Plotting the percentage in ROPE of the difference between BSEM mediation effects

[JulianGaviriaL]

I implemented the following steps:
Step1. BSEM regression on 6 variables
Step2. Computing the Mediation effects of the LV (Abeh, Nbeh) on PA_1~ TA_4 and PN_1 ~ TN_4 respectively
Step3. Comparing those mediation effects from Step2.

Question how can I plot/visualize the percentage in ROPE of the differences obtained in the Step3? Such as it's illustrated here:
https://easystats.github.io/bayestestR/articles/region_of_practical_equivalence.html

Many thanks in advance

#Step1
a1 <- "
# LV Measurement Model
  Abeh =~ 1*aff_cC + b2*aff_iC + b3*aff_iI + b4*aff_cI # Behavioral LV definitions
  Nbeh =~ 1*neu_cC + b2*neu_iC + b3*neu_iI + b4*neu_cI

 # Inter-context covariances
 Abeh ~~ Nbeh
  PA_1 ~~ PN_1 
  TA_4 ~~ TN_4
 
  # Structural Model
  PA_1 ~ Abeh 
  TA_4  ~ PA_1  + Abeh
  
  PN_1 ~  Nbeh
  TN_4  ~ PN_1  + Nbeh
"
fModel.a1. <- blavaan(theModel.aim1, data=d, n.chains=1, burnin=1000, sample=5000, bcontrol=list(cores=4))

#Step2
effects.a1 <- data.frame(
  PA_1.dir.Abeh = mchain1[, "bet_sign[13]"],
  Abeh.dir.TA_4 = mchain1[, "bet_sign[9]"],
  PA_1.dir.TA_4 = mchain1[, "bet_sign[14]"],
  PA_1.med.TA_4 = mchain1[, "bet_sign[1]"] * mchain1[, "bet_sign[9]"],
  PA_1.tot.TA_4 = mchain1[, "bet_sign[1]"] * mchain1[, "bet_sign[9]"] + mchain1[, "bet_sign[13]"],

  PN_1.dir.Nbeh = mchain1[, "bet_sign[30]"],
  Nbeh.dir.TN_4 = mchain1[, "bet_sign[35]"],
  PN_1.dir.TN_4 = mchain1[, "bet_sign[40]"],
  PN_1.med.TN_4 = mchain1[, "bet_sign[5]"] * mchain1[, "bet_sign[29]"],
  PN_1.tot.TN_4 = mchain1[, "bet_sign[5]"] * mchain1[, "bet_sign[29]"] + mchain1[, "bet_sign[00]"],
)

#Step3
compare.a1 <- data.frame(
P_1.dir.beh.Pdiff = effects.a1$PN_1.dir.Nbeh - effects.a1$PN_1.dir.Nbeh, 
beh.dir.T_4.Pdiff = effects.a1$Nbeh.dir.TN_4 - effects.a1$Abeh.dir.TA_4, 
P_1.dir.T_4.Pdiff = effects.a1$PN_1.dir.TN_4 - effects.a1$PA_1.dir.TA_4,
med.Pdiff = effects.a1$PN_1.med.TN_4 - effects.a1$PA_1.med.TA_4 , 
tot.Pdiff = effects.a1$PN_1.tot.TN_4 - effects.a1$PA_1.tot.TA_4 , 
)
res.compare.a1 <- describe_posterior(compare.a1,
                                     centrality = "all",
                                     dispersion = TRUE,
                                     ci = 0.95,
                                     ci_method = "hdi",
                                     effects = "all", 
                                     test = c("p_direction", "p_significance", "rope", "equitest"))

[mattansb]

Seeing as how you have everything in a data frame already:

library(bayestestR)

res <- data.frame(A = rnorm(3000),
                  B = rnorm(3000))

(rp <- rope(res, range = c(0, 0.5)))
#> # Proportion of samples inside the ROPE [0.00, 0.50]:
#> 
#> Parameter | inside ROPE
#> -----------------------
#> A         |     21.79 %
#> B         |     20.98 %
plot(rp)

^{Created on 2022-06-27 by the reprex package (v2.0.1)}