Package 'iNZightMR' reference manual

Title:	Tools for Exploring Multiple Response Data
Description:	Interaction and analysis of multiple response data, along with other tools for analysing these types of data including missing value analysis and calculation of standard errors for a range of covariance matrix results (proportions, multinomial, independent samples, and multiple response).
Authors:	Tom Elliott [aut, cre] , Junjie Zeng [aut] (Original author), Simon Potter [aut]
Maintainer:	Tom Elliott <[email protected]>
License:	GPL-3
Version:	2.3.0
Built:	2025-03-13 05:32:45 UTC
Source:	https://github.com/inzightvit/inzightmr

iNZightMR: Multiple Response Data Analysis

Description

The iNZightMR package provides a suite of functions which can be used in the analysis of multiple response data. It is used in the iNZight software package.

Author(s)

Junjie Zeng, Tom Elliott

Multiple response barplot

Description

Draws a barplot of a multiple response object (MRO), showing response rates for each option along with confidence intervals and comparison intervals.

Usage

barplotMR(obj, ...)

## S3 method for class 'mrocalc'
barplotMR(obj, ...)

## S3 method for class 'bymrocalc'
barplotMR(obj, g1.level = NULL, g2.level = "_MULTI", ...)

## S3 method for class 'between'
barplotMR(obj, ...)

## S3 method for class 'b2'
barplotMR(obj, g1.level = NULL, ...)
barplotMR(obj, ...)

## S3 method for class 'mrocalc'
barplotMR(obj, ...)

## S3 method for class 'bymrocalc'
barplotMR(obj, g1.level = NULL, g2.level = "_MULTI", ...)

## S3 method for class 'between'
barplotMR(obj, ...)

## S3 method for class 'b2'
barplotMR(obj, g1.level = NULL, ...)

Arguments

`obj`	an `mrocalc` object (from `mroPara()`)
`...`	additional parameters, currently not used
`g1.level`	vector of subset variable 1 levels to show
`g2.level`	vector of subset variable 2 levels to show

Methods (by class)

barplotMR(mrocalc): method for class mrocalc
barplotMR(bymrocalc): method for class bymrocalc
barplotMR(between): method for class between
barplotMR(b2): method for class b2

Author(s)

Junjie Zheng

Examples

if (requireNamespace("iNZightPlots")) {
    require(iNZightPlots)
    mr <- iNZightMR(online ~ onlinegame + onlinevideo + onlinemusic,
        data = census.at.school.5000
    )
    barplotMR(mroPara(mr))

    barplotMR(byMRO(mr, ~gender, mroPara))
}
if (requireNamespace("iNZightPlots")) {
    require(iNZightPlots)
    mr <- iNZightMR(online ~ onlinegame + onlinevideo + onlinemusic,
        data = census.at.school.5000
    )
    barplotMR(mroPara(mr))

    barplotMR(byMRO(mr, ~gender, mroPara))
}

Compute Between se's

Description

Between SEs

Usage

between(bymro)
between(bymro)

Arguments

bymro

a bymro object

Value

something about between.

Author(s)

Junjie Zheng

Examples

mr <- iNZightMR(online ~ onlinegame + onlinevideo + onlinemusic,
    data = census.at.school.5000)
(bt <- between(byMRO(mr, ~gender, mroPara)))

if (requireNamespace("iNZightPlots"))
    barplotMR(bt)
mr <- iNZightMR(online ~ onlinegame + onlinevideo + onlinemusic,
    data = census.at.school.5000)
(bt <- between(byMRO(mr, ~gender, mroPara)))

if (requireNamespace("iNZightPlots"))
    barplotMR(bt)

Calculate MRO inference for subsets

Description

Constructs a multiple response object (MRO) subset by another explanatory variable.

Usage

byMRO(mro.obj, formula, FUN, ...)
byMRO(mro.obj, formula, FUN, ...)

Arguments

`mro.obj`	an mro object (created by `iNZightMR`)
`formula`	variable for subsetting, as a formula (e.g., `~x`)
`FUN`	the function to apply to subsets
`...`	additional arguments passed to `FUN`

Value

an object with classes of by and bymrocalc

Examples

mr <- iNZightMR(online ~ onlinegame + onlinevideo + onlinemusic,
    data = census.at.school.5000
)

byMRO(mr, ~gender, mroPara)
byMRO(mr, ~ gender + handed, mroPara)
mr <- iNZightMR(online ~ onlinegame + onlinevideo + onlinemusic,
    data = census.at.school.5000
)

byMRO(mr, ~gender, mroPara)
byMRO(mr, ~ gender + handed, mroPara)

Calculate missing observation summary

Description

Calculates the summary of missingness in a data set.

