Comparisons Between Predictions Made With Different Regressor Values
Source:R/comparisons.R
comparisons.Rd
Predict the outcome variable at different regressor values (e.g., college
graduates vs. others), and compare those predictions by computing a difference,
ratio, or some other function. comparisons()
can return many quantities of
interest, such as contrasts, differences, risk ratios, changes in log odds,
slopes, elasticities, etc.
comparisons()
: unit-level (conditional) estimates.avg_comparisons()
: average (marginal) estimates.
The newdata
argument and the datagrid()
function can be used to control where statistics are evaluated in the predictor space: "at observed values", "at the mean", "at representative values", etc.
See the comparisons vignette and package website for worked examples and case studies:
Usage
comparisons(
model,
newdata = NULL,
variables = NULL,
type = NULL,
vcov = TRUE,
by = FALSE,
conf_level = 0.95,
transform_pre = "difference",
transform_post = NULL,
cross = FALSE,
wts = NULL,
hypothesis = NULL,
df = Inf,
eps = NULL,
...
)
avg_comparisons(
model,
newdata = NULL,
variables = NULL,
type = NULL,
vcov = TRUE,
by = TRUE,
conf_level = 0.95,
transform_pre = "difference",
transform_post = NULL,
cross = FALSE,
wts = NULL,
hypothesis = NULL,
df = Inf,
eps = NULL,
...
)
Arguments
- model
Model object
- newdata
NULL
, data frame, string, ordatagrid()
call. Determines the predictor values for which to compute contrasts.NULL
(default): Unit-level contrasts for each observed value in the original dataset.data frame: Unit-level contrasts for each row of the
newdata
data frame.string:
"mean": Contrasts at the Mean. Contrasts when each predictor is held at its mean or mode.
"median": Contrasts at the Median. Contrasts when each predictor is held at its median or mode.
"marginalmeans": Contrasts at Marginal Means.
"tukey": Contrasts at Tukey's 5 numbers.
"grid": Contrasts on a grid of representative numbers (Tukey's 5 numbers and unique values of categorical predictors).
datagrid()
call to specify a custom grid of regressors. For example:newdata = datagrid(cyl = c(4, 6))
:cyl
variable equal to 4 and 6 and other regressors fixed at their means or modes.newdata = datagrid(mpg = fivenum)
:mpg
variable held at Tukey's five numbers (using thefivenum
function), and other regressors fixed at their means or modes.See the Examples section and the datagrid documentation.
- variables
NULL
, character vector, or named list. The subset of variables for which to compute contrasts.NULL
: compute contrasts for all the variables in the model object (can be slow).Character vector: subset of variables (usually faster).
Named list: names identify the subset of variables of interest, and values define the type of contrast to compute. Acceptable values depend on the variable type:
Factor or character variables:
"reference": Each factor level is compared to the factor reference (base) level
"all": All combinations of observed levels
"sequential": Each factor level is compared to the previous factor level
"pairwise": Each factor level is compared to all other levels
Vector of length 2 with the two values to compare.
Logical variables:
NULL: contrast between TRUE and FALSE
Numeric variables:
Numeric of length 1: Contrast for a gap of
x
, computed at the observed value plus and minusx / 2
. For example, estimating a+1
contrast compares adjusted predictions when the regressor is equal to its observed value minus 0.5 and its observed value plus 0.5.Numeric vector of length 2: Contrast between the 2nd element and the 1st element of the
x
vector.Function which accepts a numeric vector and returns a data frame with two columns of "low" and "high" values to compare. See examples below.
"iqr": Contrast across the interquartile range of the regressor.
"sd": Contrast across one standard deviation around the regressor mean.
"2sd": Contrast across two standard deviations around the regressor mean.
"minmax": Contrast between the maximum and the minimum values of the regressor.
Examples:
variables = list(gear = "pairwise", hp = 10)
variables = list(gear = "sequential", hp = c(100, 120))
See the Examples section below for more.
