273.67 |   total
#> Subject |   331 |        Days |        7.40 |   total
#> Subject |   332 | (Intercept) |      260.44 |   total
#> Subject |   332 |        Days |       10.20 |   total
#> Subject |   333 | (Intercept) |      268.25 |   total
#> Subject |   333 |        Days |       10.24 |   total
#> Subject |   334 | (Intercept) |      244.17 |   total
#> Subject |   334 |        Days |       11.54 |   total
#> Subject |   335 | (Intercept) |      251.07 |   total
#> Subject |   335 |        Days |       -0.28 |   total
#> Subject |   337 | (Intercept) |      286.30 |   total
#> Subject |   337 |        Days |       19.10 |   total
#> Subject |   349 | (Intercept) |      226.19 |   total
#> Subject |   349 |        Days |       11.64 |   total
#> Subject |   350 | (Intercept) |      238.34 |   total
#> Subject |   350 |        Days |       17.08 |   total
#> Subject |   351 | (Intercept) |      255.98 |   total
#> Subject |   351 |        Days |        7.45 |   total
#> Subject |   352 | (Intercept) |      272.27 |   total
#> Subject |   352 |        Days |       14.00 |   total
#> Subject |   369 | (Intercept) |      254.68 |   total
#> Subject |   369 |        Days |       11.34 |   total
#> Subject |   370 | (Intercept) |      225.79 |   total
#> Subject |   370 |        Days |       15.29 |   total
#> Subject |   371 | (Intercept) |      252.21 |   total
#> Subject |   371 |        Days |        9.48 |   total
#> Subject |   372 | (Intercept) |      263.72 |   total
#> Subject |   372 |        Days |       11.75 |   total

^{Created on 2025-03-02 with reprex v2.1.1}

DominiqueMakowski · 2025-03-02T22:05:07Z

I'd call it "random_total" because it still filters out fixed effects (and it's something that applies to random effects)

strengejacke · 2025-03-02T22:06:59Z

Yeah, the name is not final yet. But what about the general output?

strengejacke · 2025-03-02T22:07:23Z

And we need methods for rstanarm and brms

DominiqueMakowski · 2025-03-02T22:09:57Z

We need to also keep the SE/CI for lme4 and other models, it is a glmmTMB particularity to not yet computing them i think

strengejacke · 2025-03-02T22:21:10Z

I don't think you can get SE/CI for the coef()?

DominiqueMakowski · 2025-03-03T08:22:29Z

It seems you're right, somehow I (mis)remembered that it was possible for lme4

(Unless with bootstrapping... that could be a longer-term feature)

https://cran.r-project.org/web/packages/merTools/vignettes/Using_predictInterval.html

DominiqueMakowski · 2025-03-03T08:35:51Z

Note: In https://hypatia.math.ethz.ch/pipermail/r-help/2007-November/145442.html, Doug Bates calles the "default" random effects (from ranef()) "conditional"

So we could have:

effects="random" (which is technically "random_conditional")
effects="random_total" OR what about effects="random_unbiased" (to link it to the BLUP terminology)

bwiernik · 2025-03-03T12:08:40Z

I'm not a fan of "random_total". I think that's kind of confusing. If something other than "total", my preference is "group_total".

R/group_level_total.R

strengejacke · 2025-03-03T13:31:13Z

brms works (at least this complex example, I bet there are dozens of not yet considered exceptions)

library(brms)
library(parameters)

x <- insight::download_model("brms_zi_4")

### Model

#>  Family: zero_inflated_poisson 
#>   Links: mu = log; zi = logit 
#> Formula: count ~ child + camper + (1 + xb | persons) + (1 + zg | ID) 
#>          zi ~ child + livebait + (1 + zg + nofish | ID)
#>    Data: fish (Number of observations: 250) 
#>   Draws: 4 chains, each with iter = 2000; warmup = 1000; thin = 1;
#>          total post-warmup draws = 4000

