| Type: | Package |
| Title: | The 'jamovi' Analyses |
| Version: | 2.7.7 |
| Date: | 2025-09-18 |
| Maintainer: | Jonathon Love <jon@thon.cc> |
| Description: | A suite of common statistical methods such as descriptives, t-tests, ANOVAs, regression, correlation matrices, proportion tests, contingency tables, and factor analysis. This package is also useable from the 'jamovi' statistical spreadsheet (see https://www.jamovi.org for more information). |
| BugReports: | https://github.com/jamovi/jmv/issues |
| License: | GPL-2 | GPL-3 [expanded from: GPL (≥ 2)] |
| Depends: | R (≥ 3.2) |
| Imports: | jmvcore (≥ 2.4.2), R6, car (≥ 3.0.0), multcomp, ggplot2 (≥ 2.2.1), PMCMR, emmeans (≥ 1.4.2), vcd, vcdExtra, GGally, BayesFactor, psych (≥ 1.7.5), GPArotation, afex (≥ 0.28-0), mvnormtest, lavaan, ggridges, ROCR, nnet, MASS, ggrepel, dplyr, magrittr, matrixStats |
| Suggests: | exact2x2, testthat (≥ 3.1.5), semPlot, carData, knitr, rmarkdown |
| Encoding: | UTF-8 |
| RoxygenNote: | 6.1.1 |
| NeedsCompilation: | no |
| Packaged: | 2025-09-17 23:47:37 UTC; c3113592 |
| Author: | Ravi Selker [aut, cph], Jonathon Love [aut, cre, cph], Damian Dropmann [aut, cph], Victor Moreno [ctb, cph], Maurizio Agosti [ctb, cph], Sebastian Jentschke [ctb, cph] |
| Repository: | CRAN |
| Date/Publication: | 2025-09-18 00:40:02 UTC |
ANOVA
Description
The Analysis of Variance (ANOVA) is used to explore the relationship between a continuous dependent variable, and one or more categorical explanatory variables.
Usage
ANOVA(data, dep, factors = NULL, effectSize = NULL,
modelTest = FALSE, modelTerms = NULL, ss = "3", homo = FALSE,
norm = FALSE, qq = FALSE, contrasts = NULL, postHoc = NULL,
postHocCorr = list("tukey"), postHocES = list(),
postHocEsCi = FALSE, postHocEsCiWidth = 95, emMeans = list(list()),
emmPlots = TRUE, emmPlotData = FALSE, emmPlotError = "ci",
emmTables = FALSE, emmWeights = TRUE, ciWidthEmm = 95, formula)
Arguments
data |
the data as a data frame |
dep |
the dependent variable from |
factors |
the explanatory factors in |
effectSize |
one or more of |
modelTest |
|
modelTerms |
a formula describing the terms to go into the model (not necessary when providing a formula, see examples) |
ss |
|
homo |
|
norm |
|
qq |
|
contrasts |
a list of lists specifying the factor and type of contrast
to use, one of |
postHoc |
a formula containing the terms to perform post-hoc tests on (see the examples) |
postHocCorr |
one or more of |
postHocES |
a possible value of |
postHocEsCi |
|
postHocEsCiWidth |
a number between 50 and 99.9 (default: 95), the width of confidence intervals for the post-hoc effect sizes |
emMeans |
a formula containing the terms to estimate marginal means for (see the examples) |
emmPlots |
|
emmPlotData |
|
emmPlotError |
|
emmTables |
|
emmWeights |
|
ciWidthEmm |
a number between 50 and 99.9 (default: 95) specifying the confidence interval width for the estimated marginal means |
formula |
(optional) the formula to use, see the examples |
Details
ANOVA assumes that the residuals are normally distributed, and that the variances of all groups are equal. If one is unwilling to assume that the variances are equal, then a Welch's test can be used instead (However, the Welch's test does not support more than one explanatory factor). Alternatively, if one is unwilling to assume that the data is normally distributed, a non-parametric approach (such as Kruskal-Wallis) can be used.