- type
string indicates the type (scale) of the predictions used to compute contrasts or slopes. This can differ based on the model type, but will typically be a string such as: "response", "link", "probs", or "zero". When an unsupported string is entered, the model-specific list of acceptable values is returned in an error message. When
type
isNULL
, the default value is used. This default is the first model-related row in themarginaleffects:::type_dictionary
dataframe.- vcov
Type of uncertainty estimates to report (e.g., for robust standard errors). Acceptable values:
FALSE: Do not compute standard errors. This can speed up computation considerably.
TRUE: Unit-level standard errors using the default
vcov(model)
variance-covariance matrix.String which indicates the kind of uncertainty estimates to return.
Heteroskedasticity-consistent:
"HC"
,"HC0"
,"HC1"
,"HC2"
,"HC3"
,"HC4"
,"HC4m"
,"HC5"
. See?sandwich::vcovHC
Heteroskedasticity and autocorrelation consistent:
"HAC"
Mixed-Models degrees of freedom: "satterthwaite", "kenward-roger"
Other:
"NeweyWest"
,"KernHAC"
,"OPG"
. See thesandwich
package documentation.
One-sided formula which indicates the name of cluster variables (e.g.,
~unit_id
). This formula is passed to thecluster
argument of thesandwich::vcovCL
function.Square covariance matrix
Function which returns a covariance matrix (e.g.,
stats::vcov(model)
)
- by
Aggregate unit-level estimates (aka, marginalize, average over). Valid inputs:
FALSE
: return the original unit-level estimates.TRUE
: aggregate estimates for each term.Character vector of column names in
newdata
or in the data frame produced by calling the function without theby
argument.Data frame with a
by
column of group labels, and merging columns shared bynewdata
or the data frame produced by calling the same function without theby
argument.See examples below.
- conf_level
numeric value between 0 and 1. Confidence level to use to build a confidence interval.
- transform_pre
string or function. How should pairs of adjusted predictions be contrasted?
string: shortcuts to common contrast functions.
Supported shortcuts strings: difference, differenceavg, differenceavgwts, dydx, eyex, eydx, dyex, dydxavg, eyexavg, eydxavg, dyexavg, dydxavgwts, eyexavgwts, eydxavgwts, dyexavgwts, ratio, ratioavg, ratioavgwts, lnratio, lnratioavg, lnratioavgwts, lnor, lnoravg, lnoravgwts, expdydx, expdydxavg, expdydxavgwts
See the Transformations section below for definitions of each transformation.
function: accept two equal-length numeric vectors of adjusted predictions (
hi
andlo
) and returns a vector of contrasts of the same length, or a unique numeric value.See the Transformations section below for examples of valid functions.
- transform_post
string or function. Transformation applied to unit-level estimates and confidence intervals just before the function returns results. Functions must accept a vector and return a vector of the same length. Support string shortcuts: "exp", "ln"
- cross
TRUE or FALSE
FALSE
: Contrasts represent the change in adjusted predictions when one predictor changes and all other variables are held constant.TRUE
: Contrasts represent the changes in adjusted predictions when all the predictors specified in thevariables
argument are manipulated simultaneously (a "cross-contrast").
- wts
string or numeric: weights to use when computing average contrasts or slopes. These weights only affect the averaging in
avg_*()
or with theby
argument, and not the unit-level estimates themselves.string: column name of the weights variable in
newdata
. When supplying a column name towts
, it is recommended to supply the original data (including the weights variable) explicitly tonewdata
.numeric: vector of length equal to the number of rows in the original data or in
newdata
(if supplied).
- hypothesis
specify a hypothesis test or custom contrast using a numeric value, vector, or matrix, a string, or a string formula.
Numeric:
Single value: the null hypothesis used in the computation of Z and p (before applying
transform_post
).Vector: Weights to compute a linear combination of (custom contrast between) estimates. Length equal to the number of rows generated by the same function call, but without the
hypothesis
argument.Matrix: Each column is a vector of weights, as describe above, used to compute a distinct linear combination of (contrast between) estimates. The column names of the matrix are used as labels in the output.
String formula to specify linear or non-linear hypothesis tests. If the
term
column uniquely identifies rows, terms can be used in the formula. Otherwise, useb1
,b2
, etc. to identify the position of each parameter. Examples:hp = drat
hp + drat = 12
b1 + b2 + b3 = 0
String:
"pairwise": pairwise differences between estimates in each row.
"reference": differences between the estimates in each row and the estimate in the first row.