### Here are the over all coefficients

coef(x)
#> $ID
#> , , Intercept
#> 
#>     Estimate Est.Error      Q2.5    Q97.5
#> 1 -0.7872119  1.042311 -2.700848 1.478536
#> 2 -0.6960301  1.036887 -2.593541 1.606411
#> 3 -0.3863236  1.028798 -2.281176 1.859450
#> 4 -1.1993232  1.035853 -3.146711 1.089575
#> 
#> , , zg
#> 
#>     Estimate  Est.Error        Q2.5     Q97.5
#> 1 0.15976724 0.08921813 -0.01071167 0.3196564
#> 2 0.24963216 0.08264031  0.08816210 0.4097783
#> 3 0.08353011 0.07200988 -0.04812716 0.2280746
#> 4 0.36529945 0.04519791  0.27709860 0.4533866
#> 
#> , , zi_Intercept
#> 
#>    Estimate Est.Error      Q2.5      Q97.5
#> 1 -7.296146  6.817151 -22.84229  0.5702874
#> 2 -9.930848  8.403730 -29.46093 -0.7190626
#> 3 -9.532446  8.373872 -28.81624  0.2686403
#> 4 -9.502189  8.415209 -30.05622  0.1605108
#> 
#> , , zi_zg
#> 
#>    Estimate Est.Error      Q2.5      Q97.5
#> 1 -8.514177  6.120933 -25.27990 -1.9568175
#> 2 -6.563639  4.558390 -16.05121 -1.8399470
#> 3 -8.958289  5.989657 -21.74333 -2.6511637
#> 4 -4.571465  3.546923 -13.44316 -0.8261003
#> 
#> , , zi_nofish1
#> 
#>      Estimate Est.Error      Q2.5    Q97.5
#> 1 -0.69059771  2.558553 -6.878615 3.391735
#> 2 -0.66113253  1.886437 -5.312619 2.324461
#> 3 -0.55772179  2.194537 -6.212487 3.026720
#> 4 -0.08506966  1.992193 -4.468332 4.176201
#> 
#> , , child
#> 
#>     Estimate Est.Error      Q2.5      Q97.5
#> 1 -0.4544075 0.1220026 -0.698228 -0.2138667
#> 2 -0.4544075 0.1220026 -0.698228 -0.2138667
#> 3 -0.4544075 0.1220026 -0.698228 -0.2138667
#> 4 -0.4544075 0.1220026 -0.698228 -0.2138667
#> 
#> , , camper1
#> 
#>    Estimate Est.Error        Q2.5     Q97.5
#> 1 0.1078043 0.1110276 -0.09501827 0.3259446
#> 2 0.1078043 0.1110276 -0.09501827 0.3259446
#> 3 0.1078043 0.1110276 -0.09501827 0.3259446
#> 4 0.1078043 0.1110276 -0.09501827 0.3259446
#> 
#> , , zi_child
#> 
#>     Estimate Est.Error      Q2.5    Q97.5
#> 1 -0.7358728  3.085961 -7.267424 4.137575
#> 2 -0.7358728  3.085961 -7.267424 4.137575
#> 3 -0.7358728  3.085961 -7.267424 4.137575
