1% Generated by roxygen2: do not edit by hand
2% Please edit documentation in R/mfx-tidiers.R
3\name{tidy.betamfx}
4\alias{tidy.betamfx}
5\title{Tidy a(n) betamfx object}
6\usage{
7\method{tidy}{betamfx}(x, conf.int = FALSE, conf.level = 0.95, ...)
8}
9\arguments{
10\item{x}{A \code{betamfx} object.}
11
12\item{conf.int}{Logical indicating whether or not to include a confidence
13interval in the tidied output. Defaults to \code{FALSE}.}
14
15\item{conf.level}{The confidence level to use for the confidence interval
16if \code{conf.int = TRUE}. Must be strictly greater than 0 and less than 1.
17Defaults to 0.95, which corresponds to a 95 percent confidence interval.}
18
19\item{...}{Additional arguments. Not used. Needed to match generic
20signature only. \strong{Cautionary note:} Misspelled arguments will be
21absorbed in \code{...}, where they will be ignored. If the misspelled
22argument has a default value, the default value will be used.
23For example, if you pass \code{conf.lvel = 0.9}, all computation will
24proceed using \code{conf.level = 0.95}. Additionally, if you pass
25\code{newdata = my_tibble} to an \code{\link[=augment]{augment()}} method that does not
26accept a \code{newdata} argument, it will use the default value for
27the \code{data} argument.}
28}
29\description{
30Tidy summarizes information about the components of a model.
31A model component might be a single term in a regression, a single
32hypothesis, a cluster, or a class. Exactly what tidy considers to be a
33model component varies across models but is usually self-evident.
34If a model has several distinct types of components, you will need to
35specify which components to return.
36}
37\details{
38The \code{mfx} package provides methods for calculating marginal effects
39for various generalized linear models (GLMs). Unlike standard linear
40models, estimated model coefficients in a GLM cannot be directly
41interpreted as marginal effects (i.e., the change in the response variable
42predicted after a one unit change in one of the regressors). This is
43because the estimated coefficients are multiplicative, dependent on both
44the link function that was used for the estimation and any other variables
45that were included in the model. When calculating marginal effects, users
46must typically choose whether they want to use i) the average observation
47in the data, or ii) the average of the sample marginal effects. See
48\code{vignette("mfxarticle")} from the \code{mfx} package for more details.
49}
50\examples{
51\dontrun{
52library(mfx)
53
54## Simulate some data
55set.seed(12345)
56n = 1000
57x = rnorm(n)
58
59## Beta outcome
60y = rbeta(n, shape1 = plogis(1 + 0.5 * x), shape2 = (abs(0.2*x)))
61## Use Smithson and Verkuilen correction
62y = (y*(n-1)+0.5)/n
63
64d = data.frame(y,x)
65mod_betamfx = betamfx(y ~ x | x, data = d)
66
67tidy(mod_betamfx, conf.int = TRUE)
68
69## Compare with the naive model coefficients of the equivalent betareg call (not run)
70# tidy(betamfx(y ~ x | x, data = d), conf.int = TRUE)
71
72augment(mod_betamfx)
73glance(mod_betamfx)
74}
75}
76\seealso{
77\code{\link[=tidy.betareg]{tidy.betareg()}}, \code{\link[mfx:betamfx]{mfx::betamfx()}}
78
79Other mfx tidiers:
80\code{\link{augment.betamfx}()},
81\code{\link{augment.mfx}()},
82\code{\link{glance.betamfx}()},
83\code{\link{glance.mfx}()},
84\code{\link{tidy.mfx}()}
85}
86\concept{mfx tidiers}
87\value{
88A \code{\link[tibble:tibble]{tibble::tibble()}} with columns:
89  \item{conf.high}{Upper bound on the confidence interval for the estimate.}
90  \item{conf.low}{Lower bound on the confidence interval for the estimate.}
91  \item{estimate}{The estimated value of the regression term.}
92  \item{p.value}{The two-sided p-value associated with the observed statistic.}
93  \item{statistic}{The value of a T-statistic to use in a hypothesis that the regression term is non-zero.}
94  \item{std.error}{The standard error of the regression term.}
95  \item{term}{The name of the regression term.}
96  \item{atmean}{TRUE if the marginal effects were originally calculated as the
97    partial effects for the average observation. If FALSE, then these were
98    instead calculated as average partial effects.}
99
100}
101