"sequential": difference between an estimate and the estimate in the next row.
"revpairwise", "revreference", "revsequential": inverse of the corresponding hypotheses, as described above.
See the Examples section below and the vignette: https://vincentarelbundock.github.io/marginaleffects/articles/hypothesis.html
- df
Degrees of freedom used to compute p values and confidence intervals. A single numeric value between 1 and
Inf
. Whendf
isInf
, the normal distribution is used. Whendf
is finite, thet
distribution is used. See insight::get_df for a convenient function to extract degrees of freedom. Ex:slopes(model, df = insight::get_df(model))
- eps
NULL or numeric value which determines the step size to use when calculating numerical derivatives: (f(x+eps)-f(x))/eps. When
eps
isNULL
, the step size is 0.0001 multiplied by the difference between the maximum and minimum values of the variable with respect to which we are taking the derivative. Changingeps
may be necessary to avoid numerical problems in certain models.- ...
Additional arguments are passed to the
predict()
method supplied by the modeling package.These arguments are particularly useful for mixed-effects or bayesian models (see the online vignettes on themarginaleffects
website). Available arguments can vary from model to model, depending on the range of supported arguments by each modeling package. See the "Model-Specific Arguments" section of the?marginaleffects
documentation for a non-exhaustive list of available arguments.
Model-Specific Arguments
Some model types allow model-specific arguments to modify the nature of marginal effects, predictions, marginal means, and contrasts.
Package | Class | Argument | Documentation |
brms | brmsfit | ndraws | brms::posterior_predict |
re_formula | |||
lme4 | merMod | include_random | insight::get_predicted |
re.form | lme4::predict.merMod | ||
allow.new.levels | lme4::predict.merMod | ||
glmmTMB | glmmTMB | re.form | glmmTMB::predict.glmmTMB |
allow.new.levels | glmmTMB::predict.glmmTMB | ||
zitype | glmmTMB::predict.glmmTMB | ||
mgcv | bam | exclude | mgcv::predict.bam |
robustlmm | rlmerMod | re.form | robustlmm::predict.rlmerMod |
allow.new.levels | robustlmm::predict.rlmerMod |
Transformations
The following transformations can be applied by supplying one of the shortcut strings to the
transform_pre
argument.
hi
is a vector of adjusted predictions for the "high" side of the
contrast. lo
is a vector of adjusted predictions for the "low" side of the
contrast. y
is a vector of adjusted predictions for the original data. x
is the predictor in the original data. eps
is the step size to use to
compute derivatives and elasticities.
Shortcut | Function |
difference | \(hi, lo) hi - lo |
differenceavg | \(hi, lo) mean(hi) - mean(lo) |
differenceavgwts | \(hi, lo, w) wmean(hi, w) - wmean(lo, w) |
dydx | \(hi, lo, eps) (hi - lo)/eps |
eyex | \(hi, lo, eps, y, x) (hi - lo)/eps * (x/y) |
eydx | \(hi, lo, eps, y, x) ((hi - lo)/eps)/y |
dyex | \(hi, lo, eps, x) ((hi - lo)/eps) * x |
dydxavg | \(hi, lo, eps) mean((hi - lo)/eps) |
eyexavg | \(hi, lo, eps, y, x) mean((hi - lo)/eps * (x/y)) |
eydxavg | \(hi, lo, eps, y, x) mean(((hi - lo)/eps)/y) |
dyexavg | \(hi, lo, eps, x) mean(((hi - lo)/eps) * x) |
dydxavgwts | \(hi, lo, eps, w) wmean((hi - lo)/eps, w) |
eyexavgwts | \(hi, lo, eps, y, x, w) wmean((hi - lo)/eps * (x/y), w) |