#> 4 -0.7358728  3.085961 -7.267424 4.137575
#> 
#> , , zi_livebait1
#> 
#>   Estimate Est.Error    Q2.5    Q97.5
#> 1 2.377975  5.663644 -5.9263 16.07785
#> 2 2.377975  5.663644 -5.9263 16.07785
#> 3 2.377975  5.663644 -5.9263 16.07785
#> 4 2.377975  5.663644 -5.9263 16.07785
#> 
#> 
#> $persons
#> , , Intercept
#> 
#>     Estimate Est.Error      Q2.5      Q97.5
#> 1 -1.4726950 0.5559270 -2.558071 -0.4086148
#> 2 -1.1021684 0.5289535 -2.150847 -0.1149768
#> 3 -1.2686065 0.5395648 -2.328640 -0.2619923
#> 4 -0.1746891 0.5134628 -1.198719  0.7418030
#> 
#> , , xb
#> 
#>    Estimate  Est.Error      Q2.5     Q97.5
#> 1 1.5626166 0.20201624 1.1732615 1.9904038
#> 2 1.2953383 0.09746668 1.0999387 1.4829048
#> 3 1.3196510 0.09650849 1.1370000 1.5147897
#> 4 0.8891383 0.04823144 0.7929538 0.9809271
#> 
#> , , zi_Intercept
#> 
#>    Estimate Est.Error      Q2.5    Q97.5
#> 1 -3.854074  4.813789 -14.58513 4.052241
#> 2 -3.854074  4.813789 -14.58513 4.052241
#> 3 -3.854074  4.813789 -14.58513 4.052241
#> 4 -3.854074  4.813789 -14.58513 4.052241
#> 
#> , , child
#> 
#>     Estimate Est.Error      Q2.5      Q97.5
#> 1 -0.4544075 0.1220026 -0.698228 -0.2138667
#> 2 -0.4544075 0.1220026 -0.698228 -0.2138667
#> 3 -0.4544075 0.1220026 -0.698228 -0.2138667
#> 4 -0.4544075 0.1220026 -0.698228 -0.2138667
#> 
#> , , camper1
#> 
#>    Estimate Est.Error        Q2.5     Q97.5
#> 1 0.1078043 0.1110276 -0.09501827 0.3259446
#> 2 0.1078043 0.1110276 -0.09501827 0.3259446
#> 3 0.1078043 0.1110276 -0.09501827 0.3259446
#> 4 0.1078043 0.1110276 -0.09501827 0.3259446
#> 
#> , , zi_child
#> 
#>     Estimate Est.Error      Q2.5    Q97.5
#> 1 -0.7358728  3.085961 -7.267424 4.137575
#> 2 -0.7358728  3.085961 -7.267424 4.137575
#> 3 -0.7358728  3.085961 -7.267424 4.137575
#> 4 -0.7358728  3.085961 -7.267424 4.137575
#> 
#> , , zi_livebait1
#> 
#>   Estimate Est.Error    Q2.5    Q97.5
#> 1 2.377975  5.663644 -5.9263 16.07785
#> 2 2.377975  5.663644 -5.9263 16.07785
#> 3 2.377975  5.663644 -5.9263 16.07785
#> 4 2.377975  5.663644 -5.9263 16.07785

