Tidy summarizes information about the components of a model. A model component might be a single term in a regression, a single hypothesis, a cluster, or a class. Exactly what tidy considers to be a model component varies across models but is usually self-evident. If a model has several distinct types of components, you will need to specify which components to return.
Usage
# S3 method for class 'margins'
tidy(x, conf.int = FALSE, conf.level = 0.95, ...)Arguments
- x
A
marginsobject returned frommargins::margins().- conf.int
Logical indicating whether or not to include a confidence interval in the tidied output. Defaults to
FALSE.- conf.level
The confidence level to use for the confidence interval if
conf.int = TRUE. Must be strictly greater than 0 and less than 1. Defaults to 0.95, which corresponds to a 95 percent confidence interval.- ...
Additional arguments. Not used. Needed to match generic signature only. Cautionary note: Misspelled arguments will be absorbed in
..., where they will be ignored. If the misspelled argument has a default value, the default value will be used. For example, if you passconf.lvel = 0.9, all computation will proceed usingconf.level = 0.95. Two exceptions here are:
Details
The margins package provides a way to obtain coefficient marginal
effects for a variety of (non-linear) models, such as logit or models with
multiway interaction terms. Note that the glance.margins() method
requires rerunning the underlying model again, which can take some time.
Similarly, an augment.margins() method is not currently supported, but
users can simply run the underlying model to obtain the same information.
Value
A tibble::tibble() with columns:
- conf.high
Upper bound on the confidence interval for the estimate.
- conf.low
Lower bound on the confidence interval for the estimate.
- estimate
The estimated value of the regression term.
- p.value
The two-sided p-value associated with the observed statistic.
- statistic
The value of a T-statistic to use in a hypothesis that the regression term is non-zero.
- std.error
The standard error of the regression term.
- term
The name of the regression term.
Examples
# load libraries for models and data
library(margins)
# example 1: logit model
mod_log <- glm(am ~ cyl + hp + wt, data = mtcars, family = binomial)
# get tidied "naive" model coefficients
tidy(mod_log)
#> # A tibble: 4 × 5
#> term estimate std.error statistic p.value
#> <chr> <dbl> <dbl> <dbl> <dbl>
#> 1 (Intercept) 19.7 8.12 2.43 0.0152
#> 2 cyl 0.488 1.07 0.455 0.649
#> 3 hp 0.0326 0.0189 1.73 0.0840
#> 4 wt -9.15 4.15 -2.20 0.0276
# convert to marginal effects with margins()
marg_log <- margins(mod_log)
# get tidied marginal effects
tidy(marg_log)
#> # A tibble: 3 × 5
#> term estimate std.error statistic p.value
#> <chr> <dbl> <dbl> <dbl> <dbl>
#> 1 cyl 0.0215 0.0470 0.457 0.648
#> 2 hp 0.00143 0.000618 2.32 0.0204
#> 3 wt -0.403 0.115 -3.49 0.000487
tidy(marg_log, conf.int = TRUE)
#> # A tibble: 3 × 7
#> term estimate std.error statistic p.value conf.low conf.high
#> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 cyl 0.0215 0.0470 0.457 0.648 -0.0706 0.114
#> 2 hp 0.00143 0.000618 2.32 0.0204 0.000222 0.00265
#> 3 wt -0.403 0.115 -3.49 0.000487 -0.629 -0.176
# requires running the underlying model again. quick for this example
glance(marg_log)
#> # A tibble: 1 × 8
#> null.deviance df.null logLik AIC BIC deviance df.residual nobs
#> <dbl> <int> <dbl> <dbl> <dbl> <dbl> <int> <int>
#> 1 43.2 31 -4.92 17.8 23.7 9.84 28 32
# augmenting `margins` outputs isn't supported, but
# you can get the same info by running on the underlying model
augment(mod_log)
#> # A tibble: 32 × 11
