Package 'coreStatsNMR' reference manual

Title:	Statistical Functions for Core Analysis Tasks at NMR
Description:	A set of statistical functions for use at NMR Group when completing core analysis tasks: frequency tables, cross-tabs, t-tests, proportion tests, etc.
Authors:	Julian Ricardo [aut, cre], Jerrad Pierce [ctb], Matt Woundy [ctb]
Maintainer:	Julian Ricardo <[email protected]>
License:	file LICENSE
Version:	1.3.6-7
Built:	2024-10-30 05:01:53 UTC
Source:	https://gitlab.com/NMRgroup/corestatsnmr

Area-weighted R-value

Description

Returns a single area-weighted R-value from the two vectors it accepts as arguments: one vector of R-values, and the other of areas associated with each R-value

Usage

aggRval(r_val, area)
aggRval(r_val, area)

Arguments

`r_val`	Vector of r-values
`area`	Vector of area associated with each r-value

Value

Single area-weighted r-value

Examples

aggRval(c(2,5,20), c(10,10,50))
aggRval(c(2,5,20), c(10,10,50))

Calculate confidence interval (using normal distribution)

Description

Calculate confidence interval (using normal distribution)

Usage

confInterval(x, conf_lvl = 0.9)
confInterval(x, conf_lvl = 0.9)

Arguments

`x`	Numerical vector
`conf_lvl`	A number from 0 to 1 indicating confidence level, defaults to 0.9 or 90%

Value

A dataframe summarizing the sample mean and confidence interval

Examples

confInterval(runif(100)); confInterval(runif(1e3))
confInterval(runif(100)); confInterval(runif(1e3))

Calculate the design effect for adjusting cluster sampling sizes

Description

Calculate the design effect for adjusting cluster sampling sizes

Usage

designEffect(n_obs, icc)
designEffect(n_obs, icc)

Arguments

`n_obs`	number. Observations in a cluster (e.g. average lamps in a home)
`icc`	number. Intraclass correlation (similarity of clustered data)

Value

A correction factor for sample sizes drawn from clustered units.

References

http://faculty.smu.edu/slstokes/stat6380/deff doc.pdf

Examples

designEffect(35, 0.75)
designEffect(35, 0.75)

Generate weights for data from sample and population counts

Description

Assumes data is provided in columns listing each category that is part of the weighting scheme, then a column for the sample n and a column for the general population.

Usage

makeWeights(data, sampleVal, populationVal, digits = 5, checkCols = FALSE)

## S3 method for class 'data.frame'
makeWeights(data, sampleVal, populationVal, digits = 5, checkCols = FALSE)
makeWeights(data, sampleVal, populationVal, digits = 5, checkCols = FALSE)

## S3 method for class 'data.frame'
makeWeights(data, sampleVal, populationVal, digits = 5, checkCols = FALSE)

Arguments

`data`	A data.frame (or data.table) to add weights to.
`sampleVal`	A string selecting the column in the data with sample counts
`populationVal`	A string selecting the column in the data with population counts
`digits`	A number of digits to use when rounding proportion weights
`checkCols`	A boolean that toggles whether to calculate checks on proportion and population (included as additional columns)

Value

A dataframe with population and proportion weights, as well as optional intermediate calculations.

Examples

myData <- data.frame(HairEyeColor)

myData$Population <- round(runif(nrow(myData),10000,20000),0)

makeWeights(data=myData,sampleVal="Freq",populationVal = "Population")

myData <- data.frame(HairEyeColor)

myData$Population <- round(runif(nrow(myData),10000,20000),0)

makeWeights(data=myData,sampleVal="Freq",populationVal = "Population")

Get the mode of a vector of values

Description

Get the mode of a vector of values

Usage

mode(x, show_all = FALSE)
mode(x, show_all = FALSE)

Arguments

`x`	A vector of values to calculate the mode from
`show_all`	A boolean, if FALSE (default) returns a single mode or NA if there are none/multiple. If TRUE, returns multiple modes, if they exist

Value

The mode(s) of the supplied vector.

