How can I calculate the marginal effects for each group level of random effects-CodePudding

I'm trying to figure out if I can use a multilevel model for a project, and I need to simulate some data, fit a random effects model then generate the marginal effects for each group (participant or Subject in my case). However I cannot seem to generate varying marginal effects from my model fit, despite it matching a working example (e.g., the sleepstudy).

The sleepstudy provides a good starting example to work from:

library(tidyverse)
library(lme4)
library(ggeffects)

data("sleepstudy")
m <- lmer(Reaction ~ Days   (1   Days | Subject), data = sleepstudy)
me <- ggpredict(m, terms = c("Days", "Subject [sample=8]"), type = "random")
plot(me)

The plot shows eight different slopes from eight different Subjects

However if I simulate data which approximately matches the sleep study, my model fit doesn't generate varying slopes and I don't know why not:

N = 18 # Number of people
Nt = 9 # Number of trials

d <- tibble(
  Subject = factor(sprintf("d", 1:N)),       # create subject numbers
  subj_b0 = rnorm(n = N, mean = 250, sd = 20),  # create random intercepts
  subj_b1 = rnorm(n = N, mean = 10, sd = 6)     # create random slopes       
) %>%                                           
  mutate(Days = list(0:Nt)) %>%
  unnest(Days) %>%
  mutate(
    Y = subj_b0   Days*subj_b1,
    Y = Y   rnorm(n = nrow(.), sd = 15)          # Add error
  )    

fit <- lmer(Y ~ 1   Days   (1   Days|Subject), data = d)
p <- ggpredict(fit, terms = c("Days", "Subject [sample=8]"), type = "random")
plot(p)

I can't for the life of me figure out why ggpredict() is returning the same slope for each Subject.

The model summary confirms there is variation around the slope for days, and as best I can tell the other estimated parameters are similar to the sleepstudy.

summary(fit)

Linear mixed model fit by REML ['lmerMod']
Formula: Y ~ 1   Days   (1   Days | Subject)
   Data: d

REML criterion at convergence: 1553.3

Scaled residuals: 
     Min       1Q   Median       3Q      Max 
-2.57322 -0.57070 -0.01073  0.59463  3.02611 

Random effects:
 Groups   Name        Variance Std.Dev. Corr
 Subject  (Intercept) 745.22   27.30        
          Days         37.95    6.16    0.12
 Residual             176.54   13.29        
Number of obs: 180, groups:  Subject, 18

Fixed effects:
            Estimate Std. Error t value
(Intercept)  254.517      6.693  38.030
Days           9.033      1.492   6.053

sleepstudy model summary for comparison:

summary(m)

Linear mixed model fit by REML ['lmerMod']
Formula: Reaction ~ Days   (1   Days | Subject)
   Data: sleepstudy

REML criterion at convergence: 1743.6

Scaled residuals: 
    Min      1Q  Median      3Q     Max 
-3.9536 -0.4634  0.0231  0.4634  5.1793 

Random effects:
 Groups   Name        Variance Std.Dev. Corr
 Subject  (Intercept) 612.10   24.741       
          Days         35.07    5.922   0.07
 Residual             654.94   25.592       
Number of obs: 180, groups:  Subject, 18

Fixed effects:
            Estimate Std. Error t value
(Intercept)  251.405      6.825  36.838
Days          10.467      1.546   6.771

CodePudding user response：

It seems that ggpredict() cannot handle factors (as random effects), which are "numeric", but have levels like 001, 002 etc. This seems to be a bug, and I will see how to fix this.

The workaround would be to change the levels for Subject:

library(ggeffects)
library(easystats)
library(lme4)

N = 18 # Number of people
Nt = 9 # Number of trials

d <- data.frame(
  Subject = factor(1:N),                        # create subject numbers
  subj_b0 = rnorm(n = N, mean = 250, sd = 20),  # create random intercepts
  subj_b1 = rnorm(n = N, mean = 10, sd = 6)     # create random slopes       
) |>
  data_modify(Days = list(0:Nt)) |>
  tidyr::unnest(Days) |>
  data_modify(
    Y = subj_b0   Days*subj_b1,
    Y = Y   rnorm(n = N * (Nt   1), sd = 15)          # Add error
  )

fit <- lmer(Y ~ 1   Days   (1   Days|Subject), data = d)
p <- ggpredict(fit, terms = c("Days", "Subject [sample=8]"), type = "random")
plot(p)

^{Created on 2023-11-03 with reprex v2.0.2}