uwLeveledBoxPlot: Box plots with Two factor variables
In jbirstler/biostatrpts: Functions for UW-Biostatistics Report Creation

Description Usage Arguments Details Author(s) See Also Examples

This function will produce a graphic and statistical testing for a numeric variable when stratified by two factor variables. Generally one of the factor variables is considered the treatment variable. Also produced is code for a corresponding LaTeX table.

uwLeveledBoxPlot(
  allData,
  trxName,
  metricName,
  lowfactName,
  hifactName = NULL,
  delta = FALSE,
  deltaPct = FALSE,
  baseLevel = NULL,
  idName = NULL,
  LatexFileName = NULL,
  LatexCaption = NULL,
  lowfacttableName = "",
  hifacttableName = "",
  showTab = FALSE,
  pTitle = NULL,
  yLab = NULL,
  yLim = NULL,
  plotMean = FALSE,
  pOutliers = FALSE,
  lWhisker = 0.05,
  hWhisker = 0.95,
  lHinge = 0.25,
  hHinge = 0.75,
  boxWex = 0.75,
  LabelCex = 0.7,
  Legend = FALSE,
  LegendLoc = "topright",
  LegendCex = 0.7,
  printPVals = TRUE,
  PVinTableOnly = FALSE,
  PValCex = 0.6,
  pTest = c("wilcox", "t.test"),
  abbrevN = 1,
  pairwise = TRUE,
  pAdjust = NULL,
  trxControl = NULL,
  pExclude = NULL,
  pInclude = list(list(NULL, NULL)),
  Ropen = FALSE,
  numDec = 2,
  ...
)

`allData`	Data frame in R that contains all of the variables
`trxName`	String. Name of the treatment variable in allData
`metricName`	String. Name of the numeric variable in allData that will be represented by boxes
`lowfactName`	String. Name of the low-level factor variable in allData that will be used to stratify. Make the factor an ordered factor if order matters in the reporting.
`hifactName`	String. Name of the upper-level factor variable in allData that will be used to stratify first, and then stratify on lowfactName.
`delta`	Logical. TRUE looks at differences from baseLevel and other levels of hifactName or lowfactName is no hifactName is given
`deltaPct`	Logical. TRUE gives percentage change from baseLevel. delta must also be TRUE
`baseLevel`	String. Level in hifactName or lowfactName if no hifactName is given that is considered the baseline level
`idName`	String. Variable in allData that gives patient identifiers. Only necessary if delta=TRUE
`LatexFileName`	String. Giving the file path and name, ending in .tex, where the LaTeX code for the table should be saved. If NULL, table is printed in the R session but not saved.
`LatexCaption`	uwLatex, Caption for Latex table
`lowfacttableName`	String to be given to the lowfactName in the LaTeX table
`hifacttableName`	String to be given to the hifactName in the LaTeX table
`showTab`	Logical. TRUE -> Displays the compiled LaTeX table using dvix
`pTitle`	String. Title of the Plot and caption of the LaTeX table
`yLab`	String. Label for the y-axis.
`yLim`	Limits of y-axix. In the form: c(start,finish)
`plotMean`	Logical. TRUE -> A line is plotted in the box plots showing the mean value. A point is always plotted showing the median value.
`pOutliers`	Logical. TRUE -> Show points outside of whiskers
`lWhisker`	Lower whisker level. Default is .05 for 5th percentile
`hWhisker`	Upper whisker level. Default is .95 for 95th percentile
`lHinge`	Lower level of box. Default is .25 for 25th percentile
`hHinge`	Upper level of box. Default is .75 for 75th percentile
`boxWex`	Number for width of boxes
`LabelCex`	Numeric. Gives magnification level of the labels in the x-axis. Less than 1 shrinks them, greater than 1 expands them.
`Legend`	Logical. TRUE puts a legend of the trxName levels in the plot and removes the labeling in the x-axis for trxName levels
`LegendLoc`	Location of the legend if Legend=TRUE, see help(legend)
`LegendCex`	Numeric. Gives magnification level of the legend
`printPVals`	Logical. TRUE -> Prints p-values on the top of the graph and adds them to the LaTeX table
`PVinTableOnly`	Logical. TRUE Suppresses the printing of p-values in the top margin of the graph, but they are still printed in the LaTeX table. Useful when p-values are hard to read in the margin
`PValCex`	Numeric. Gives magnification level of the p-values in the top margin
`pTest`	printPVals must be TRUE. Chose type of comparison test
`abbrevN`	Whole number indicating how many letters should abbreviation of the treatments levels should be.
`pairwise`	(logical) TRUE pairwise comparisons should be made between treatment levels. FALSE multi-level test occurs between all levels of the treatment
`pAdjust`	NULL for none, "h" for "holm", or "b" for "bonferroni". P-value adjustment methods. See uwPVals for more information.
`trxControl`	Must be defined if pairwise=TRUE. A treatment level is indicated here as the control level, two-way comparisons with the control and non-controls will then be made.
`pExclude`	only applies if pairwise=TRUE. This should be a vector of non-control treatment levels that will be looked at one at a time and that non-control's contrast will not be reported
`pInclude`	List specifying contrasts want to be examined. This is most useful when wanting to combine treatment levels together. See uwPVals for more details
`Ropen`	Logical. TRUE collapses data across all treatment arms.
`numDec`	Numeric from 0 to 20 indicating number of decimals to be reported in the tables
`...`	Any other argument that can be passed to uwLatex other than mat, file, or col.just which are already set