eydxavgwts | \(hi, lo, eps, y, x, w) wmean(((hi - lo)/eps)/y, w) |
dyexavgwts | \(hi, lo, eps, x, w) wmean(((hi - lo)/eps) * x, w) |
ratio | \(hi, lo) hi/lo |
ratioavg | \(hi, lo) mean(hi)/mean(lo) |
ratioavgwts | \(hi, lo) wmean(hi)/wmean(lo) |
lnratio | \(hi, lo) log(hi/lo) |
lnratioavg | \(hi, lo) log(mean(hi)/mean(lo)) |
lnratioavgwts | \(hi, lo) log(wmean(hi)/wmean(lo)) |
lnor | \(hi, lo) log((hi/(1 - hi))/(lo/(1 - lo))) |
lnoravg | \(hi, lo) log((mean(hi)/(1 - mean(hi)))/(mean(lo)/(1 - mean(lo)))) |
lnoravgwts | \(hi, lo, w) log((wmean(hi, w)/(1 - wmean(hi, w)))/(wmean(lo, w)/(1 - wmean(lo, w)))) |
expdydx | \(hi, lo, eps) ((exp(hi) - exp(lo))/exp(eps))/eps |
expdydxavg | \(hi, lo, eps) mean(((exp(hi) - exp(lo))/exp(eps))/eps) |
expdydxavgwts | \(hi, lo, eps, w) wmean(((exp(hi) - exp(lo))/exp(eps))/eps, w) |
Bayesian posterior summaries
By default, credible intervals in bayesian models are built as equal-tailed intervals. This can be changed to a highest density interval by setting a global option:
options("marginaleffects_posterior_interval" = "eti")
options("marginaleffects_posterior_interval" = "hdi")
By default, the center of the posterior distribution in bayesian models is identified by the median. Users can use a different summary function by setting a global option:
options("marginaleffects_posterior_center" = "mean")
options("marginaleffects_posterior_center" = "median")
When estimates are averaged using the by
argument, the tidy()
function, or
the summary()
function, the posterior distribution is marginalized twice over.
First, we take the average across units but within each iteration of the
MCMC chain, according to what the user requested in by
argument or
tidy()/summary()
functions. Then, we identify the center of the resulting
posterior using the function supplied to the
"marginaleffects_posterior_center"
option (the median by default).
Examples
library(marginaleffects)
# Linear model
tmp <- mtcars
tmp$am <- as.logical(tmp$am)
mod <- lm(mpg ~ am + factor(cyl), tmp)
comparisons(mod, variables = list(cyl = "reference")) |> tidy()
#> # A tibble: 2 × 9
#> type term contrast estim…¹ std.e…² stati…³ p.value conf.…⁴ conf.…⁵
#> <chr> <chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 response cyl mean(6) - mea… -6.16 1.54 -4.01 6.11e- 5 -9.17 -3.15
#> 2 response cyl mean(8) - mea… -10.1 1.45 -6.93 4.11e-12 -12.9 -7.22
#> # … with abbreviated variable names ¹estimate, ²std.error, ³statistic,
#> # ⁴conf.low, ⁵conf.high
comparisons(mod, variables = list(cyl = "sequential")) |> tidy()
#> # A tibble: 2 × 9
#> type term contrast estim…¹ std.e…² stati…³ p.value conf.…⁴ conf.…⁵
#> <chr> <chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 response cyl mean(6) - mean… -6.16 1.54 -4.01 6.11e-5 -9.17 -3.15
#> 2 response cyl mean(8) - mean… -3.91 1.47 -2.66 7.81e-3 -6.79 -1.03
#> # … with abbreviated variable names ¹estimate, ²std.error, ³statistic,
#> # ⁴conf.low, ⁵conf.high
comparisons(mod, variables = list(cyl = "pairwise")) |> tidy()
#> # A tibble: 3 × 9
#> type term contrast estim…¹ std.e…² stati…³ p.value conf.…⁴ conf.…⁵
#> <chr> <chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 response cyl mean(6) - mea… -6.16 1.54 -4.01 6.11e- 5 -9.17 -3.15
#> 2 response cyl mean(8) - mea… -10.1 1.45 -6.93 4.11e-12 -12.9 -7.22
#> 3 response cyl mean(8) - mea… -3.91 1.47 -2.66 7.81e- 3 -6.79 -1.03