### formatted output from parameters

model_parameters(x, effects = "total")
#> Group   | Level | Parameter | Median |          95% CI |     pd |     Component | Effects
#> -----------------------------------------------------------------------------------------
#> persons |     1 | Intercept |  -1.45 | [ -2.56, -0.41] | 99.62% |   conditional |   total
#> persons |     2 | Intercept |  -1.07 | [ -2.15, -0.11] | 98.35% |   conditional |   total
#> persons |     3 | Intercept |  -1.27 | [ -2.33, -0.26] | 98.72% |   conditional |   total
#> persons |     4 | Intercept |  -0.13 | [ -1.20,  0.74] | 62.60% |   conditional |   total
#> persons |     1 |        xb |   1.57 | [  1.17,  1.99] |   100% |   conditional |   total
#> persons |     2 |        xb |   1.30 | [  1.10,  1.48] |   100% |   conditional |   total
#> persons |     3 |        xb |   1.32 | [  1.14,  1.51] |   100% |   conditional |   total
#> persons |     4 |        xb |   0.89 | [  0.79,  0.98] |   100% |   conditional |   total
#> persons |     1 | Intercept |  -3.56 | [-14.59,  4.05] | 83.65% | zero_inflated |   total
#> persons |     2 | Intercept |  -3.56 | [-14.59,  4.05] | 83.65% | zero_inflated |   total
#> persons |     3 | Intercept |  -3.56 | [-14.59,  4.05] | 83.65% | zero_inflated |   total
#> persons |     4 | Intercept |  -3.56 | [-14.59,  4.05] | 83.65% | zero_inflated |   total
#> ID      |     1 | Intercept |  -0.89 | [ -2.70,  1.48] | 79.97% |   conditional |   total
#> ID      |     2 | Intercept |  -0.81 | [ -2.59,  1.61] | 78.05% |   conditional |   total
#> ID      |     3 | Intercept |  -0.53 | [ -2.28,  1.86] | 70.40% |   conditional |   total
#> ID      |     4 | Intercept |  -1.32 | [ -3.15,  1.09] | 87.22% |   conditional |   total
#> ID      |     1 |        zg |   0.16 | [ -0.01,  0.32] | 96.67% |   conditional |   total
#> ID      |     2 |        zg |   0.25 | [  0.09,  0.41] | 99.85% |   conditional |   total
#> ID      |     3 |        zg |   0.08 | [ -0.05,  0.23] | 86.28% |   conditional |   total
#> ID      |     4 |        zg |   0.37 | [  0.28,  0.45] |   100% |   conditional |   total
#> ID      |     1 | Intercept |  -5.98 | [-22.84,  0.57] | 95.97% | zero_inflated |   total
#> ID      |     2 | Intercept |  -7.43 | [-29.46, -0.72] | 98.72% | zero_inflated |   total
#> ID      |     3 | Intercept |  -7.23 | [-28.82,  0.27] | 96.85% | zero_inflated |   total
#> ID      |     4 | Intercept |  -7.13 | [-30.06,  0.16] | 97.20% | zero_inflated |   total
#> ID      |     1 |        zg |  -7.23 | [-25.28, -1.96] |   100% | zero_inflated |   total
#> ID      |     3 |        zg |  -7.13 | [-21.74, -2.65] |   100% | zero_inflated |   total
#> ID      |     2 |   nofish1 |  -0.23 | [ -5.31,  2.32] | 61.22% | zero_inflated |   total
#> ID      |     4 |   nofish1 |  -0.08 | [ -4.47,  4.18] | 54.62% | zero_inflated |   total

DominiqueMakowski · 2025-03-03T13:47:04Z

In this example the zi term is on the conditional component, but what if you have it on say another auxiliarity component, in a model like:

bf(Resp ~ Formula1,
   sigma ~ Formula_with_zi)

strengejacke · 2025-03-03T13:49:39Z

From ?lme4:::coef.merMod:

I still would prefer "total" for now.

strengejacke · 2025-03-03T13:50:42Z

In this example the zi term is on the conditional component, but what if you have it on say another auxiliarity component, in a model like:
bf(Resp ~ Formula1,
   sigma ~ Formula_with_zi)

How can we make a "nested" formula? (i.e. sigma ~ zi_formula)?

bwiernik · 2025-03-03T14:12:50Z

In this example the zi term is on the conditional component, but what if you have it on say another auxiliarity component, in a model like:
bf(Resp ~ Formula1,

sigma ~ Formula_with_zi)
How can we make a "nested" formula? (i.e. sigma ~ zi_formula)?

You need to use brms' nonlinear formula syntax. See the last example here https://rdrr.io/cran/brms/man/brmsformula.html

DominiqueMakowski · 2025-03-03T14:29:53Z

data <- iris
data$Group <- as.factor(rep(c("G1", "G2", "G3"), each = 50))

m4 <- brms::brm(brms::bf(Sepal.Width ~ Petal.Width + (Petal.Width | Group),
     sigma ~ Petal.Width + (Petal.Width | Group)),
    data = data, refresh=0)

strengejacke · 2025-03-03T15:19:15Z

data <- iris
data$Group <- as.factor(rep(c("G1", "G2", "G3"), each = 50))

m4 <- brms::brm(brms::bf(Sepal.Width ~ Petal.Width + (Petal.Width | Group),
     sigma ~ Petal.Width + (Petal.Width | Group)),
    data = data, refresh=0)

That's no zero-inflation formula.