Value
A results object containing:
results$main | a table of ANOVA results | ||||
results$model | The underlying aov object |
||||
results$assump$homo | a table of homogeneity tests | ||||
results$assump$norm | a table of normality tests | ||||
results$assump$qq | a q-q plot | ||||
results$contrasts | an array of contrasts tables | ||||
results$postHoc | an array of post-hoc tables | ||||
results$emm | an array of the estimated marginal means plots + tables | ||||
results$residsOV | an output | ||||
Tables can be converted to data frames with asDF or as.data.frame. For example:
results$main$asDF
as.data.frame(results$main)
Examples
data('ToothGrowth')
ANOVA(formula = len ~ dose * supp, data = ToothGrowth)
#
# ANOVA
#
# ANOVA
# -----------------------------------------------------------------------
# Sum of Squares df Mean Square F p
# -----------------------------------------------------------------------
# dose 2426 2 1213.2 92.00 < .001
# supp 205 1 205.4 15.57 < .001
# dose:supp 108 2 54.2 4.11 0.022
# Residuals 712 54 13.2
# -----------------------------------------------------------------------
#
ANOVA(
formula = len ~ dose * supp,
data = ToothGrowth,
emMeans = ~ supp + dose:supp, # est. marginal means for supp and dose:supp
emmPlots = TRUE, # produce plots of those marginal means
emmTables = TRUE) # produce tables of those marginal means
Big 5
Description
Big 5
Tooth Growth
Description
Tooth Growth
ANCOVA
Description
The Analysis of Covariance (ANCOVA) is used to explore the relationship between a continuous dependent variable, one or more categorical explanatory variables, and one or more continuous explanatory variables (or covariates). It is essentially the same analysis as ANOVA, but with the addition of covariates.
Usage
ancova(data, dep, factors = NULL, covs = NULL, effectSize = NULL,
modelTest = FALSE, modelTerms = NULL, ss = "3", homo = FALSE,
norm = FALSE, qq = FALSE, contrasts = NULL, postHoc = NULL,
postHocCorr = list("tukey"), postHocES = list(),
postHocEsCi = FALSE, postHocEsCiWidth = 95, emMeans = list(list()),
emmPlots = TRUE, emmPlotData = FALSE, emmPlotError = "ci",
emmTables = FALSE, emmWeights = TRUE, ciWidthEmm = 95, formula)
Arguments
data |
the data as a data frame |
dep |
the dependent variable from |
factors |
the explanatory factors in |
covs |
the explanatory covariates (not necessary when providing a formula, see examples) |
effectSize |
one or more of |
modelTest |
|
modelTerms |
a formula describing the terms to go into the model (not necessary when providing a formula, see examples) |
ss |
|
homo |
|
norm |
|
qq |
|
contrasts |
a list of lists specifying the factor and type of contrast
to use, one of |
postHoc |
a formula containing the terms to perform post-hoc tests on (see the examples) |
postHocCorr |
one or more of |
postHocES |
a possible value of |
postHocEsCi |
|
postHocEsCiWidth |
a number between 50 and 99.9 (default: 95), the width of confidence intervals for the post-hoc effect sizes |
emMeans |
a formula containing the terms to estimate marginal means for (see the examples) |
emmPlots |
|
emmPlotData |
|
emmPlotError |
|
emmTables |
|
emmWeights |
|
ciWidthEmm |
a number between 50 and 99.9 (default: 95) specifying the confidence interval width for the estimated marginal means |
formula |
(optional) the formula to use, see the examples |
Value
A results object containing:
results$main | a table of ANCOVA results | ||||
results$model | The underlying aov object |
||||
results$assump$homo | a table of homogeneity tests | ||||
results$assump$norm | a table of normality tests | ||||
results$assump$qq | a q-q plot | ||||
results$contrasts | an array of contrasts tables | ||||
results$postHoc | an array of post-hoc tables | ||||
results$emm | an array of the estimated marginal means plots + tables | ||||
results$residsOV | an output | ||||
Tables can be converted to data frames with asDF or as.data.frame. For example:
results$main$asDF
as.data.frame(results$main)
Examples
data('ToothGrowth')
ancova(formula = len ~ supp + dose, data = ToothGrowth)
#
# ANCOVA
#
# ANCOVA
# -----------------------------------------------------------------------
# Sum of Squares df Mean Square F p
# -----------------------------------------------------------------------
# supp 205 1 205.4 11.4 0.001
# dose 2224 1 2224.3 124.0 < .001
# Residuals 1023 57 17.9
# -----------------------------------------------------------------------
#
ancova(
formula = len ~ supp + dose,
data = ToothGrowth,
postHoc = ~ supp,
emMeans = ~ supp)
One-Way ANOVA (Non-parametric)
Description
The Kruskal-Wallis test is used to explore the relationship between a continuous dependent variable, and a categorical explanatory variable. It is analagous to ANOVA, but with the advantage of being non-parametric and having fewer assumptions. However, it has the limitation that it can only test a single explanatory variable at a time.