#> # … with abbreviated variable names ¹estimate, ²std.error, ³statistic,
#> # ⁴conf.low, ⁵conf.high
# GLM with different scale types
mod <- glm(am ~ factor(gear), data = mtcars)
comparisons(mod, type = "response") |> tidy()
#> # A tibble: 2 × 9
#> type term contrast estim…¹ std.e…² stati…³ p.value conf.…⁴ conf.…⁵
#> <chr> <chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 response gear mean(4) - mea… 0.667 0.117 5.68 1.38e- 8 0.436 0.897
#> 2 response gear mean(5) - mea… 1 0.157 6.39 1.70e-10 0.693 1.31
#> # … with abbreviated variable names ¹estimate, ²std.error, ³statistic,
#> # ⁴conf.low, ⁵conf.high
comparisons(mod, type = "link") |> tidy()
#> # A tibble: 2 × 9
#> type term contrast estim…¹ std.e…² stati…³ p.value conf.…⁴ conf.…⁵
#> <chr> <chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 link gear mean(4) - mean(3) 0.667 0.117 5.68 1.38e- 8 0.436 0.897
#> 2 link gear mean(5) - mean(3) 1 0.157 6.39 1.70e-10 0.693 1.31
#> # … with abbreviated variable names ¹estimate, ²std.error, ³statistic,
#> # ⁴conf.low, ⁵conf.high
# Contrasts at the mean
comparisons(mod, newdata = "mean")
#>
#> Term Contrast Estimate Std. Error z Pr(>|z|) 2.5 % 97.5 %
#> gear 4 - 3 0.6667 0.1174 5.676 1.3751e-08 0.4365 0.8969
#> gear 5 - 3 1.0000 0.1566 6.386 1.7026e-10 0.6931 1.3069
#>
#> Prediction type: response
#> Columns: rowid, type, term, contrast, estimate, std.error, statistic, p.value, conf.low, conf.high, predicted, predicted_hi, predicted_lo, am, gear
#>
# Contrasts between marginal means
comparisons(mod, newdata = "marginalmeans")
#>
#> Term Contrast Estimate Std. Error z Pr(>|z|) 2.5 % 97.5 %
#> gear 4 - 3 0.6667 0.1174 5.676 1.3751e-08 0.4365 0.8969
#> gear 5 - 3 1.0000 0.1566 6.386 1.7026e-10 0.6931 1.3069
#>
#> Prediction type: response
#> Columns: type, term, contrast, estimate, std.error, statistic, p.value, conf.low, conf.high, predicted, predicted_hi, predicted_lo
#>
# Contrasts at user-specified values
comparisons(mod, newdata = datagrid(am = 0, gear = tmp$gear))
#>
#> Term Contrast Estimate Std. Error z Pr(>|z|) 2.5 % 97.5 % am gear
#> gear 4 - 3 0.6667 0.1174 5.676 1.3751e-08 0.4365 0.8969 0 4
#> gear 4 - 3 0.6667 0.1174 5.676 1.3751e-08 0.4365 0.8969 0 3
#> gear 4 - 3 0.6667 0.1174 5.676 1.3751e-08 0.4365 0.8969 0 5
#> gear 5 - 3 1.0000 0.1566 6.386 1.7026e-10 0.6931 1.3069 0 4
#> gear 5 - 3 1.0000 0.1566 6.386 1.7026e-10 0.6931 1.3069 0 3
#> gear 5 - 3 1.0000 0.1566 6.386 1.7026e-10 0.6931 1.3069 0 5
#>
#> Prediction type: response
#> Columns: rowid, type, term, contrast, estimate, std.error, statistic, p.value, conf.low, conf.high, predicted, predicted_hi, predicted_lo, am, gear
#>
comparisons(mod, newdata = datagrid(am = unique, gear = max))
#>
#> Term Contrast Estimate Std. Error z Pr(>|z|) 2.5 % 97.5 % am gear
#> gear 4 - 3 0.6667 0.1174 5.676 1.3751e-08 0.4365 0.8969 1 5
#> gear 4 - 3 0.6667 0.1174 5.676 1.3751e-08 0.4365 0.8969 0 5
#> gear 5 - 3 1.0000 0.1566 6.386 1.7026e-10 0.6931 1.3069 1 5
#> gear 5 - 3 1.0000 0.1566 6.386 1.7026e-10 0.6931 1.3069 0 5
#>
#> Prediction type: response
#> Columns: rowid, type, term, contrast, estimate, std.error, statistic, p.value, conf.low, conf.high, predicted, predicted_hi, predicted_lo, am, gear
#>
m <- lm(mpg ~ hp + drat + factor(cyl) + factor(am), data = mtcars)