DominiqueMakowski · 2025-03-03T15:22:30Z

I wonder if the zi parts should not be a different Parameter rather than Component

strengejacke · 2025-03-03T15:22:39Z

Assuming that auxiliary components have no zero-inflation components, that stuff should work now, too. I bet there are more cases that won't work and which are hard to anticipate, so we can "fix on demand".

strengejacke · 2025-03-03T15:40:04Z

I wonder if the zi parts should not be a different Parameter rather than Component

That would be a different behaviour than what we usually have across the easystats verse.

strengejacke · 2025-03-03T15:41:14Z

x <- insight::download_model("brms_zi_4")
parameters::model_parameters(x)
#> # Fixed Effects
#> 
#> Parameter   | Median |         95% CI |     pd |  Rhat |     ESS
#> ----------------------------------------------------------------
#> (Intercept) |  -0.67 | [-2.54,  1.78] | 72.47% | 1.020 |  214.00
#> child       |  -0.45 | [-0.70, -0.21] |   100% | 1.001 | 2571.00
#> camper1     |   0.11 | [-0.10,  0.33] | 81.30% | 1.025 |  174.00
#> 
#> # Zero-Inflation
#> 
#> Parameter   | Median |          95% CI |     pd |  Rhat |     ESS
#> -----------------------------------------------------------------
#> (Intercept) |  -3.56 | [-14.59,  4.05] | 83.65% | 1.003 | 1061.00
#> child       |  -0.63 | [ -7.27,  4.14] | 59.80% | 1.035 |  103.00
#> livebait1   |   1.52 | [ -5.93, 16.08] | 67.30% | 1.006 |  648.00
#> 
#> Uncertainty intervals (equal-tailed) computed using a MCMC distribution
#>   approximation.
#> 
#> The model has a log- or logit-link. Consider using `exponentiate =
#>   TRUE` to interpret coefficients as ratios.

^{Created on 2025-03-03 with reprex v2.1.1}

DominiqueMakowski · 2025-03-03T18:04:22Z

NEWS.md

+## Changes
+
+* The `effects` argument in `model_parameters()` for classes `merMod`, `glmmTMB`,
+  `brmsfit` and `stanreg` gets an additional `"total"` option, to return the


I think now that "random_unbiased" would be the best name...

Yeah, it's not finalized yet, we can change the name, if we find a name we're all happy with.

@ your component comment: We always had it composed like this, where the parameter names are "clean" and the related component is indicated in the Component column.

Should we now revise estimate_grouplevel()? It seems it could just be a wrapper around model_parameters(..., effects = "random_<whatever>"), or is estimate_grouplevel() doing more? The implementation here supports lme4/glmmTMB/rstanarm/brms.

But random_unbiased suggests the other effects are biased. But they're just deviations from the overall mean. Just curious, why are you so hesitant with "total" or "overall", since it's really just summing up fixed+random effects...

And: we should have the same options for the type argument in estimate_grouplevel(), or not?

Should we now revise estimate_grouplevel()?

Yes!

But random_unbiased

The main thing is to avoid effects="total" because it's too vague. Since it concerns random effects, I'd at least name it "random_total" so that it's clear what's it about

Now about unbiased, well I'm not an expert of the underlying mathematics, but:

"It simply computes the "total" of random + fixed": in general yes, but I'm not 100% that is always exactly the case. As @bwiernik mentioned, sometimes there is no fixed effect and I'm not sure what happens then. Also, In the coef.brmsfit code, there seem to be some edgecases with some models etc. etc. So I'm not sure it is always just as simple as the sum.

"unbiased": initially I wasn't a fan either. But 1) it has the benefit of directly linking to the now well-established literature about BLUPs (best linear unbiased). And one could understand "biased" as "relative to the fixed effect", whereas "unbiased" means "absolute". Technically we could even consider "random_relative" and "random_absolute" 🤷

Do BLUPs refer to the sum of fixed and random effects? I thought it was just a description of the estimation method of random effects in general?