Usage
anovaNP(data, deps, group, es = FALSE, pairs = FALSE,
pairsDunn = FALSE, formula)
Arguments
data |
the data as a data frame |
deps |
a string naming the dependent variable in |
group |
a string naming the grouping or independent variable in
|
es |
|
pairs |
|
pairsDunn |
|
formula |
(optional) the formula to use, see the examples |
Value
A results object containing:
results$table | a table of the test results | ||||
results$comparisons | an array of pairwise comparison tables | ||||
results$comparisonsDunn | an array of pairwise comparison tables | ||||
Tables can be converted to data frames with asDF or as.data.frame. For example:
results$table$asDF
as.data.frame(results$table)
Examples
data('ToothGrowth')
anovaNP(formula = len ~ dose, data=ToothGrowth)
#
# ONE-WAY ANOVA (NON-PARAMETRIC)
#
# Kruskal-Wallis
# -------------------------------
# X² df p
# -------------------------------
# len 40.7 2 < .001
# -------------------------------
#
One-Way ANOVA
Description
The Analysis of Variance (ANOVA) is used to explore the relationship between a continuous dependent variable, and one or more categorical explanatory variables. This 'One-Way ANOVA' is a simplified version of the 'normal' ANOVA, allowing only a single explanatory factor, however also providing a Welch's ANOVA. The Welch's ANOVA has the advantage that it need not assume that the variances of all groups are equal.
Usage
anovaOneW(data, deps, group, welchs = TRUE, fishers = FALSE,
miss = "perAnalysis", desc = FALSE, descPlot = FALSE,
norm = FALSE, qq = FALSE, eqv = FALSE, phMethod = "none",
phMeanDif = TRUE, phSig = TRUE, phTest = FALSE, phFlag = FALSE,
formula)
Arguments
data |
the data as a data frame |
deps |
a string naming the dependent variables in |
group |
a string naming the grouping or independent variable in
|
welchs |
|
fishers |
|
miss |
|
desc |
|
descPlot |
|
norm |
|
qq |
|
eqv |
|
phMethod |
|
phMeanDif |
|
phSig |
|
phTest |
|
phFlag |
|
formula |
(optional) the formula to use, see the examples |
Details
For convenience, this method allows specifying multiple dependent variables, resulting in multiple independent tests.
Note that the Welch's ANOVA is the same procedure as the Welch's independent samples t-test.
Value
A results object containing:
results$anova | a table of the test results | ||||
results$desc | a table containing the group descriptives | ||||
results$assump$norm | a table containing the normality tests | ||||
results$assump$eqv | a table of homogeneity of variances tests | ||||
results$plots | an array of groups of plots | ||||
results$postHoc | an array of post-hoc tables | ||||
Tables can be converted to data frames with asDF or as.data.frame. For example:
results$anova$asDF
as.data.frame(results$anova)
Examples
data('ToothGrowth')
dat <- ToothGrowth
dat$dose <- factor(dat$dose)
anovaOneW(formula = len ~ dose, data = dat)
#
# ONE-WAY ANOVA
#
# One-Way ANOVA (Welch's)
# ----------------------------------------
# F df1 df2 p
# ----------------------------------------
# len 68.4 2 37.7 < .001
# ----------------------------------------
#
Repeated Measures ANOVA
Description
The Repeated Measures ANOVA is used to explore the relationship between a continuous dependent variable and one or more categorical explanatory variables, where one or more of the explanatory variables are 'within subjects' (where multiple measurements are from the same subject). Additionally, this analysis allows the inclusion of covariates, allowing for repeated measures ANCOVAs as well.
Usage
anovaRM(data, rm = list(list(label = "RM Factor 1", levels =
list("Level 1", "Level 2"))), rmCells = NULL, bs = NULL,
cov = NULL, effectSize = NULL, depLabel = "Dependent",
rmTerms = NULL, bsTerms = NULL, ss = "3", spherTests = FALSE,
spherCorr = list("none"), leveneTest = FALSE, qq = FALSE,
contrasts = NULL, postHoc = NULL, postHocCorr = list("tukey"),
emMeans = list(list()), emmPlots = TRUE, emmTables = FALSE,
emmWeights = TRUE, ciWidthEmm = 95, emmPlotData = FALSE,
emmPlotError = "ci", groupSumm = FALSE)
Arguments
data |
the data as a data frame |
rm |
a list of lists, where each list describes the |
rmCells |
a list of lists, where each list decribes a repeated measure
(as a string) from |
bs |
a vector of strings naming the between subjects factors from
|
cov |
a vector of strings naming the covariates from |
effectSize |
one or more of |
depLabel |
a string (default: 'Dependent') describing the label used for the dependent variable throughout the analysis |
rmTerms |
a list of character vectors describing the repeated measures terms to go into the model |
bsTerms |
a list of character vectors describing the between subjects terms to go into the model |
ss |
|
spherTests |
|
spherCorr |
one or more of |
leveneTest |
|
qq |
|
contrasts |
in development |
postHoc |
a list of character vectors describing the post-hoc tests that need to be computed |
postHocCorr |
one or more of |
emMeans |
a list of lists specifying the variables for which the estimated marginal means need to be calculate. Supports up to three variables per term. |
emmPlots |
|
emmTables |
|
emmWeights |
|
ciWidthEmm |
a number between 50 and 99.9 (default: 95) specifying the confidence interval width for the estimated marginal means |
emmPlotData |
|
emmPlotError |
|
groupSumm |
|
Details
This analysis requires that the data be in 'wide format', where each row represents a subject (as opposed to long format, where each measurement of the dependent variable is represented as a row).