Usage

calcmissing(obj, ...)

## S3 method for class 'data.frame'
calcmissing(obj, MRO.case = FALSE, ...)

## S3 method for class 'mro'
calcmissing(obj, ...)
calcmissing(obj, ...)

## S3 method for class 'data.frame'
calcmissing(obj, MRO.case = FALSE, ...)

## S3 method for class 'mro'
calcmissing(obj, ...)

Arguments

`obj`	An object
`...`	additional arguments
`MRO.case`	does something with rownames

Value

Missing value object

Methods (by class)

calcmissing(data.frame): Method for a dataframe
calcmissing(mro): accepts a whole mr.object , which is first mro.mat, second element labels, third element the input data frame.

Author(s)

Junjie Zeng

Examples

calcmissing(census.at.school.5000[, 1:20])
calcmissing(census.at.school.5000[, 1:20])

Census at School 5000

Description

A dataset containing 5000 observations from a New Zealand census of school students. It includes binary response variables.

Usage

census.at.school.5000
census.at.school.5000

Format

A data frame with 72 variables and 5000 rows.

X: unique identifier for each observation
gender: their biological gender
age: their age, years
country: The country the student is from
country_en: Country code
country_mi: A different country code
ethniceng: binary for ethnicity english
ethnicmri: binary for ethnicity maori
ethnicwsm: binary for ethnicity wsm
ethniccok: binary for ethnicity cok
ethnicton: binary for ethnicity tonga
ethnicniu: binary for ethnicity niue
ethnicchn: binary for ethnicity china
ethnicind: binary for ethnicity india
ethnicother: factor for other ethnicity
ethnicother_en: factor for other ethnicity_en
ethnicother_mi: factor for other ethnicity_mi
languages: how many languages they know
handed: left, right, or ambi
height: height measurement, cm
rightfoot: length of the right foot, mm
armspan: their armspan measurement, cm
wrist: wrist measurement
neck: neck measurement
popliteal: another measurement
indexfinger: index finger measurement
ringfinger: ring finger measurement
hairlength: the length of their hair
travel: travel method used to get to school
timetravel: how long they spend travelling
bagweightraw: the weight of the bag
bagweight: weight of the bag
bagcarry: factor of how they're carrying
favlearning: their favourite subject
favlearningmo: their favourite subject?
memory: a memory score
reaction: a reaction score
sport: what sport they play
sport_en: what sport they play
techtv: binary for use of TV
techmp3: binary for use of mp3
techinternet: binary for use of the internet
techmobinternet: binary for use of mobile internet
techfacebook: binary for use of facebook
techtwitter: binary for use of twitter
techbebo: binary for use of bebo
techmyspace: binary for use of myspace
techskype: binary for use of skype
techconsole: binary for use of a console
technone: binary for use of no technology
cellmonths: how many months they've had a cellphone?
onlinemusic: binary for if they listen to music online
onlinevideo: binary for if they watch video online
onlinegame: binary for if they play games online
onlinefriend: binary for if they talk to friends online
onlineschool: binary for if they access school online
onlineother: binary for if they do anything else online
onlinenone: binary for if they do nothing online
bedtime: hours spent in bed
waketime: hours spent awake
favtvshow: the name of their favourite TV show
favtvshow_en: the name of their favourite TV show
importwarm: binary about warm
importpollution: binary about pollution
importrecycling: binary about recycling
importwater: binary about water
importlifestyle: binary about lifestyle
importenergy: binary about energy
importgovern: binary about government
importcomputer: binary about computer
region: number of the region they're in
year: their school year

Source

http://new.censusatschool.org.nz/

Create iNZightMR multiple response object (MRO)

Description

Creates a multiple response object (MRO) containing binary response matrix (zeros and ones) as well as the input data source.

Usage

iNZightMR(frm, data, Labels = NULL, inverse = FALSE, ...)
iNZightMR(frm, data, Labels = NULL, inverse = FALSE, ...)