Source

https://stackoverflow.com/questions/56552709/r-no-mode-and-exclude-na?noredirect=1#comment99692066_56552709

Pairwise proportion comparisons

Description

Pairwise proportion comparisons

Usage

pairPropTest(
  data,
  indexVar,
  valVar,
  grpVar,
  alpha = 0.1,
  n.min = 10,
  p.adjust.method = p.adjust.methods,
  counts = FALSE
)
pairPropTest(
  data,
  indexVar,
  valVar,
  grpVar,
  alpha = 0.1,
  n.min = 10,
  p.adjust.method = p.adjust.methods,
  counts = FALSE
)

Arguments

`data`	A dataset to calculate proportions for and test for statistically significant differences.
`indexVar`	string. Selects an index column for the dataset
`valVar`	string. Selects the column containing counts of successes in data
`grpVar`	string. Selects the column containing counts of trials in data
`alpha`	number. Significance level (e.g. 0.05 for 95-pct confidence level)
`n.min`	number. Minimum counts to consider
`p.adjust.method`	string. Method for adjusting p-values. See ?p.adjust for more details.
`counts`	Boolean. Toggles whether function returns significance results or counts (for diagnostic purposes)

Value

A dataframe showing p-values and statistically significant differences for the pairs of variables chosen

Pairwise T-test comparisons

Description

Pairwise T-test comparisons

Usage

pairTtest(
  data,
  valVar,
  grpVar,
  alpha = 0.1,
  n.min = 10,
  p.adjust.method = p.adjust.methods
)
pairTtest(
  data,
  valVar,
  grpVar,
  alpha = 0.1,
  n.min = 10,
  p.adjust.method = p.adjust.methods
)

Arguments

`data`	A dataset to calculate difference testing for and test for statistically significant differences.
`valVar`	string. Selects the column containing counts of successes in data
`grpVar`	string. Selects the column containing counts of trials in data
`alpha`	number. Significance level (e.g. 0.05 for 95-pct confidence level)
`n.min`	number. Minimum counts to consider
`p.adjust.method`	string. Method for adjusting p-values. See ?p.adjust for more details.

Value

A dataframe showing p-values and statistically significant differences for the pairs of variables chosen

Generating a penetration table

Description

Generating a penetration table

Generating a weighted proportion table (2-way)

Usage

penTable(
  data,
  index,
  x,
  y,
  totWeightVar = NULL,
  inGroupWeightVar = NULL,
  only_ns,
  accuracy,
  normwt = TRUE,
  tot.label = "Total"
)

## S3 method for class 'data.frame'
penTable(
  data,
  index,
  x,
  y,
  totWeightVar = NULL,
  inGroupWeightVar = NULL,
  only_ns = FALSE,
  accuracy = 1,
  normwt = TRUE,
  tot.label = "Total"
)
penTable(
  data,
  index,
  x,
  y,
  totWeightVar = NULL,
  inGroupWeightVar = NULL,
  only_ns,
  accuracy,
  normwt = TRUE,
  tot.label = "Total"
)

## S3 method for class 'data.frame'
penTable(
  data,
  index,
  x,
  y,
  totWeightVar = NULL,
  inGroupWeightVar = NULL,
  only_ns = FALSE,
  accuracy = 1,
  normwt = TRUE,
  tot.label = "Total"
)

Arguments

`data`	A dataset to calculate weighted proportions
`index`	string. Selects an index column for the dataset
`x`	string. Selects the first variable to find proportions for
`y`	string. Selects the second variable to find proportions for
`totWeightVar`	string. A string selecting the column to weight the population
`inGroupWeightVar`	string. A string selecting the column to use for in-group weights
`only_ns`	Boolean. Toggles whether to return penetration table or intermediate table of n's.
`accuracy`	number. A number to round to. Use (e.g.) 0.01 to show 2 decimal places of precision. If NULL, the default, uses a heuristic that should ensure breaks have the minimum number of digits needed to show the difference between adjacent values.
`normwt`	Boolean. if TRUE, normalize weights so that the total weighted count is the same as the unweighted one
`tot.label`	string. A string label for totals column

Value

A data.frame or data.table showing a penetration table

Proportion comparisons

Description

Proportion comparisons

Usage

propTest(
  data,
  indexVar,
  valVar,
  grpVar,
  counts = NULL,
  alpha = 0.1,
  n.min = 10,
  alternative = c("two.sided", "less", "greater")
)
propTest(
  data,
  indexVar,
  valVar,
  grpVar,
  counts = NULL,
  alpha = 0.1,
  n.min = 10,
  alternative = c("two.sided", "less", "greater")
)

Arguments

`data`	A dataset to calculate proportions for and test for statistically significant differences.
`valVar`	string. Selects the column containing counts of successes in data
`grpVar`	string. Selects the column containing counts of trials in data
`counts`	vector. Optional vector of strings containing columns counts for successes and trials (otherwise, function calculates counts from valVar and grpVar)
`alpha`	number. Significance level (e.g. 0.05 for 95-pct confidence level)
`n.min`	number. Minimum counts to consider
`alternative`	string. Specifies the alternative hypothesis. See ?prop.test