A non-parametric equivalent of the repeated measures ANOVA also exists; the Friedman test. However, it has the limitation of only being able to test a single factor.
Value
A results object containing:
results$rmTable | a table | ||||
results$bsTable | a table | ||||
results$assump$spherTable | a table | ||||
results$assump$leveneTable | a table | ||||
results$assump$qq | a q-q plot | ||||
results$contrasts | an array of tables | ||||
results$postHoc | an array of tables | ||||
results$emm | an array of the estimated marginal means plots + tables | ||||
results$groupSummary | a summary of the groups | ||||
Tables can be converted to data frames with asDF or as.data.frame. For example:
results$rmTable$asDF
as.data.frame(results$rmTable)
Examples
data('bugs', package = 'jmv')
anovaRM(
data = bugs,
rm = list(
list(
label = 'Frightening',
levels = c('Low', 'High'))),
rmCells = list(
list(
measure = 'LDLF',
cell = 'Low'),
list(
measure = 'LDHF',
cell = 'High')),
rmTerms = list(
'Frightening'))
#
# REPEATED MEASURES ANOVA
#
# Within Subjects Effects
# -----------------------------------------------------------------------
# Sum of Squares df Mean Square F p
# -----------------------------------------------------------------------
# Frightening 126 1 126.11 44.2 < .001
# Residual 257 90 2.85
# -----------------------------------------------------------------------
# Note. Type 3 Sums of Squares
#
#
#
# Between Subjects Effects
# -----------------------------------------------------------------
# Sum of Squares df Mean Square F p
# -----------------------------------------------------------------
# Residual 954 90 10.6
# -----------------------------------------------------------------
# Note. Type 3 Sums of Squares
#
Repeated Measures ANOVA (Non-parametric)
Description
The Friedman test is used to explore the relationship between a continuous dependent variable and a categorical explanatory variable, where the explanatory variable is 'within subjects' (where multiple measurements are from the same subject). It is analagous to Repeated Measures ANOVA, but with the advantage of being non-parametric, and not requiring the assumptions of normality or homogeneity of variances. However, it has the limitation that it can only test a single explanatory variable at a time.
Usage
anovaRMNP(data, measures, pairs = FALSE, desc = FALSE, plots = FALSE,
plotType = "means")
Arguments
data |
the data as a data frame |
measures |
a vector of strings naming the repeated measures variables |
pairs |
|
desc |
|
plots |
|
plotType |
|
Value
A results object containing:
results$table | a table of the Friedman test results | ||||
results$comp | a table of the pairwise comparisons | ||||
results$desc | a table containing the descriptives | ||||
results$plot | a descriptives plot | ||||
Tables can be converted to data frames with asDF or as.data.frame. For example:
results$table$asDF
as.data.frame(results$table)
Examples
data('bugs', package = 'jmv')
anovaRMNP(bugs, measures = vars(LDLF, LDHF, HDLF, HDHF))
#
# REPEATED MEASURES ANOVA (NON-PARAMETRIC)
#
# Friedman
# ------------------------
# X² df p
# ------------------------
# 55.8 3 < .001
# ------------------------
#
bugs
Description
bugs
Author(s)
Ryan, Wilde & Crist (2013)
References
Calculate the Greenhouse-Geisser correction for repeated measures ANOVA
Description
Calculate the Greenhouse-Geisser correction for repeated measures ANOVA
Usage
calcGG(SSPE, P)
Arguments
SSPE |
A matrix representing the sum of squares for the pure error |
P |
The design matrix for the effect |
Value
A numeric value representing the Greenhouse-Geisser correction factor
Calculate the Huynh-Feldt correction for repeated measures ANOVA
Description
Calculate the Huynh-Feldt correction for repeated measures ANOVA
Usage
calcHF(gg, error.df, p)
Arguments
gg |
The Greenhouse-Geisser correction |
error.df |
The degrees of freedom for the error term |
p |
The number of levels in the repeated measures factor |
Value
A numeric value representing the Huynh-Feldt correction
