dotwhisker (version 0.2.0.5)

small_multiple: Generate a 'Small Multiple' Plot of Regression Results

Description

small_multiple is a function for plotting regression results of multiple models as a 'small multiple' plot

Usage

small_multiple(x, dodge_size = 0.06, alpha = 0.05)

Arguments

x

Either a tidy data.frame including results from multiple models (see 'Details') or a list of model objects that can be tidied with tidy

dodge_size

A number (typically between 0 and 0.3; the default is .06) indicating how much horizontal separation should appear between different submodels' coefficients when multiple submodels are graphed in a single plot. Lower values tend to look better when the number of models is small, while a higher value may be helpful when many submodels appear on the same plot.

alpha

A number setting the criterion of the confidence intervals. The default value is .05, corresponding to 95-percent confidence intervals.

Value

The function returns a ggplot object.

Details

Kastellec and Leoni (2007) small_multiple takes a tidy data.frame of regression results or a list of model objects and generates a dot-and-whisker plot of the results of a single variable across the multiple models.

Tidy data.frames to be plotted should include the variables term (names of predictors), estimate (corresponding estimates of coefficients or other quantities of interest), std.error (corresponding standard errors), and model (identifying the corresponding model). In place of std.error one may substitute lb (the lower bounds of the confidence intervals of each estimate) and ub (the corresponding upper bounds).

Alternately, small_multiple accepts as input a list of model objects that can be tidied by tidy.

Optionally, more than one set of results can be clustered to facilitate comparison within each model; one example of when this may be desireable is to compare results across samples. In that case, the data.frame should also include a variable submodel identifying the submodel of the results.

Examples

library(broom)
library(dplyr)

# Generate a tidy data.frame of regression results from six models

m <- list()
ordered_vars <- c("wt", "cyl", "disp", "hp", "gear", "am")
m[[1]] <- lm(mpg ~ wt, data = mtcars)
m123456_df <- m[[1]] %>% tidy %>% by_2sd(mtcars) %>%
  mutate(model = "Model 1")

for (i in 2:6) {
 m[[i]] <- update(m[[i-1]], paste(". ~ . +", ordered_vars[i]))
 m123456_df <- rbind(m123456_df, m[[i]] %>% tidy %>% by_2sd(mtcars) %>%
   mutate(model = paste("Model", i)))
}

# Generate a 'small multiple' plot
small_multiple(m123456_df)


## Using submodels to compare results across different samples
# Generate a tidy data.frame of regression results from five models on
# the mtcars data subset by transmission type (am)
ordered_vars <- c("wt", "cyl", "disp", "hp", "gear")
mod <- "mpg ~ wt"
by_trans <- mtcars %>% group_by(am) %>%  # group data by transmission
  do(tidy(lm(mod, data = .))) %>%        # run model on each group
  rename(submodel = am) %>%              # make submodel variable
  mutate(model = "Model 1")              # make model variable

for (i in 2:5) {
  mod <- paste(mod, "+", ordered_vars[i])
  by_trans <- rbind(by_trans, mtcars %>% group_by(am) %>%
                   do(tidy(lm(mod, data = .))) %>%
                   rename(submodel = am) %>%
                   mutate(model = paste("Model", i)))
}

small_multiple(by_trans) +
theme_bw() + ylab("Coefficient Estimate") +
    geom_hline(yintercept = 0, colour = "grey60", linetype = 2) +
    theme(axis.text.x  = element_text(angle = 45, hjust = 1),
          legend.position=c(0, 0), legend.justification=c(0, 0),
          legend.title = element_text(size=9),
          legend.background = element_rect(color="gray90"),
          legend.margin = unit(-3, "pt"),
          legend.key.size = unit(10, "pt")) +
    scale_colour_hue(name = "Transmission",
    breaks = c(0, 1),
    labels = c("Automatic", "Manual"))