https://github.com/cran/bayestestR
Raw File
Tip revision: d8462ad2168ad7ee61c0d7e679174e775f01a9be authored by Dominique Makowski on 18 January 2020, 07:10:02 UTC
version 0.5.0
Tip revision: d8462ad
rope_range.Rd
% Generated by roxygen2: do not edit by hand
% Please edit documentation in R/rope_range.R
\name{rope_range}
\alias{rope_range}
\alias{rope_range.default}
\title{Find Default Equivalence (ROPE) Region Bounds}
\usage{
rope_range(x, ...)

\method{rope_range}{default}(x, verbose = TRUE, ...)
}
\arguments{
\item{x}{A \code{stanreg}, \code{brmsfit} or \code{BFBayesFactor} object.}

\item{...}{Currently not used.}

\item{verbose}{Toggle warnings.}
}
\description{
This function attempts at automatically finding suitable "default"
values for the Region Of Practical Equivalence (ROPE).
}
\details{
\emph{Kruschke (2018)} suggests that the region of practical equivalence
could be set, by default, to a range from \code{-0.1} to \code{0.1} of a standardized
parameter (negligible effect size according to \emph{Cohen, 1988}).
\itemize{
\item For \strong{linear models (lm)}, this can be generalised to
\ifelse{html}{\out{-0.1 * SD<sub>y</sub>, 0.1 * SD<sub>y</sub>}}{\eqn{[-0.1*SD_{y}, 0.1*SD_{y}]}}.
\item For \strong{logistic models}, the parameters expressed in log odds ratio can be
converted to standardized difference through the formula
\ifelse{html}{\out{&pi;/&radic;(3)}}{\eqn{\pi/\sqrt{3}}}, resulting in a
range of \code{-0.18} to \code{0.18}.
\item For other models with \strong{binary outcome}, it is strongly recommended to
manually specify the rope argument. Currently, the same default is applied
that for logistic models.
\item For models from \strong{count data}, the residual variance is used. This is a
rather experimental threshold and is probably often similar to \verb{-0.1, 0.1},
but should be used with care!
\item For \strong{t-tests}, the standard deviation of the response is used, similarly
to linear models (see above).
\item For \strong{correlations}, \verb{-0.05, 0.05} is used, i.e., half the value of a
negligible correlation as suggested by Cohen's (1988) rules of thumb.
\item For all other models, \verb{-0.1, 0.1} is used to determine the ROPE limits,
but it is strongly advised to specify it manually.
}
}
\examples{
\dontshow{if (require("rstanarm") && require("brms") && require("BayesFactor")) (if (getRversion() >= "3.4") withAutoprint else force)(\{ # examplesIf}
\donttest{
model <- suppressWarnings(rstanarm::stan_glm(
  mpg ~ wt + gear,
  data = mtcars,
  chains = 2,
  iter = 200,
  refresh = 0
))
rope_range(model)

model <- suppressWarnings(
  rstanarm::stan_glm(vs ~ mpg, data = mtcars, family = "binomial", refresh = 0)
)
rope_range(model)

model <- brms::brm(mpg ~ wt + cyl, data = mtcars)
rope_range(model)

model <- BayesFactor::ttestBF(mtcars[mtcars$vs == 1, "mpg"], mtcars[mtcars$vs == 0, "mpg"])
rope_range(model)

model <- lmBF(mpg ~ vs, data = mtcars)
rope_range(model)
}
\dontshow{\}) # examplesIf}
}
\references{
Kruschke, J. K. (2018). Rejecting or accepting parameter values
in Bayesian estimation. Advances in Methods and Practices in Psychological
Science, 1(2), 270-280. \doi{10.1177/2515245918771304}.
}
back to top