Value

A dataframe showing p-values and statistically significant differences for the chosen variables

Generate a statistical summary table, with optional grouping

Description

Generate a statistical summary table, with optional grouping

Usage

statsTable(
  data,
  summVar,
  groupVar = NULL,
  stats,
  accuracy = NULL,
  totCol = TRUE,
  totWeightVar = NULL,
  inGroupWeightVar = NULL,
  drop0trailing = FALSE,
  colOrder = NULL
)

## S3 method for class 'data.frame'
statsTable(
  data,
  summVar,
  groupVar = NULL,
  stats,
  accuracy = 1,
  totCol = TRUE,
  totWeightVar = NULL,
  inGroupWeightVar = NULL,
  drop0trailing = FALSE,
  colOrder = NULL
)

## S3 method for class 'data.table'
statsTable(
  data,
  summVar,
  groupVar = NULL,
  stats,
  accuracy = 1,
  totCol = TRUE,
  totWeightVar = NULL,
  inGroupWeightVar = NULL,
  drop0trailing = FALSE,
  colOrder = NULL
)
statsTable(
  data,
  summVar,
  groupVar = NULL,
  stats,
  accuracy = NULL,
  totCol = TRUE,
  totWeightVar = NULL,
  inGroupWeightVar = NULL,
  drop0trailing = FALSE,
  colOrder = NULL
)

## S3 method for class 'data.frame'
statsTable(
  data,
  summVar,
  groupVar = NULL,
  stats,
  accuracy = 1,
  totCol = TRUE,
  totWeightVar = NULL,
  inGroupWeightVar = NULL,
  drop0trailing = FALSE,
  colOrder = NULL
)

## S3 method for class 'data.table'
statsTable(
  data,
  summVar,
  groupVar = NULL,
  stats,
  accuracy = 1,
  totCol = TRUE,
  totWeightVar = NULL,
  inGroupWeightVar = NULL,
  drop0trailing = FALSE,
  colOrder = NULL
)

Arguments

`data`	A data.frame (or data.table) to use for statistical summary
`summVar`	A string selecting the column in 'data' to summarize
`groupVar`	A string or list of strings selecting the (optional) columns in 'data' to grop by
`stats`	A list of strings selecting summary stats functions (i.e. mean, sd, sum)
`accuracy`	A number to round to. Use (e.g.) 0.01 to show 2 decimal places of precision. If NULL, the default, uses a heuristic that should ensure breaks have the minimum number of digits needed to show the difference between adjacent values.
`totCol`	A boolean toggling whether to include a total column
`totWeightVar`	A string selecting the column to weight the population
`inGroupWeightVar`	A string selecting the column to use for in-group weights
`drop0trailing`	A boolean toggling whether to include trailing zeros in the output (converts to strings)
`colOrder`	To be deprecated

Value

A data.frame with statistical summary results describing the selected variable.

Examples

library(dplyr)

statsTable(iris,
             summVar = "Sepal.Length",
             groupVar = "Species",
             stats = c("n", "min", "max", "weighted.mean", "median", "sd"),
             accuracy = 2)
library(dplyr)

statsTable(iris,
             summVar = "Sepal.Length",
             groupVar = "Species",
             stats = c("n", "min", "max", "weighted.mean", "median", "sd"),
             accuracy = 2)

Conducting stratified random sampling

Description

Conducting stratified random sampling

Stratified random sampling

Usage

stratRandSample(
  data,
  group,
  size,
  select = NULL,
  replace = FALSE,
  bothSets = FALSE,
  keep.rownames = FALSE
)

## S3 method for class 'data.frame'
stratRandSample(
  data,
  group,
  size,
  select = NULL,
  replace = FALSE,
  bothSets = FALSE,
  keep.rownames = NULL
)

## S3 method for class 'data.table'
stratRandSample(
  data,
  group,
  size,
  select = NULL,
  replace = FALSE,
  bothSets = FALSE,
  keep.rownames = FALSE
)
stratRandSample(
  data,
  group,
  size,
  select = NULL,
  replace = FALSE,
  bothSets = FALSE,
  keep.rownames = FALSE
)

## S3 method for class 'data.frame'
stratRandSample(
  data,
  group,
  size,
  select = NULL,
  replace = FALSE,
  bothSets = FALSE,
  keep.rownames = NULL
)

## S3 method for class 'data.table'
stratRandSample(
  data,
  group,
  size,
  select = NULL,
  replace = FALSE,
  bothSets = FALSE,
  keep.rownames = FALSE
)