When there are not fixef, it's still the sum, with fixef's being zero:

m1 <- lme4::lmer(Reaction ~ Days + (Days | Subject), lme4::sleepstudy) m2 <- lme4::lmer(Reaction ~ 1 + (Days | Subject), lme4::sleepstudy) coef(m1) #> $Subject #> (Intercept) Days #> 308 253.6637 19.6662617 #> 309 211.0064 1.8476053 #> 310 212.4447 5.0184295 #> 330 275.0957 5.6529356 #> 331 273.6654 7.3973743 #> 332 260.4447 10.1951090 #> 333 268.2456 10.2436499 #> 334 244.1725 11.5418676 #> 335 251.0714 -0.2848792 #> 337 286.2956 19.0955511 #> 349 226.1949 11.6407181 #> 350 238.3351 17.0815038 #> 351 255.9830 7.4520239 #> 352 272.2688 14.0032871 #> 369 254.6806 11.3395008 #> 370 225.7921 15.2897709 #> 371 252.2122 9.4791297 #> 372 263.7197 11.7513080 #> #> attr(,"class") #> [1] "coef.mer" lme4::ranef(m1) #> $Subject #> (Intercept) Days #> 308 2.2585509 9.1989758 #> 309 -40.3987381 -8.6196806 #> 310 -38.9604090 -5.4488565 #> 330 23.6906196 -4.8143503 #> 331 22.2603126 -3.0699116 #> 332 9.0395679 -0.2721770 #> 333 16.8405086 -0.2236361 #> 334 -7.2326151 1.0745816 #> 335 -0.3336684 -10.7521652 #> 337 34.8904868 8.6282652 #> 349 -25.2102286 1.1734322 #> 350 -13.0700342 6.6142178 #> 351 4.5778642 -3.0152621 #> 352 20.8636782 3.5360011 #> 369 3.2754656 0.8722149 #> 370 -25.6129993 4.8224850 #> 371 0.8070461 -0.9881562 #> 372 12.3145921 1.2840221 #> #> with conditional variances for "Subject" lme4::fixef(m1) #> (Intercept) Days #> 251.40510 10.46729 coef(m2) #> $Subject #> Days (Intercept) #> 308 20.6010269 249.4572 #> 309 0.2135211 218.3598 #> 310 3.8918664 217.5142 #> 330 4.1171150 282.0069 #> 331 6.1589652 279.2383 #> 332 9.5231905 263.4683 #> 333 9.5157838 271.5210 #> 334 11.2238002 245.6038 #> 335 -2.5963842 261.4733 #> 337 19.6695601 283.7125 #> 349 11.4860944 226.8907 #> 350 17.7185058 235.4685 #> 351 6.3674017 260.8638 #> 352 13.8582027 272.9217 #> 369 10.9022191 256.6483 #> 370 15.7360708 223.7837 #> 371 8.7573841 255.4600 #> 372 11.3074308 265.7171 #> #> attr(,"class") #> [1] "coef.mer" lme4::ranef(m2) #> $Subject #> (Intercept) Days #> 308 -8.304993 20.6010269 #> 309 -39.402398 0.2135211 #> 310 -40.247926 3.8918664 #> 330 24.244775 4.1171150 #> 331 21.476108 6.1589652 #> 332 5.706141 9.5231905 #> 333 13.758848 9.5157838 #> 334 -12.158381 11.2238002 #> 335 3.711084 -2.5963842 #> 337 25.950356 19.6695601 #> 349 -30.871497 11.4860944 #> 350 -22.293637 17.7185058 #> 351 3.101614 6.3674017 #> 352 15.159485 13.8582027 #> 369 -1.113833 10.9022191 #> 370 -33.978438 15.7360708 #> 371 -2.302158 8.7573841 #> 372 7.954974 11.3074308 #> #> with conditional variances for "Subject" lme4::fixef(m2) #> (Intercept) #> 257.7622

^{Created on 2025-03-03 with reprex v2.1.1}

So why is it called BLUPs then? Why such a complicated name for "simply" taking the random parameters

They are the best linear unbiased predictions of the unknown parameters (random or total slope coefficients), the same way that the OLS coefficients are the best linear unbiased estimates (BLUEs) of those unknown coefficients.

