使用 R 包交互 (jtools) 对 merMod 对象进行简单斜率测试

Question

我正在尝试对使用 lmer 获得的混合效应模型进行简单的斜率分析。

模型类似如下：

data(Orthodont,package="nlme")
mod <- lme4::lmer(distance ~ age*Sex + (1|Subject), data=Orthodont)

当尝试使用包 interactions 中的函数 sim_slopes 进行简单的斜率分析时，我收到以下错误消息。

interactions::sim_slopes(model=mod, pred=age, modx=Sex)

Error: One of the requested columns does not exist.
Backtrace:
1. interactions::sim_slopes(model = mod, pred = age, modx = Sex)
5. jtools:::summ.merMod(...)
6. jtools:::create_table(...)
In addition: Warning message:
Johnson-Neyman intervals are not available for factor moderators.

有没有人知道问题出在哪里？

请注意，它与包 reghelper 配合使用效果很好，但我需要这些测试的 p 值，reghelper 没有提供。

reghelper::simple_slopes(mod)

sessionInfo()的输出是：

R version 3.6.0 (2019-04-26) Platform: x86_64-pc-linux-gnu (64-bit) Running under: Ubuntu 18.04.2 LTS

Matrix products: default BLAS:
/usr/lib/x86_64-linux-gnu/blas/libblas.so.3.7.1 LAPACK: /usr/lib/x86_64-linux-gnu/lapack/liblapack.so.3.7.1

locale: [1] LC_CTYPE=en_CA.UTF-8 LC_NUMERIC=C
LC_TIME=en_CA.UTF-8 [4] LC_COLLATE=en_CA.UTF-8
LC_MONETARY=en_CA.UTF-8 LC_MESSAGES=en_CA.UTF-8 [7] LC_PAPER=en_CA.UTF-8 LC_NAME=C LC_ADDRESS=C
[10] LC_TELEPHONE=C LC_MEASUREMENT=en_CA.UTF-8 LC_IDENTIFICATION=C

attached base packages: [1] stats graphics grDevices utils
datasets methods base

loaded via a namespace (and not attached):
[1] Rcpp_1.0.1 magrittr_1.5 splines_3.6.0 MASS_7.3-51.1
[5] munsell_0.5.0 colorspace_1.4-1 lattice_0.20-38 rlang_0.3.4
[9] minqa_1.2.4 plyr_1.8.4 tools_3.6.0 grid_3.6.0
[13] gtable_0.3.0 nlme_3.1-140 cli_1.1.0 assertthat_0.2.1
[17] digest_0.6.19 lme4_1.1-21 lazyeval_0.2.2 tibble_2.1.2
[21] crayon_1.3.4 Matrix_1.2-17 reghelper_0.3.4 nloptr_1.2.1

[25] ggplot2_3.1.1 interactions_1.1.0 jtools_2.0.1 pander_0.6.3
[29] compiler_3.6.0 pillar_1.4.1 scales_1.0.0 boot_1.3-20
[33] pkgconfig_2.0.2

Answer 1

（不是答案，但对于评论来说太长了。如果合适，稍后会删除。）

在干净的 R 会话中，这对我有用：

data(Orthodont,package="nlme")
mod <- lme4::lmer(distance ~ age*Sex + (1|Subject), data=Orthodont)
interactions::sim_slopes(model=mod, pred=age, modx=Sex)
sessionInfo()

SIMPLE SLOPES ANALYSIS
Slope of age when Sex = Female:
Est. S.E. t val. p
------ ------ -------- ------
0.48 0.09 5.13 0.00
Slope of age when Sex = Male:
Est. S.E. t val. p
------ ------ -------- ------
0.78 0.08 10.12 0.00
Warning message: Johnson-Neyman intervals are not available for factor moderators.

这是我的sessionInfo()：重要的东西看起来和你的一样（lme4 1.1-21，interactions 1.1.0，jtools 2.0.1） , 但它肯定不相同...... ??

R Under development (unstable) (2019-05-21 r76566)
Platform: x86_64-pc-linux-gnu (64-bit)
Running under: Ubuntu 16.04.6 LTS

[matrix product and locale info deleted]   

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base     

loaded via a namespace (and not attached):
 [1] Rcpp_1.0.1         magrittr_1.5       splines_3.7.0      MASS_7.3-51.4     
 [5] tidyselect_0.2.5   munsell_0.5.0      colorspace_1.4-1   lattice_0.20-38   
 [9] R6_2.4.0           rlang_0.3.4        minqa_1.2.4        plyr_1.8.4        
[13] dplyr_0.8.1        tools_3.7.0        grid_3.7.0         gtable_0.3.0      
[17] nlme_3.1-140       cli_1.1.0          digest_0.6.19      assertthat_0.2.1  
[21] lme4_1.1-21        lazyeval_0.2.2     tibble_2.1.2       numDeriv_2016.8-1 
[25] crayon_1.3.4       Matrix_1.2-17      purrr_0.3.2        nloptr_1.2.1      
[29] ggplot2_3.1.1      jtools_2.0.1       interactions_1.1.0 glue_1.3.1        
[33] pander_0.6.3       compiler_3.7.0     pillar_1.4.1       scales_1.0.0      
[37] lmerTest_3.1-0     boot_1.3-22        pkgconfig_2.0.2

Answer 2

我通过分离 lmerTest 解决了这个问题。

使用 R 包交互 (jtools) 对 merMod 对象进行简单斜率测试

Simple slopes test for merMod object with R package interactions (jtools)

interaction

r

mixed-models