Select a Statistical Audit Sample

This function takes a data frame and performs statistical selection according to one of four algorithms: fixed interval sampling, cell sampling, random sampling, and modified sieve sampling. Selection is done on the level of two possible sampling units: items (records / rows) or monetary units. The function returns an object of class jfaSelection which can be used with associated summary() and a plot() methods.

For more details on how to use this function, see the package vignette: vignette('jfa', package = 'jfa')

selection(data, size, units = c('items', 'values'),
          method = c('interval', 'cell', 'random', 'sieve'), values = NULL,
          order = NULL, decreasing = FALSE, randomize = FALSE,
          replace = FALSE, start = 1)

Arguments

data: a data frame containing the population of items the auditor wishes to sample from.
size: an integer larger than 0 specifying the number of sampling units that need to be selected from the population. Can also be an object of class jfaPlanning.
units: a character specifying the sampling units used. Possible options are items (default) for selection on the level of items (rows) or values for selection on the level of monetary units.
method: a character specifying the sampling algorithm used. Possible options are interval (default) for fixed interval sampling, cell for cell sampling, random for random sampling, or sieve for modified sieve sampling.
values: a character specifying the name of a column in data containing the book values of the items.
order: a character specifying the name of a column in data containing the ranks of the items. The items in the data are ordered according to these values in the order indicated by decreasing.
decreasing: if order is specified, a logical specifying whether to order the items from smallest to largest. Defaults to FALSE.
randomize: a logical specifying whether the items in the data should be randomly shuffled before selection. Defaults to FALSE. Note that specifying if randomize = TRUE overrules order.
replace: if method = 'random', a logical specifying whether sampling should be performed with replacement. Defaults to FALSE.
start: if method = 'interval', an integer larger than 0 specifying the starting point of the algorithm.

Value

An object of class jfaSelection containing:

data: a data frame containing the input data.
sample: a data frame containing the selected sample of items.
n.req: an integer indicating the requested sample size.
n.units: an integer indicating the total number of obtained sampling units.
n.items: an integer indicating the total number of obtained sample items.
N.units: an integer indicating the total number of sampling units in the population.
N.items: an integer indicating the total number of items in the population.
interval: if method = 'interval', a numeric value indicating the size of the selection interval.
units: a character indicating the sampling units that were used to create the selection.
method: a character indicating the the algorithm that was used to create the selection.
values: if values is specified, a character indicating the name of the book value column.
start: if method = 'interval', an integer indicating the starting point in the interval.
data.name: a character string giving the name of the data.

Details

The first part of this section elaborates on the two possible options for the units argument:

items: In record sampling each item in the population is seen as a sampling unit. An item of $5000 is therefore equally likely to be selected as an item of $500.
values: In monetary unit sampling each monetary unit in the population is seen as a sampling unit. An item of $5000 is therefore ten times more likely to be selected as an item of $500.

The second part of this section elaborates on the four possible options for the method argument:

interval: In fixed interval sampling the sampling units in the population are divided into a number (equal to the sample size) of intervals. From each interval one sampling unit is selected according to a fixed starting point (specified by start).
cell: In cell sampling the sampling units in the population are divided into a number (equal to the sample size) of intervals. From each interval one sampling unit is selected with equal probability.
random: In random sampling each sampling unit in the population is drawn with equal probability.
sieve: In modified sieve sampling each item in the population is selected proportional to its value (Hoogduin, Hall, & Tsay, 2010).

References

Hoogduin, L. A., Hall, T. W., & Tsay, J. J. (2010). Modified sieve sampling: A method for single-and multi-stage probability-proportional-to-size sampling. Auditing: A Journal of Practice & Theory, 29(1), 125-148.

Leslie, D. A., Teitlebaum, A. D., & Anderson, R. J. (1979). Dollar-unit Sampling: A Practical Guide for Auditors. Copp Clark Pitman; Belmont, Calif.: distributed by Fearon-Pitman.

Wampler, B., & McEacharn, M. (2005). Monetary-unit sampling using Microsoft Excel. The CPA journal, 75(5), 36.

Author

Koen Derks, k.derks@nyenrode.nl

Examples

data("BuildIt")

# Select 100 items using random sampling
selection(data = BuildIt, size = 100, method = "random")
#> 
#> 	Audit Sample Selection
#> 
#> data:  BuildIt
#> number of sampling units = 100, number of items = 100
#> sample selected via method 'items' + 'random'

# Select 150 monetary units using fixed interval sampling
selection(
  data = BuildIt, size = 150, units = "values",
  method = "interval", values = "bookValue"
)
#> 
#> 	Audit Sample Selection
#> 
#> data:  BuildIt
#> number of sampling units = 150, number of items = 150
#> sample selected via method 'values' + 'interval'