comparisons(m, variables = "hp", newdata = datagrid(FUN_factor = unique, FUN_numeric = median))
#>
#> Term Contrast Estimate Std. Error z Pr(>|z|) 2.5 % 97.5 %
#> hp +1 -0.04523 0.01489 -3.037 0.0023856 -0.07441 -0.01604
#> hp +1 -0.04523 0.01489 -3.037 0.0023856 -0.07441 -0.01604
#> hp +1 -0.04523 0.01489 -3.037 0.0023856 -0.07441 -0.01604
#> hp +1 -0.04523 0.01489 -3.037 0.0023856 -0.07441 -0.01604
#> hp +1 -0.04523 0.01489 -3.037 0.0023856 -0.07441 -0.01604
#> hp +1 -0.04523 0.01489 -3.037 0.0023856 -0.07441 -0.01604
#>
#> Prediction type: response
#> Columns: rowid, type, term, contrast, estimate, std.error, statistic, p.value, conf.low, conf.high, predicted, predicted_hi, predicted_lo, mpg, hp, drat, cyl, am, eps
#>
# Numeric contrasts
mod <- lm(mpg ~ hp, data = mtcars)
comparisons(mod, variables = list(hp = 1)) |> tidy()
#> # A tibble: 1 × 9
#> type term contrast estimate std.error statistic p.value conf.low conf.…¹
#> <chr> <chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 response hp mean(+1) -0.0682 0.0101 -6.74 1.56e-11 -0.0881 -0.0484
#> # … with abbreviated variable name ¹conf.high
comparisons(mod, variables = list(hp = 5)) |> tidy()
#> # A tibble: 1 × 9
#> type term contrast estimate std.error statistic p.value conf.low conf.…¹
#> <chr> <chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 response hp mean(+5) -0.341 0.0506 -6.74 1.56e-11 -0.440 -0.242
#> # … with abbreviated variable name ¹conf.high
comparisons(mod, variables = list(hp = c(90, 100))) |> tidy()
#> # A tibble: 1 × 9
#> type term contrast estim…¹ std.e…² stati…³ p.value conf.…⁴ conf.…⁵
#> <chr> <chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 response hp mean(100) - m… -0.682 0.101 -6.74 1.56e-11 -0.881 -0.484
#> # … with abbreviated variable names ¹estimate, ²std.error, ³statistic,
#> # ⁴conf.low, ⁵conf.high
comparisons(mod, variables = list(hp = "iqr")) |> tidy()
#> # A tibble: 1 × 9
#> type term contrast estim…¹ std.e…² stati…³ p.value conf.…⁴ conf.…⁵
#> <chr> <chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 response hp mean(Q3) - me… -5.70 0.845 -6.74 1.56e-11 -7.35 -4.04
#> # … with abbreviated variable names ¹estimate, ²std.error, ³statistic,
#> # ⁴conf.low, ⁵conf.high
comparisons(mod, variables = list(hp = "sd")) |> tidy()
#> # A tibble: 1 × 9
#> type term contrast estim…¹ std.e…² stati…³ p.value conf.…⁴ conf.…⁵
#> <chr> <chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 response hp mean(x + sd/2… -4.68 0.694 -6.74 1.56e-11 -6.04 -3.32
#> # … with abbreviated variable names ¹estimate, ²std.error, ³statistic,
#> # ⁴conf.low, ⁵conf.high
comparisons(mod, variables = list(hp = "minmax")) |> tidy()
#> # A tibble: 1 × 9
#> type term contrast estim…¹ std.e…² stati…³ p.value conf.…⁴ conf.…⁵
#> <chr> <chr> <chr> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 response hp mean(Max) - m… -19.3 2.86 -6.74 1.56e-11 -24.9 -13.7
#> # … with abbreviated variable names ¹estimate, ²std.error, ³statistic,
#> # ⁴conf.low, ⁵conf.high
# using a function to specify a custom difference in one regressor
dat <- mtcars
dat$new_hp <- 49 * (dat$hp - min(dat$hp)) / (max(dat$hp) - min(dat$hp)) + 1
modlog <- lm(mpg ~ log(new_hp) + factor(cyl), data = dat)
fdiff <- \(x) data.frame(x, x + 10)
avg_comparisons(modlog, variables = list(new_hp = fdiff))
#> Warning: These variables were not found: new_hp. Try specifying the `newdata`
#> argument explicitly.
