ggbootUV: Univariate bootstrap plot
In MiguelRodo/ggboot: Bootstrap univariate and multivariate plots

ggbootUV

R Documentation

Univariate bootstrap plot

Description

Plots a sample statistic, such as the mean or median, and associated bootstrapped 95% confidence intervals for specified subsets of observations. Facilitates hypothesis testing by allowing a) application of the Benjamini-Hochberg procedure to control the alse discovery rate and b) specification of a minimum effect size.

Usage

ggbootUV(data, resp, xAxis, col = NULL, facet = NULL, nullBFactor = 10,
  diff = NULL, calcStat = "mean", B = 10, seB = 200,
  pMethod = "percT", fdr = 0.01, altSide, nSide = 1, ciMethod = "bca",
  alpha = 0.05, minEff = 0, hLineVec = NULL, nullValue = 0,
  xLab = NULL, yLab = "Response", xAxisLabVec = NULL, rotXText = FALSE,
  remXAxisMarks = FALSE, pointSize = 2, errBarSize = 1,
  errBarSizeRatio = 2, errBarAlpha = 0.7, errBarAlphaRatio = 2,
  eqErrBarWidth = TRUE, errBarLineType = "dotted", colLabName = NULL,
  colourVec = c("darkorchid1", "springgreen2", "maroon1", "dodgerblue", "red",
  "yellow3", "cyan2", "orange2"), colLabVec = NULL, facetLabVec = NULL,
  facetScale = "free", nCol = 2, plotTblName = NULL,
  fontScaleFactor = 1, lineScaleFactor = 1, hLineFactor = 1,
  bootT = TRUE, trim = 0.2, bootSECorr = FALSE)

Arguments

`data`	dataframe. Has one row per observation and the column names specified below.
`resp`	unquoted expression. Response variable name.
`xAxis`	unquoted expression. X-axis variable name.
`col`	unquoted expression. Colour variable name.
`facet`	unquoted expression. Facet variable name.
`diff`	unquoted expression. Difference variable name. Must be a binary variable, preferrably ('0','1') as then it is clear that the difference is the amount by which the observations labelled '1' are greater than the observations labelled '0'.
`calcStat`	function olr character. Bootstrap statistic function. Functions other than mean or median used at own risk.
`B`	integer. Number of bootstrap samples per subgroup.
`pMethod`	"perc" or "percT". Method to use for calculating p-values. Both uses the bootstrap percentile method, but "perc" does this on the mean or median, whilst "percT" does this on the t-statistic based on the mean or median.
`fdr`	numeric. False discovery rate.
`nSide`	1 or 2. Number of sides of the hypothesis tests.
`ciMethod`	"perc" or "bca". method to use to calculate confidence intervals. Both use bootstrap samples, but "perc" uses the percentile method and "bca" the bias-corrected and adjusted method.
`alpha`	numeric. 1-`alpha` is the coverage of the confidence intervals.
`minEff`	numeric. Minimum effect size. Default is 0.
`hLineVec`	numeric vector. Heights of dashed horizontal lines to plot. Useful for indicating minimum effect size.
`nullValue`	numeric. Null hypothesis value for statistic.
`xLab, yLab`	characeter. X- and y-axis titles. Optional.
`xAxisLabVec`	named character vector. Labels for x-axis variable. Optional.
`rotXText`	logical. If TRUE, then the x-axis text is rotated 90 degrees clockwise.
`remXAxisMarks`	logical. If TRUE, then the x-axis ticks and text are removed.
`pointSize`	numeric. Size of plot points
`errBarSize`	numeric. Size of error bar
`errBarSizeRatio`	numeric. How many times larger the size is of the statistically significant error bars than the non-statistically significant bars are.
`errBarAlpha`	numeric. Error bar alpha
`errBarAlphaRatio`	numeric. How many times greater the alpha is of the statistically significant error bars than the non-statistically significant bars are.
`eqErrBarWidth`	logical. If `TRUE`, then the error bars are made the same width in each facet. Otherwise the error bars automatically span the entire facet, however many x-axis entries there are.
`colLabName`	character. Name for colour legend. Optional.
`colourVec`	character vector. Colours for colour scale. Optional. Label with factor levels to match levels to colours manually.
`colLabVec`	named character vector. Labels for colour variable. Optional.
`facetScale`	'free', 'free_x' or 'free_y'. Facet scale.
`nCol`	numeric. Number of columns to use.
`plotTblName`	character. If not NULL, then the dataframe used to plot is saved to global environment with name `plotTblName`.
`fontScaleFactor`	numeric. Factor by which to scale the text size from the default.
`lineScaleFactor`	numeric. Factor by which to scale the line size from the default.
`hLineFactor`	numeric. Factor by which to scale the horizontal line size from the default.
`altSde`	'high', 'low' or 'both'. Character specifying the alternative hypothesis relative to the null hypothesis. Note that you must still specify nSide, even if altSide is specified.
`facetlabeVec`	named character vector. Labels for facet fariable. Optional.
`sigIndType`	character vector. Must contain at least one of `"lineType"`, `"alpha"` and `"size"`. The statistically significant and non-statistically significant error bars will then differ by this aesthetic.