Arguments

`data`	A data.frame (or data.table) to use for allocating sample
`group`	string. The column(s) that represent strata
`size`	number. If <1, the proportion to take from each stratum. If an integer 1+, the number of samples to take from each stratum. If size is a vector of integers, the number of samples taken for each stratum. Recommended in latter case to use a named vector
`select`	list. Named list specifying a subset of strata to use in sampling
`replace`	boolean. Toggling whether to sample with replacement
`bothSets`	boolean. Toggling whether to return list of sampled and unsampled portions of data
`keep.rownames`	For data.tables only. See ?data.table. Adapted from https://gist.github.com/mrdwab/6424112 and https://gist.github.com/mrdwab/933ffeaa7a1d718bd10a

Value

A sample of the data passed to the function, optionally accounting for strata.

Tidy a weighted chi-squared contingency table test

Description

Tidy a weighted chi-squared contingency table test

Usage

## S3 method for class 'wtd.chi.sq'
tidy(x)
## S3 method for class 'wtd.chi.sq'
tidy(x)

Arguments

`x`	An htest object, such as those created by weights::wtd.chi.sq

Value

A tibble::tibble() with columns for method, coefficients, estimated values, p-value, and other statistics

Tidy a weighted t-test object

Description

Tidy a weighted t-test object

Usage

## S3 method for class 'wtd.t.test'
tidy(x)
## S3 method for class 'wtd.t.test'
tidy(x)

Arguments

`x`	An htest object, such as those created by weights::wtd.t.test()

Value

A tibble::tibble() with columns for method, coefficients, estimated values, p-value, and other statistics

Generating a weighted frequency table (2-way)

Description

Generating a weighted frequency table (2-way)

Usage

wtdFreqTable(
  data,
  x,
  y,
  totWeightVar = NULL,
  inGroupWeightVar = NULL,
  accuracy = 1,
  normwt = TRUE,
  tot.label = "Statewide",
  colOrder = NULL
)

## S3 method for class 'data.frame'
wtdFreqTable(
  data,
  x,
  y,
  totWeightVar = NULL,
  inGroupWeightVar = NULL,
  accuracy = 1,
  normwt = TRUE,
  tot.label = "Total",
  colOrder = NULL
)
wtdFreqTable(
  data,
  x,
  y,
  totWeightVar = NULL,
  inGroupWeightVar = NULL,
  accuracy = 1,
  normwt = TRUE,
  tot.label = "Statewide",
  colOrder = NULL
)

## S3 method for class 'data.frame'
wtdFreqTable(
  data,
  x,
  y,
  totWeightVar = NULL,
  inGroupWeightVar = NULL,
  accuracy = 1,
  normwt = TRUE,
  tot.label = "Total",
  colOrder = NULL
)

Arguments

`data`	A dataset to calculate weighted frequencies. Only operational for data.frame for now
`x`	string. Selects the first variable to find frequencies for
`y`	string. Selects the second variable to find frequencies for
`totWeightVar`	string. A string selecting the column to weight the population
`inGroupWeightVar`	string. A string selecting the column to use for in-group weights
`accuracy`	number. A number to round to. Use (e.g.) 0.01 to show 2 decimal places of precision. If NULL, the default, uses a heuristic that should ensure breaks have the minimum number of digits needed to show the difference between adjacent values.
`normwt`	Boolean. if TRUE, normalize weights so that the total weighted count is the same as the unweighted one
`tot.label`	string. Label for totals column
`colOrder`	vector. Vector of strings to set the order for the colum given by variable x

Value

A data.frame showing a two-way weighted frequency table

Weighted pairwise proportion comparisons

Description

Weighted pairwise proportion comparisons

Usage

wtdPairPropTest(
  data,
  indexVar,
  valVar,
  grpVar,
  weightVar,
  alpha = 0.1,
  n.min = 10,
  p.adjust.method = p.adjust.methods,
  counts = FALSE
)
wtdPairPropTest(
  data,
  indexVar,
  valVar,
  grpVar,
  weightVar,
  alpha = 0.1,
  n.min = 10,
  p.adjust.method = p.adjust.methods,
  counts = FALSE
)

Arguments

`data`	A dataset to calculate proportions for and test for statistically significant differences.
`indexVar`	string. Selects an index column for the dataset
`valVar`	string. Selects the column containing counts of successes in data
`grpVar`	string. Selects the column containing counts of trials in data
`weightVar`	string. Selects the column containing weights in the data
`alpha`	number. Significance level (e.g. 0.05 for 95-pct confidence level)
`n.min`	number. Minimum counts to consider
`p.adjust.method`	string. Method for adjusting p-values. See ?p.adjust for more details.
`counts`	Boolean. Toggles whether function returns significance results or counts (for diagnostic purposes)