Arguments

`frm`	formula containing the response variables
`data`	a data.frame containing response and explanatory variables
`Labels`	labels for the response categories; by default, the function will attempt to. Can also be the function `substrsplit`, which will detect a common base in the variables (see Details)
`inverse`	if `TRUE`, binary responses will be reversed (see details)
`...`	additional arguments, passed to `model.frame`

Details

The individual response variable names can be detected from the variable name by passing Labels = substrsplit. For example, in 'ethniceng' and 'ethnicmri', 'ethnic' is common to both, so the labels will be 'eng' and 'mri'.

If a user wants to inverse the response (zeros becomes ones), then pass inverse = TRUE. This is useful when the responses are characters (such as "correct" and "wrong", where correct would be given a zero) and the order needs reversing (so that correct is 1 instead).

Value

An mro object containing a multiple response binary matrix and input data source

Examples

mr <- iNZightMR(online ~ onlinegame + onlinevideo + onlinemusic,
    data = census.at.school.5000)

# users can also override the variable names
iNZightMR(online ~ onlinegame + onlinevideo + onlinemusic,
    Labels = c("gaming", "youtube", "spotify"),
    data = census.at.school.5000
)
mr <- iNZightMR(online ~ onlinegame + onlinevideo + onlinemusic,
    data = census.at.school.5000)

# users can also override the variable names
iNZightMR(online ~ onlinegame + onlinevideo + onlinemusic,
    Labels = c("gaming", "youtube", "spotify"),
    data = census.at.school.5000
)

Margin of Error Calculation

Description

Computes the margin of error for various objects.

Usage

moecalc(
  x,
  factorname = NULL,
  levelnames = NULL,
  coef.idx = NULL,
  est = NULL,
  ci = NULL,
  base = TRUE,
  basename = "base",
  conf.level = 1.96
)
moecalc(
  x,
  factorname = NULL,
  levelnames = NULL,
  coef.idx = NULL,
  est = NULL,
  ci = NULL,
  base = TRUE,
  basename = "base",
  conf.level = 1.96
)

Arguments

`x`	the object for which we compute margins of error
`factorname`	name of factor
`levelnames`	names of factor levels
`coef.idx`	index of coefficient to use
`est`	estimates
`ci`	confidence intervals
`base`	baseline
`basename`	name of baseline
`conf.level`	level of confidence to use

Details

If x is a model, must have factorname or coefficient index (coef.idx) If input factorname, will compute ErrBars by factorname (for given model) If input coefficient index, will compute ErrBars simply by index only (even they are not factor) If x is ses.moecalc object, will compute ErrBars simply by given ses.moecalc object

Value

a moecalc object

Examples

fit <- lm(Sepal.Length ~ Species, data = iris)
(mc <- moecalc(fit, "Species"))
summary(mc)
plot(mc)
fit <- lm(Sepal.Length ~ Species, data = iris)
(mc <- moecalc(fit, "Species"))
summary(mc)
plot(mc)

Calculate MRO inference

Description

Calculates required proportions, their differences, variance-covariance matrices, standard errors of differences, and comparison intervals for differences, over all of the data. To compute values over various subsets of another explanatory variable, see by.

Usage

mroPara(obj, conf.levels = 1.96, nonparallel = NULL)
mroPara(obj, conf.levels = 1.96, nonparallel = NULL)

Arguments

`obj`	an MRO object created by `iNZightMR`
`conf.levels`	confidence level to use, default is 1.96 for 95% intervals
`nonparallel`	Should these things be parallel?

Value

An object of class mrocalc

Examples

mr <- iNZightMR(online ~ onlinegame + onlinevideo + onlinemusic,
    data = census.at.school.5000)
mrp <- mroPara(mr)
mr <- iNZightMR(online ~ onlinegame + onlinevideo + onlinemusic,
    data = census.at.school.5000)
mrp <- mroPara(mr)

Missing Value plot

Description

Plot of Missing Value combinations

Usage

plotcombn(obj)
plotcombn(obj)

Arguments

obj

a calcmissing object

Value

summarised info for plot

Author(s)

Junjie Zeng

Examples

plotcombn(census.at.school.5000[, 10:25])
plotcombn(census.at.school.5000[, 10:25])

Independent Binomial Proportions

Description

Compute SEs for Independent Binomial Proportions

Usage

seBinprops(ns, phats)
seBinprops(ns, phats)

Arguments

`ns`	the number of observations in the independent groups
`phats`	the proportions of TRUE/1's etc.

