report.quanti: Descriptive "Quantitative" statistics (mean, SD, median...)...

In ClinReport: Statistical Reporting for Scientific Publications, Epidemiological Studies and Clinical Trials

Description

report.quanti Returns quantitative descriptive statistics such as mean, median, standard deviation etc...

For more examples see the website: ClinReport website

Usage

report.quanti(data, y, x1 = NULL, x2 = NULL, y.label = y,
  round = 2, total = F, scientific = F, digits = NULL,
  at.row = NULL, subjid = NULL, geomean = F, add.mad = F,
  default.stat = T, func.stat = NULL, stat.name = "Statistics",
  func.stat.name = "", drop.x1 = NULL, drop.x2 = NULL)

Arguments

data	Data.frame object
y	Character indicating a numerical vector in the data frame passed to data argument
x1	Character indicating a factor in the data (levels will be displayed in columns)
x2	Character indicating a factor in the data (levels will be displayed in lines)
y.label	Character Indicates the label for y parameter to be displayed in the title of the table
round	Numeric to indicate how to round statistics
total	Logical to indicate if a "Total" column should be added
scientific	Logical Indicates if statistics should be displayed in scientific notations or not
digits	Numeric (used if scientific=TRUE) to indicate how many digits to use in scientific notation
at.row	Character Used to space the results (see examples)
subjid	Character Indicates the column in which there is the subject Id to add the number of subjects in the column header if x1 and x2 are not null.
geomean	Logical If yes geometric mean is calculated instead of arithmetic mean: exp(mean(log(x),na.rm=TRUE)) for x>0
add.mad	Logical If yes the Median Absolute Deviance is added to the median statistics (see function mad)
default.stat	Logical (default to TRUE). If FALSE you can specify your own example
func.stat	Function. If specified then default.stat=FALSE and only the specified statistic is reported
stat.name	Character. Indicates the name of the variable that report the statistics Default = "Statistics"
func.stat.name	Character. Used only if default.stat=FALSE. Indicates the name of specific statistic you want to report
drop.x1	Character. Indicates one or several levels of the x1 factor that you want to drop in the result
drop.x2	Character. Indicates one or several levels of the x2 factor that you want to drop in the result

Details

This function computes and reports quantitative statistics on y. And can gives the statistics by level of two factors (x1 in columns and/or x2 in rows). See the example to show the results. If total=TRUE, the last column is the statistics performed overall levels of x1 for each levels of x2. Quantiles are calculated using type 3 (SAS presumed definition) algorithms, but even though, some differences between SAS and R can appear on quantile values.

"geomean" compute the geometric mean defined as exp(mean(log(y))). The values below or equal 0 are removed and a message is printed to indicate how many values were deleted to calculate the geometric mean.

N returns the number of observations (including NA values)

stat.name is automatically transformed using make.names function.

Value

A desc object.

See Also

report.quali report.doc desc

Examples

 
data(datafake)

# Quantitative statistics with no factor

report.quanti(data=datafake,y="y_numeric",total=TRUE,y.label="Awesome results")

#' # Quantitative statistics with no factor with geometric mean (option geomean=TRUE)

report.quanti(data=datafake,y="y_numeric",y.label="Awesome results",geomean=TRUE)

# Quantitative statistics with one factor

report.quanti(data=datafake,y="y_numeric",x1="GROUP")

# One factor with total column

report.quanti(data=datafake,y="y_numeric",x1="GROUP",total=TRUE)

# Quantitative statistics with two factors

report.quanti(data=datafake,y="y_numeric",x1="GROUP",x2="TIMEPOINT")

# Quantitative statistics with two factors and a total column

report.quanti(data=datafake,y="y_numeric",x1="GROUP",x2="TIMEPOINT",total=TRUE)

# Add median absolute deviance to the median statistics

report.quanti(data=datafake,y="y_numeric",x1="GROUP",x2="TIMEPOINT",total=TRUE,add.mad=TRUE)

# Quantitative statistics with spacing rows (option at.row)

report.quanti(data=datafake,y="y_numeric",x1="GROUP",
x2="TIMEPOINT",total=TRUE,at.row="TIMEPOINT")

# Add number of subjects in headers (option subjid)

tab=report.quanti(data=datafake,y="y_numeric",x1="GROUP",
x2="TIMEPOINT",total=TRUE,at.row="TIMEPOINT",subjid="SUBJID")

# Print tab output
tab


#Getting a specific statistic and not the default ones

mystat=function(x) quantile(x,0.99,na.rm=TRUE)

tab=report.quanti(data=datafake,y="y_numeric",x1="GROUP",
x2="TIMEPOINT",total=TRUE,subjid="SUBJID",
func.stat=mystat,func.stat.name="99% quantile")
tab

mystat2=function(x) mean(x,na.rm=TRUE)/sd(x,na.rm=TRUE)

tab=report.quanti(data=datafake,y="y_numeric",x1="GROUP",
total=TRUE,subjid="SUBJID",func.stat=mystat2,
func.stat.name="Coefficient of variation")
tab

mode=function(x) {
  x=na.omit(x)
  ux <- unique(x)
  ux[which.max(tabulate(match(x, ux)))]
}

tab=report.quanti(data=datafake,y="y_numeric",
func.stat=mode,func.stat.name="Mode")


#Getting raw output
tab$raw.output

#Getting a data.frame version of the output
tab$output

ClinReport documentation built on Sept. 3, 2019, 5:07 p.m.