Value

A dataframe showing p-values and statistically significant differences for the pairs of variables chosen

Pairwise Weighted T-Test comparisons

Description

Pairwise Weighted T-Test comparisons

Usage

wtdPairTtest(
  data,
  valVar,
  grpVar,
  weightVar,
  alpha = 0.1,
  n.min = 10,
  p.adjust.method = p.adjust.methods
)
wtdPairTtest(
  data,
  valVar,
  grpVar,
  weightVar,
  alpha = 0.1,
  n.min = 10,
  p.adjust.method = p.adjust.methods
)

Arguments

`data`	A dataset to calculate difference testing for and test for statistically significant differences.
`valVar`	string. Selects the column containing counts of successes in data
`grpVar`	string. Selects the column containing counts of trials in data
`weightVar`	string. Selects the column containing weights in the data
`alpha`	number. Significance level (e.g. 0.05 for 95-pct confidence level)
`n.min`	number. Minimum counts to consider
`p.adjust.method`	string. Method for adjusting p-values. See ?p.adjust for more details.

Value

A dataframe showing p-values and statistically significant differences for the pairs of variables chosen

Generating a weighted proportion table (2-way)

Description

Generating a weighted proportion table (2-way)

Usage

wtdPropTable(
  data,
  x,
  y,
  totWeightVar = NULL,
  inGroupWeightVar = NULL,
  pct_format = TRUE,
  accuracy = 0.1,
  normwt = TRUE,
  tot.label = "Total",
  colOrder = NULL
)

## S3 method for class 'data.frame'
wtdPropTable(
  data,
  x,
  y,
  totWeightVar = NULL,
  inGroupWeightVar = NULL,
  pct_format = TRUE,
  accuracy = 0.1,
  normwt = TRUE,
  tot.label = "Total",
  colOrder = NULL
)
wtdPropTable(
  data,
  x,
  y,
  totWeightVar = NULL,
  inGroupWeightVar = NULL,
  pct_format = TRUE,
  accuracy = 0.1,
  normwt = TRUE,
  tot.label = "Total",
  colOrder = NULL
)

## S3 method for class 'data.frame'
wtdPropTable(
  data,
  x,
  y,
  totWeightVar = NULL,
  inGroupWeightVar = NULL,
  pct_format = TRUE,
  accuracy = 0.1,
  normwt = TRUE,
  tot.label = "Total",
  colOrder = NULL
)

Arguments

`data`	A dataset to calculate weighted proportions
`x`	string. Selects the first variable to find proportions for
`y`	string. Selects the second variable to find proportions for
`totWeightVar`	string. A string selecting the column to weight the population
`inGroupWeightVar`	string. A string selecting the column to use for in-group weights
`pct_format`	boolean. Toggles whether proportions are given as decimals or percents (converts to strings)
`accuracy`	number. A number to round to. Use (e.g.) 0.01 to show 2 decimal places of precision. If NULL, the default, uses a heuristic that should ensure breaks have the minimum number of digits needed to show the difference between adjacent values.
`normwt`	Boolean. if TRUE, normalize weights so that the total weighted count is the same as the unweighted one
`tot.label`	string. A string label for totals column
`colOrder`	vector. A vector of strings to set the order for the column given by variable x

Value

A data.frame or data.table showing a two-way weighted proportion table

Package 'coreStatsNMR'

Help Index

Area-weighted R-value

Description

Usage

Arguments

Value

Examples

Calculate confidence interval (using normal distribution)

Description

Usage

Arguments

Value

Examples

Calculate the design effect for adjusting cluster sampling sizes

Description

Usage

Arguments

Value

References

Examples

Generate weights for data from sample and population counts

Description

Usage

Arguments

Value

Examples

Get the mode of a vector of values

Description

Usage

Arguments

Value

Source

Pairwise proportion comparisons

Description

Usage

Arguments

Value

Pairwise T-test comparisons

Description

Usage

Arguments

Value

Generating a penetration table

Description

Usage

Arguments

Value

Proportion comparisons

Description

Usage

Arguments

Value

Generate a statistical summary table, with optional grouping

Description

Usage

Arguments

Value

Examples

Conducting stratified random sampling

Description

Usage

Arguments

Value

Tidy a weighted chi-squared contingency table test

Description

Usage

Arguments

Value

Tidy a weighted t-test object

Description

Usage

Arguments

Value

Generating a weighted frequency table (2-way)

Description

Usage

Arguments

Value

Weighted pairwise proportion comparisons