dots_bars {dvmisc}

Plot Points +/- Error Bars

Description

Creates plot showing user-specified points (e.g. means, medians, regression coefficients) along with user-specified error bars (e.g. standard deviations, min/max, 95% confidence intervals).

Usage

dots_bars(y = NULL, bars = NULL, bars.lower = y - bars,
  bars.upper = y + bars, truth = NULL, group.labels = NULL,
  group.dividers = TRUE, subgroup.spacing = 1,
  subgroup.labels = NULL, subgroup.pch = NULL, subgroup.col = NULL,
  points.list = NULL, arrows.list = NULL, xaxis.list = NULL,
  yaxis.list = xaxis.list, abline.dividers.list = NULL,
  abline.truth.list = NULL, legend.list = NULL, ...)

Arguments

y

Numeric vector of y-values for different groups, or numeric matrix where each column contains y-values for clustered subgroups within a group.

bars

Numeric vector or matrix (matching whichever type y is) specifying the length of the error bar for each group/subgroup (i.e. distance from point to one end of error bar).

bars.lower

Numeric vector or matrix (matching whichever type y is) specifying the position of the lower end of the error bar for each group/subgroup.

bars.upper

Numeric vector or matrix (matching whichever type y is) specifying the position of the upper end of the error bar for each group/subgroup.

truth

Numeric value specifying true value of parameter being estimated. If specified, a horizontal reference line is added to the plot.

group.labels

Character vector of labels for the groups.

group.dividers

Logical value for whether to add vertical lines distinguishing the groups.

subgroup.spacing

Numeric value controlling the amount of spacing between subgroups, with values > 1 corresponding to more spacing.

subgroup.labels

Character vector giving labels for the subgroups.

subgroup.pch

Plotting symbol for different subgroups within each group.

subgroup.col

Plotting color for different subgroups within each group.

points.list

Optional list of inputs to pass to points.

arrows.list

Optional list of inputs to pass to arrows.

xaxis.list

Optional list of inputs to pass to axis for x-axis.

yaxis.list

Optional list of inputs to pass to axis for y-axis.

abline.dividers.list

Optional list of inputs to pass to abline for group dividers. Only used if group.dividers = TRUE.

abline.truth.list

Optional list of inputs to pass to abline for horizontal line at true value of parameter. Only used if truth is specified.

legend.list

Optional list of inputs to pass to legend.

...

Additional arguments to pass to plot function.

Value

Plot showing points +/- error bars across groups/subgroups.

Examples

# Generate 100 values from normal distributions with different means, and 
# graph mean +/- standard deviation across groups 
dat <- cbind(rnorm(100, 2), rnorm(100, 2.5), rnorm(100, 1.75))
means <- apply(dat, 2, mean)
sds <- apply(dat, 2, sd)
fig1 <- dots_bars(y = means, bars = sds, main = "Mean +/- SD by Group",
                  ylab = "Mean +/- SD")
                  
# Simulate BMI values for males and females in 3 different age groups, and 
# graph mean +/- 95% CI
sex <- as.factor(c(rep("Male", 300), rep("Female", 300)))
age <- as.factor(rep(c("Young", "Middle", "Old"), 2))
bmi <- c(rnorm(100, 25, 4), rnorm(100, 26, 4.25), rnorm(100, 27, 4.5),
         rnorm(100, 26.5, 4.5), rnorm(100, 27.25, 4.75), rnorm(100, 28, 5))
dat <- data.frame(sex = sex, age = age, bmi = bmi)
means <- tapply(dat$bmi, dat[, c("sex", "age")], mean)
ci.lower <- tapply(dat$bmi, dat[, c("sex", "age")],
                   function(x) t.test(x)$conf.int[1])
ci.upper <- tapply(dat$bmi, dat[, c("sex", "age")],
                   function(x) t.test(x)$conf.int[2])
fig2 <- dots_bars(y = means, bars.lower = ci.lower, bars.upper = ci.upper,
                  main = "BMI by Sex and Age",
                  ylab = "BMI (mean +/- CI)",
                  xlab = "Age group")