#> Error: There is no valid predictor variable. Please change the `variables`
#> argument or supply a new data frame to the `newdata` argument.
# Adjusted Risk Ratio: see the contrasts vignette
mod <- glm(vs ~ mpg, data = mtcars, family = binomial)
avg_comparisons(mod, transform_pre = "lnratioavg", transform_post = exp)
#>
#> Term Contrast Estimate Pr(>|z|) 2.5 % 97.5 %
#> mpg mean(+1) 1.135 2.3808e-10 1.091 1.18
#>
#> Prediction type: response
#> Post-transformation: transform_post
#> Columns: type, term, contrast, estimate, p.value, conf.low, conf.high, predicted, predicted_hi, predicted_lo
#>
# Adjusted Risk Ratio: Manual specification of the `transform_pre`
avg_comparisons(
mod,
transform_pre = function(hi, lo) log(mean(hi) / mean(lo)),
transform_post = exp)
#>
#> Term Contrast Estimate Pr(>|z|) 2.5 % 97.5 %
#> mpg +1 1.135 2.3808e-10 1.091 1.18
#>
#> Prediction type: response
#> Pre-transformation: transform_pre
#> Post-transformation: transform_post
#> Columns: type, term, contrast, estimate, p.value, conf.low, conf.high, predicted, predicted_hi, predicted_lo
#>
# cross contrasts
mod <- lm(mpg ~ factor(cyl) * factor(gear) + hp, data = mtcars)
avg_comparisons(mod, variables = c("cyl", "gear"), cross = TRUE)
#> Warning: Model matrix is rank deficient. Some variance-covariance parameters are
#> missing.
#>
#> C: cyl C: gear Estimate Std. Error z Pr(>|z|) 2.5 % 97.5 %
#> 6 - 4 4 - 3 -0.6306 3.405 -0.1852 0.85307 -7.303 6.042
#> 6 - 4 5 - 3 2.6778 4.617 0.5800 0.56193 -6.372 11.727
#> 8 - 4 4 - 3 3.3475 6.427 0.5208 0.60249 -9.250 15.945
#> 8 - 4 5 - 3 5.5250 5.868 0.9416 0.34642 -5.976 17.026
#>
#> Prediction type: response
#> Columns: type, term, contrast_cyl, contrast_gear, estimate, std.error, statistic, p.value, conf.low, conf.high
#>
# variable-specific contrasts
avg_comparisons(mod, variables = list(gear = "sequential", hp = 10))
#> Warning: Model matrix is rank deficient. Some variance-covariance parameters are
#> missing.
