R, ggplot2, jtools: johnson neyman plot 错误

Question

我正在拼命尝试为以下交互绘制 Johnson-Neyman 图：

nxsc_20 <-lm(meandec20 ~ centered_nep*centered_selfcontrol + factor(study), data = allstudies_wide)

我得到以下输出：

Coefficients:
                                  Estimate Std. Error t value Pr(>|t|)    
(Intercept)                        0.28264    0.02260  12.508  < 2e-16 ***
centered_nep                       0.08998    0.01192   7.551 7.41e-14 ***
centered_selfcontrol               0.01894    0.01021   1.856  0.06364 .  
factor(study)2                     0.03462    0.02531   1.368  0.17146    
factor(study)3                     0.35767    0.02635  13.573  < 2e-16 ***
factor(study)4                     0.33224    0.03709   8.956  < 2e-16 ***
centered_nep:centered_selfcontrol  0.03706    0.01300   2.850  0.00443 **

现在我尝试制作 JN-Plot

johnson_neyman(nxsc_20, meandec20, centered_selfcontrol, alpha = 0.05, plot = TRUE)

我收到以下错误：

Fehler in vmat[pred, pred] : Indizierung außerhalb der Grenzen (Error in vmat[pred, pred] : indexing out of range)

谁能帮我解决这个问题？

非常感谢！

Answer 1

pred = 选项不适用于响应变量，而是用于您要在 y 轴上绘制的预测变量。这将起作用：

library(interactions)

allstudies_wide = data.frame(meandec20=rnorm(500),centered_nep = runif(500),
centered_selfcontrol = runif(500), study = sample(1:4,500,replace=TRUE))

nxsc_20 <-lm(meandec20 ~ centered_nep*centered_selfcontrol + factor(study), 
data = allstudies_wide)

johnson_neyman(model = nxsc_20, pred = centered_nep,modx = centered_selfcontrol)

R, ggplot2, jtools: johnson neyman plot 错误

R, ggplot2, jtools: johnson neyman plot error

plot

interaction

r

ggplot2