ucr.base.tab: Create a baseline table
In renlund/ucR: Ucr Stat Functions

ucr.base.tab

R Documentation

Create a baseline table

Description

Creates a baseline table from a data frame. The data may be partitioned in two or more groups. A column for the "combined" group (all data) can also be included.

Usage

ucr.base.tab(
  data,
  group.name = NULL,
  combined.name = "Combined",
  x.names = setdiff(names(data), group.name),
  num.format = "median",
  median.format = "iqr",
  mean.format = "par",
  factor.format = "count.perc",
  perc.method = "group",
  print.perc = T,
  print.perc.space = F,
  omit.perc.decimal = F,
  use.texttt = F,
  omit.ref.level = F,
  separate.factor.row = F,
  show.missing = "none",
  digits = 1,
  spec.digits = NULL,
  include.combined = T,
  include.n = T,
  include.p = T,
  test.x.names = x.names,
  num.test = "nonparam",
  factor.test = "fisher",
  min.p = "0.001"
)

Arguments

`data`	The data frame.
`group.name`	Name of the variable in 'data' that defines the groups, or NULL (default) if there are no groups. The group variable must be a factor without missing..
`combined.name`	Column heading for the combined group. Defaults to "Combined".
`x.names`	Names and order of the variables in 'data' that shall be included in the table. By default, all variables except the group variable, if any, are included.
`num.format`	How to present numerical variables. Allowed values: "median": Median with IQR or range. "mean": Mean with standard deviation. "both": Both. This gives "median stuff (mean stuff)".
`median.format`	What measure of dispersion to give together with medians. Allowed values: "iqr": Interquartile range (Q1 and Q3). "range": Range (min and max). A numerical value strictly between 0 and 50. Example: If median.format = 10, then the 10th and 90th percentiles will be reported.
`mean.format`	How to present standard deviations. Allowed values: "par": Mean (SD). "pm": Mean plus/minus SD.
`factor.format`	How to present factor variables. Allowed values: count.perc: Count (percentage). perc.count: Percentage (count).
`perc.method`	How to compute percentages for factor variables. Allowed values: "group": Computes n(group, level) / n(group). "level": Computes n(group, level) / n(level). For the combined group, this would always give 100%. Hence, percentages are omitted in this case. "total": Percentage n(group, level) / n.
`print.perc`	TRUE if a percent sign ('%') is to be printed for percentages.
`print.perc.space`	TRUE if a space is to be printed between the number and the percent sign, i.e. "12.3 %" rather than "12.3%". N/A if percent sign is disabled.
`omit.perc.decimal`	TRUE if the decimal is to be omitted for percentages above 1%, e.g. 13% rather than 13.3%.
`use.texttt`	logical; use latex texttt
`omit.ref.level`	TRUE if the reference (first) level of dichotomous factor variables should be omitted from the table.
`separate.factor.row`	TRUE if a separate row is added (first) for each factor, containing the variable name, number of inidividuals/missing etc.
`show.missing`	How should missing values, if any, per group be shown? "none": Don't show missing values. "in.row": Show the number of missing values within square brackets after the usual data. "sep.row": Show the number of missing values is a separate row.
`digits`	An integer, giving the number of decimals for numeric variables, unless otherwise specified through the parameter 'spec.digits'.
`spec.digits`	A list, whose names form a subset of 'x.names', and whose values are the number of decimals for those (numeric) variables. The value NULL causes 'digits' to be used throughout.
`include.combined`	TRUE if the combined group should be included in a separate column.
`include.n`	TRUE if number of valid (non-NA) observations per variable should be included in a separate column.
`include.p`	TRUE if P-values are to be included in the table. They refer to univariate tests for identical distribution in the different groups. P-values can only be included if there are groups.
`test.x.names`	Names of variables to perform tests for.
`num.test`	Which kind of test to use for numerical variables. Allowed values: "nonparam": A non-parametrical test: Wilcoxon (Mann-Whitney) test for two groups. Kruskal-Wallis test for three or more groups. "param": A parametrical test: t-test for two groups. ANOVA for three or more groups. "nonparam.trend": The non-parametric Jonckheere-Terpstra trend test, treating the group variable as ordinal. Only for 3+ groups. "param.trend": A parametric trend test (likelihood ratio test for simple linear regression), converting the group variable to linear scores. Only for 3+ groups.
`factor.test`	Which test to use for factor variables. Allowed values: "fisher": Fisher's exact test. "pearson": Pearson's chi2 test. "trend": Linear-by-linear association test, converting both the group variable and the baseline variable (possibly dichotomous) to linear scores. Only for 3+ groups.
`min.p`	Smallest P-value to be displayed. If smaller, then a text like "< 0.001" is given. NOTE: This value must be a string, e.g. "0.001" or "1e-3".