Special Functions used in uwLeveledBoxPlot: uwPVals and uwLatex

Plotting points in the boxes represent the median value. A line would represent the mean if plotMean=TRUE.

University of Wisconsin-Madison Biostatistics and Medical Informatics Department, Scott Hetzel M.S. and Frontier Science and Technology Research Foundation, Patrick Lenon

uwBoxPlot(), uwLatex()

trt1 <- rep(rep(c("A", "B", "C"), 120), 3)
fctr1 <- as.factor(rep(c(rep("Female", 180), rep("Male", 180)), 3))
mtrc1 <- rnorm(360 * 3, 10, 4)
upfct <- c(rep("Baseline", 360), rep("Week 12", 360), rep("Week 24", 360))
ids <- as.character(c(1:360, 1:360, 1:360))
dat1 <- data.frame(trt1, fctr1, mtrc1, upfct, ids)

lyt <- layout(matrix(c(1, 1, 2, 2, 3, 3), nrow = 3, byrow = TRUE))

uwLeveledBoxPlot(
  allData = dat1,
  trxName = "trt1",
  metricName = "mtrc1",
  lowfactName = "fctr1",
  hifactName = "upfct",
  delta = FALSE,
  baseLevel = "Baseline",
  idName = "ids",
  LatexFileName = NULL,
  showTab = FALSE,
  Legend = FALSE,
  pTitle = "Example of uwLeveledBoxPlot",
  yLab = "Random Normal",
  plotMean = TRUE, pOutliers = FALSE,
  lWhisker = .05, hWhisker = .95,
  lHinge = .25, hHinge = .75,
  boxWex = 0.75,
  printPVals = TRUE,
  PVinTableOnly = FALSE,
  PValCex = 0.5,
  Ropen = FALSE,
  pTest = "t",
  abbrevN = 1,
  pairwise = TRUE,
  trxControl = NULL,
  pExclude = NULL,
  yLim = NULL
)

uwLeveledBoxPlot(
  allData = dat1,
  trxName = "trt1",
  metricName = "mtrc1",
  lowfactName = "fctr1",
  hifactName = "upfct",
  delta = TRUE,
  deltaPct = FALSE,
  baseLevel = "Baseline",
  idName = "ids",
  LatexFileName = NULL,
  showTab = FALSE,
  Legend = TRUE,
  LegendLoc = "topleft",
  pTitle = "Example of uwLeveledBoxPlot with delta=TRUE",
  yLab = "Random Normal",
  plotMean = FALSE, pOutliers = FALSE,
  lWhisker = .05, hWhisker = .95,
  lHinge = .25, hHinge = .75,
  boxWex = 0.75,
  printPVals = TRUE,
  PVinTableOnly = FALSE,
  PValCex = 0.5,
  Ropen = FALSE,
  pTest = "w",
  abbrevN = 1,
  pairwise = TRUE,
  trxControl = NULL,
  pExclude = NULL,
  pInclude = list(list(c("A", "B"), "C"), list("A", c("B", "C")), list("A", "C")),
  yLim = NULL
)

uwLeveledBoxPlot(
  allData = dat1,
  trxName = "trt1",
  metricName = "mtrc1",
  lowfactName = "fctr1",
  hifactName = "upfct",
  delta = TRUE,
  deltaPct = TRUE,
  baseLevel = "Baseline",
  idName = "ids",
  LatexFileName = NULL,
  showTab = FALSE,
  Legend = TRUE,
  LegendLoc = "topleft",
  pTitle = "Example of uwLeveledBoxPlot with deltaPct=TRUE",
  yLab = "Random Normal",
  plotMean = FALSE, pOutliers = FALSE,
  lWhisker = .05, hWhisker = .95,
  lHinge = .25, hHinge = .75,
  boxWex = 0.75,
  printPVals = TRUE,
  PVinTableOnly = FALSE,
  PValCex = 0.5,
  Ropen = FALSE,
  pTest = "w",
  abbrevN = 1,
  pairwise = TRUE,
  trxControl = NULL,
  pExclude = NULL,
  pInclude = list(list(c("A", "B"), "C"), list("A", c("B", "C")), list("A", "C")),
  yLim = NULL
)