#>
#> Term Contrast Estimate Std. Error z Pr(>|z|) 2.5 % 97.5 %
#> gear 4 - 3 3.4095 2.587 1.3177 0.187591 -1.662 8.4806
#> gear 5 - 4 2.6277 2.747 0.9566 0.338744 -2.756 8.0113
#> hp +10 -0.5741 0.225 -2.5518 0.010718 -1.015 -0.1331
#>
#> Prediction type: response
#> Columns: type, term, contrast, estimate, std.error, statistic, p.value, conf.low, conf.high
#>
# hypothesis test: is the `hp` marginal effect at the mean equal to the `drat` marginal effect
mod <- lm(mpg ~ wt + drat, data = mtcars)
comparisons(
mod,
newdata = "mean",
hypothesis = "wt = drat")
#>
#> Term Estimate Std. Error z Pr(>|z|) 2.5 % 97.5 %
#> wt=drat -6.225 1.052 -5.919 3.2398e-09 -8.287 -4.164
#>
#> Prediction type: response
#> Columns: type, term, estimate, std.error, statistic, p.value, conf.low, conf.high
#>
# same hypothesis test using row indices
comparisons(
mod,
newdata = "mean",
hypothesis = "b1 - b2 = 0")
#>
#> Term Estimate Std. Error z Pr(>|z|) 2.5 % 97.5 %
#> b1-b2=0 -6.225 1.052 -5.919 3.2398e-09 -8.287 -4.164
#>
#> Prediction type: response
#> Columns: type, term, estimate, std.error, statistic, p.value, conf.low, conf.high
#>
# same hypothesis test using numeric vector of weights
comparisons(
mod,
newdata = "mean",
hypothesis = c(1, -1))
#>
#> Term Estimate Std. Error z Pr(>|z|) 2.5 % 97.5 %
#> custom -6.225 1.052 -5.919 3.2398e-09 -8.287 -4.164
#>
#> Prediction type: response
#> Columns: type, term, estimate, std.error, statistic, p.value, conf.low, conf.high
#>
# two custom contrasts using a matrix of weights
lc <- matrix(c(
1, -1,
2, 3),
ncol = 2)
comparisons(
mod,
newdata = "mean",
hypothesis = lc)
#>
#> Term Estimate Std. Error z Pr(>|z|) 2.5 % 97.5 %
#> custom -6.225 1.052 -5.9190 3.2398e-09 -8.287 -4.164
#> custom -5.238 5.624 -0.9315 0.35162 -16.261 5.784
#>
#> Prediction type: response
#> Columns: type, term, estimate, std.error, statistic, p.value, conf.low, conf.high
#>
# `by` argument
mod <- lm(mpg ~ hp * am * vs, data = mtcars)
comparisons(mod, by = TRUE)
#>
#> Term Contrast Estimate Std. Error z Pr(>|z|) 2.5 % 97.5 %
#> hp +1 -0.06878 0.0182 -3.7797 0.00015703 -0.1044 -0.03311
#> am 1 - 0 4.69801 1.0601 4.4316 9.3531e-06 2.6202 6.77580
#> vs 1 - 0 -0.29433 2.3379 -0.1259 0.89981407 -4.8766 4.28789
#>
#> Prediction type: response
#> Columns: type, term, contrast, estimate, std.error, statistic, p.value, conf.low, conf.high
#>
mod <- lm(mpg ~ hp * am * vs, data = mtcars)
avg_comparisons(mod, variables = "hp", by = c("vs", "am"))
#>
#> Term Contrast vs am Estimate Std. Error z Pr(>|z|) 2.5 % 97.5 %
#> hp mean(+1) 0 1 -0.03685 0.01240 -2.971 0.0029676 -0.06116 -0.012540
#> hp mean(+1) 1 1 -0.11115 0.04634 -2.399 0.0164473 -0.20197 -0.020336
#> hp mean(+1) 1 0 -0.09941 0.05345 -1.860 0.0628876 -0.20417 0.005344
#> hp mean(+1) 0 0 -0.04215 0.02477 -1.702 0.0887904 -0.09070 0.006394
#>
#> Prediction type: response
#> Columns: type, term, contrast, vs, am, estimate, std.error, statistic, p.value, conf.low, conf.high, predicted, predicted_hi, predicted_lo
#>
library(nnet)
mod <- multinom(factor(gear) ~ mpg + am * vs, data = mtcars, trace = FALSE)
by <- data.frame(
group = c("3", "4", "5"),
by = c("3,4", "3,4", "5"))
comparisons(mod, type = "probs", by = by)
#>
#> Term By Estimate Std. Error z Pr(>|z|) 2.5 % 97.5 %
#> mpg 3,4 0.0004635 0.005796 0.07997 0.9362637 -0.01090 0.01182
#> am 3,4 -0.2227925 0.079558 -2.80038 0.0051042 -0.37872 -0.06686
#> vs 3,4 0.1021021 0.073236 1.39416 0.1632708 -0.04144 0.24564
#> mpg 5 -0.0009270 0.011592 -0.07997 0.9362637 -0.02365 0.02179
#> am 5 0.4455851 0.159116 2.80038 0.0051042 0.13372 0.75745
#> vs 5 -0.2042041 0.146472 -1.39416 0.1632708 -0.49128 0.08287
#>
#> Prediction type: probs
#> Columns: type, term, by, estimate, std.error, statistic, p.value, conf.low, conf.high
#>