That all makes sense. Incredible that I never found any clear clarification in a lot of reading. AI companies should trains their LLMs on the easystats' github issues & PRs it's a goldmine of knowledge lol

Then:

effects="random"

effects="random_total"

it's a goldmine of knowledge lol

😂

strengejacke added 9 commits March 2, 2025 19:29

Add option to pull coef() instead of ranef() for random effects

d5ed019

Fixes #1067

version

bc6310c

fix

3c3d411

fix

f969ad8

fix

4fd4a4a

implement for glmmTMB

f7b175c

remove deprecated

a8d5d9f

docs

41f2940

docs

4d5001f

strengejacke mentioned this pull request Mar 2, 2025

Add option to pull coef() instead of ranef() for random effects #1067

Closed

strengejacke added 4 commits March 2, 2025 23:24

styler, wordlist

484122d

styler

76fb2e8

update snapshot

a21f0f5

add test

3252d15

strengejacke mentioned this pull request Mar 3, 2025

CRAN update before 2025-03-24 easystats/modelbased#430

Closed

brms

725ef7b

strengejacke added 5 commits March 3, 2025 14:16

add brms method

16d3e01

fix

8953dce

fix

bafb30d

fix

c2a91fc

brms implementation

2e599c4

DominiqueMakowski reviewed Mar 3, 2025

View reviewed changes

R/group_level_total.R Show resolved Hide resolved

add missing component

f20a8cd

add rstanarm method

6eed492

add rstanarm method

cc0197d

strengejacke added 2 commits March 3, 2025 14:51

news

8440615

setup air

9c01d30

brms auxilary

720f2e4

add test

e888c46

mini styler

02b3c10

strengejacke added 4 commits March 3, 2025 16:43

fix

9a8bdf8

Apple sucks

23431f1

Use remotes

1a65f02

Update snap

8d4ee41

strengejacke merged commit 90b6d4c into main Mar 3, 2025
16 of 23 checks passed

strengejacke deleted the strengejacke/issue1067 branch March 3, 2025 17:59

DominiqueMakowski reviewed Mar 3, 2025

View reviewed changes

DominiqueMakowski mentioned this pull request Mar 3, 2025

Possible bug with estimate_grouplevel() for brms model easystats/modelbased#229

Closed

Uh oh!

Conversation

strengejacke commented Mar 2, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

strengejacke commented Mar 2, 2025

Uh oh!

DominiqueMakowski commented Mar 2, 2025

Uh oh!

strengejacke commented Mar 2, 2025

Uh oh!

strengejacke commented Mar 2, 2025

Uh oh!

DominiqueMakowski commented Mar 2, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

strengejacke commented Mar 2, 2025

Uh oh!

DominiqueMakowski commented Mar 3, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

DominiqueMakowski commented Mar 3, 2025

Uh oh!

bwiernik commented Mar 3, 2025

Uh oh!

Uh oh!

strengejacke commented Mar 3, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

DominiqueMakowski commented Mar 3, 2025

Uh oh!

strengejacke commented Mar 3, 2025

Uh oh!

strengejacke commented Mar 3, 2025

Uh oh!

bwiernik commented Mar 3, 2025

Uh oh!

DominiqueMakowski commented Mar 3, 2025

Uh oh!

strengejacke commented Mar 3, 2025

Uh oh!

DominiqueMakowski commented Mar 3, 2025

Uh oh!

strengejacke commented Mar 3, 2025

Uh oh!

strengejacke commented Mar 3, 2025

Uh oh!

strengejacke commented Mar 3, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

strengejacke commented Mar 2, 2025 •

edited

Loading

DominiqueMakowski commented Mar 2, 2025 •

edited

Loading

DominiqueMakowski commented Mar 3, 2025 •

edited

Loading

strengejacke commented Mar 3, 2025 •

edited

Loading

strengejacke commented Mar 3, 2025 •

edited

Loading