#> .rownames am cyl hp wt .fitted .resid .hat .sigma
#> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 Mazda RX4 1 6 110 2.62 2.24 0.449 0.278 0.595
#> 2 Mazda RX4 Wag 1 6 110 2.88 -0.0912 1.22 0.352 0.529
#> 3 Datsun 710 1 4 93 2.32 3.46 0.249 0.0960 0.602
#> 4 Hornet 4 Drive 0 6 110 3.22 -3.20 -0.282 0.0945 0.601
#> 5 Hornet Sporta… 0 8 175 3.44 -2.17 -0.466 0.220 0.595
#> 6 Valiant 0 6 105 3.46 -5.61 -0.0856 0.0221 0.604
#> 7 Duster 360 0 8 245 3.57 -1.07 -0.766 0.337 0.576
#> 8 Merc 240D 0 4 62 3.19 -5.51 -0.0897 0.0376 0.603
#> 9 Merc 230 0 4 95 3.15 -4.07 -0.184 0.122 0.603
#> 10 Merc 280 0 6 123 3.44 -4.84 -0.126 0.0375 0.603
#> # ℹ 22 more rows
#> # ℹ 2 more variables: .cooksd <dbl>, .std.resid <dbl>
# example 2: threeway interaction terms
mod_ie <- lm(mpg ~ wt * cyl * disp, data = mtcars)
# get tidied "naive" model coefficients
tidy(mod_ie)
#> # A tibble: 8 × 5
#> term estimate std.error statistic p.value
#> <chr> <dbl> <dbl> <dbl> <dbl>
#> 1 (Intercept) 108. 23.3 4.62 0.000109
#> 2 wt -24.8 8.47 -2.92 0.00744
#> 3 cyl -10.8 4.34 -2.49 0.0201
#> 4 disp -0.593 0.213 -2.79 0.0102
#> 5 wt:cyl 2.91 1.42 2.05 0.0514
#> 6 wt:disp 0.184 0.0685 2.69 0.0127
#> 7 cyl:disp 0.0752 0.0268 2.81 0.00979
#> 8 wt:cyl:disp -0.0233 0.00861 -2.71 0.0123
# convert to marginal effects with margins()
marg_ie0 <- margins(mod_ie)
# get tidied marginal effects
tidy(marg_ie0)
#> # A tibble: 3 × 5
#> term estimate std.error statistic p.value
#> <chr> <dbl> <dbl> <dbl> <dbl>
#> 1 cyl -3.85 1.46 -2.65 0.00812
#> 2 disp -0.0295 0.0174 -1.70 0.0900
#> 3 wt -2.01 1.17 -1.72 0.0860
glance(marg_ie0)
#> # A tibble: 1 × 12
#> r.squared adj.r.squared sigma statistic p.value df logLik AIC
#> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 0.896 0.865 2.21 29.4 2.75e-10 7 -66.2 150.
#> # ℹ 4 more variables: BIC <dbl>, deviance <dbl>, df.residual <int>,
#> # nobs <int>
# marginal effects evaluated at specific values of a variable (here: cyl)
marg_ie1 <- margins(mod_ie, at = list(cyl = c(4,6,8)))
# summarize model fit with tidiers
tidy(marg_ie1)
#> # A tibble: 9 × 7
#> term at.variable at.value estimate std.error statistic p.value
#> <chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 cyl cyl 4 -3.85 1.46 -2.65 0.00808
#> 2 cyl cyl 6 -3.85 1.46 -2.65 0.00814
#> 3 cyl cyl 8 -3.85 1.46 -2.65 0.00812
#> 4 disp cyl 4 0.000978 0.0314 0.0312 0.975
#> 5 disp cyl 6 0.00134 0.0182 0.0737 0.941
#> 6 disp cyl 8 0.00170 0.0120 0.141 0.888
#> 7 wt cyl 4 7.91 5.06 1.56 0.118
#> 8 wt cyl 6 2.96 2.52 1.18 0.239
#> 9 wt cyl 8 -1.98 2.40 -0.825 0.409
# marginal effects of one interaction variable (here: wt), modulated at
# specific values of the two other interaction variables (here: cyl and drat)
marg_ie2 <- margins(mod_ie,
variables = "wt",
at = list(cyl = c(4,6,8), drat = c(3, 3.5, 4)))
# summarize model fit with tidiers
tidy(marg_ie2)
#> # A tibble: 18 × 7
#> term at.variable at.value estimate std.error statistic p.value
#> <chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 wt cyl 4 7.91 5.06 1.56 0.118
#> 2 wt drat 3 7.91 5.06 1.56 0.118
#> 3 wt cyl 4 7.91 5.06 1.56 0.118
#> 4 wt drat 3.5 7.91 5.06 1.56 0.118
#> 5 wt cyl 4 7.91 5.06 1.56 0.118
#> 6 wt drat 4 7.91 5.06 1.56 0.118
#> 7 wt cyl 6 2.96 2.52 1.18 0.239
#> 8 wt drat 3 2.96 2.52 1.18 0.239
#> 9 wt cyl 6 2.96 2.52 1.18 0.239
#> 10 wt drat 3.5 2.96 2.52 1.18 0.239
#> 11 wt cyl 6 2.96 2.52 1.18 0.239
#> 12 wt drat 4 2.96 2.52 1.18 0.239
#> 13 wt cyl 8 -1.98 2.40 -0.825 0.409
#> 14 wt drat 3 -1.98 2.40 -0.825 0.409
#> 15 wt cyl 8 -1.98 2.40 -0.825 0.409
#> 16 wt drat 3.5 -1.98 2.40 -0.825 0.409
#> 17 wt cyl 8 -1.98 2.40 -0.825 0.409
#> 18 wt drat 4 -1.98 2.40 -0.825 0.409