Value

an ses.moecalc object

Author(s)

Junjie Zeng

Examples

seBinprops(c(50, 30), c(0.3, 0.7))
seBinprops(c(50, 30), c(0.3, 0.7))

Compute standard error for covariance matrix

Description

Compute the standard error information for a given covariance matrix.

Usage

seCovs(covs, addbase = FALSE)
seCovs(covs, addbase = FALSE)

Arguments

`covs`	covariance matrix
`addbase`	logical, is there a baseline?

Value

an ses.moecalc object

Author(s)

Junjie Zeng

Examples

seCovs(cov(iris[, -5]))
seCovs(cov(iris[, -5]))

Independent Standard errors given

Description

Returns ses.moecalc for given SEs

Usage

seIndepSes(ses)
seIndepSes(ses)

Arguments

ses

the standard errors

Value

an ses.moecalc object

Author(s)

Junjie Zeng

Examples

seIndepSes(c(0.02, 0.05, 0.1))
seIndepSes(c(0.02, 0.05, 0.1))

Compute SE for Multinomial proportions

Description

SEs for Multinomial Proportions

Usage

seMNprops(n, phat)
seMNprops(n, phat)

Arguments

`n`	the number of observations in each group
`phat`	the estimates proportions for each group

Value

an ses.moecalc object

Author(s)

Junjie Zeng

Examples

phat <- table(iris$Species) / nrow(iris)
seMNprops(nrow(iris), phat)
phat <- table(iris$Species) / nrow(iris)
seMNprops(nrow(iris), phat)

Multiple binary response

Description

SE's for multiple binary response

Usage

seMRprops(obj)
seMRprops(obj)

Arguments

obj

something that can be turned into a matrix

Value

an ses.moecalc object

Author(s)

Junjie Zeng

Examples

x <- data.frame(
    v1 = rbinom(20, 1, 0.8),
    v2 = rbinom(20, 1, 0.3),
    v3 = rbinom(20, 1, 0.5)
)
seMRprops(x)
x <- data.frame(
    v1 = rbinom(20, 1, 0.8),
    v2 = rbinom(20, 1, 0.3),
    v3 = rbinom(20, 1, 0.5)
)
seMRprops(x)

Extract Common Name from variables

Description

Help mro variables extract common name out

Usage

substrsplit(obj, split_string = "")
substrsplit(obj, split_string = "")

Arguments

`obj`	It can be a vector or data frame, however, `substrsplit` is usually used in the `iNZightMR` function.
`split_string`	Specification of a string to allow splits on (e.g., use " " to only allow splitting at words).

Value

A list with common character and unique variable name respectively

Examples

substrsplit(c("varx", "vary", "varz"))

strings <- c("Do you like eating eggs?", "Do you like elephants?" , "Do you like elections?")
substrsplit(strings)
substrsplit(strings, split_string = " ")
substrsplit(c("varx", "vary", "varz"))

strings <- c("Do you like eating eggs?", "Do you like elephants?" , "Do you like elections?")
substrsplit(strings)
substrsplit(strings, split_string = " ")

Package 'iNZightMR'

Help Index

iNZightMR: Multiple Response Data Analysis

Description

Author(s)

Multiple response barplot

Description

Usage

Arguments

Methods (by class)

Author(s)

Examples

Compute Between se's

Description

Usage

Arguments

Value

Author(s)

Examples

Calculate MRO inference for subsets

Description

Usage

Arguments

Value

See Also

Examples

Calculate missing observation summary

Description

Usage

Arguments

Value

Methods (by class)

Author(s)

See Also

Examples

Census at School 5000

Description

Usage

Format

Source

Create iNZightMR multiple response object (MRO)

Description

Usage

Arguments

Details

Value

See Also

Examples

Margin of Error Calculation

Description

Usage

Arguments

Details

Value

Examples

Calculate MRO inference

Description

Usage

Arguments

Value

See Also

Examples

Missing Value plot

Description

Usage

Arguments

Value

Author(s)

Examples

Independent Binomial Proportions

Description

Usage

Arguments

Value

Author(s)

Examples

Compute standard error for covariance matrix

Description

Usage

Arguments