Perform Mauchly's Test of Sphericity
Description
Perform Mauchly's Test of Sphericity
Usage
calcMauchlyTest(SSD, P, df)
Arguments
SSD |
A matrix representing the Sum of Squares and Cross-Products for the Differences |
P |
The design matrix for the effect |
df |
Numeric value representing the degrees of freedom for error |
Value
A named numeric vector with the Mauchly's test statistic (W) and the p-value
Calculate Univariate Tests for ANOVA
Description
Calculate Univariate Tests for ANOVA
Usage
calcUnivariateTests(SSP, SSPE, P, df, error_df)
Arguments
SSP |
Sum of Squares and Products matrix for the term |
SSPE |
Sum of Squares and Products Error matrix |
P |
Design matrix for the term |
df |
Degrees of freedom for the term |
error_df |
Degrees of freedom for the error |
Value
A named list containing univariate test results
Confirmatory Factor Analysis
Description
Confirmatory Factor Analysis
Usage
cfa(data, factors = list(list(label = "Factor 1", vars = list())),
resCov, miss = "fiml", constrain = "facVar", estTest = TRUE,
ci = FALSE, ciWidth = 95, stdEst = FALSE, factCovEst = TRUE,
factInterceptEst = FALSE, resCovEst = FALSE,
resInterceptEst = FALSE, fitMeasures = list("cfi", "tli", "rmsea"),
modelTest = TRUE, pathDiagram = FALSE, corRes = FALSE,
hlCorRes = 0.1, mi = FALSE, hlMI = 3)
Arguments
data |
the data as a data frame |
factors |
a list containing named lists that define the |
resCov |
a list of lists specifying the residual covariances that need to be estimated |
miss |
|
constrain |
|
estTest |
|
ci |
|
ciWidth |
a number between 50 and 99.9 (default: 95) specifying the
confidence interval width that is used as |
stdEst |
|
factCovEst |
|
factInterceptEst |
|
resCovEst |
|
resInterceptEst |
|
fitMeasures |
one or more of |
modelTest |
|
pathDiagram |
|
corRes |
|
hlCorRes |
a number (default: 0.1), highlight values in the
|
mi |
|
hlMI |
a number (default: 3), highlight values in the
|
Value
A results object containing:
results$factorLoadings | a table containing the factor loadings | ||||
results$factorEst$factorCov | a table containing factor covariances estimates | ||||
results$factorEst$factorIntercept | a table containing factor intercept estimates | ||||
results$resEst$resCov | a table containing residual covariances estimates | ||||
results$resEst$resIntercept | a table containing residual intercept estimates | ||||
results$modelFit$test | a table containing the chi-square test for exact fit | ||||
results$modelFit$fitMeasures | a table containing fit measures | ||||
results$modelPerformance$corRes | a table containing residuals for the observed correlation matrix | ||||
results$modelPerformance$modIndices | a group | ||||
results$pathDiagram | an image containing the model path diagram | ||||
results$modelSyntax | the lavaan syntax used to fit the model | ||||
Tables can be converted to data frames with asDF or as.data.frame. For example:
results$factorLoadings$asDF
as.data.frame(results$factorLoadings)
Examples
data <- lavaan::HolzingerSwineford1939
jmv::cfa(
data = data,
factors = list(
list(label="Visual", vars=c("x1", "x2", "x3")),
list(label="Textual", vars=c("x4", "x5", "x6")),
list(label="Speed", vars=c("x7", "x8", "x9"))),
resCov = NULL)
#
# CONFIRMATORY FACTOR ANALYSIS
#
# Factor Loadings
# -----------------------------------------------------------------
# Factor Indicator Estimate SE Z p
# -----------------------------------------------------------------
# Visual x1 0.900 0.0832 10.81 < .001
# x2 0.498 0.0808 6.16 < .001
# x3 0.656 0.0776 8.46 < .001
# Textual x4 0.990 0.0567 17.46 < .001
# x5 1.102 0.0626 17.60 < .001
# x6 0.917 0.0538 17.05 < .001
# Speed x7 0.619 0.0743 8.34 < .001
# x8 0.731 0.0755 9.68 < .001
# x9 0.670 0.0775 8.64 < .001
# -----------------------------------------------------------------
#
#
# FACTOR ESTIMATES
#
# Factor Covariances
# --------------------------------------------------------------
# Estimate SE Z p
# --------------------------------------------------------------
# Visual Visual 1.000 a
# Textual 0.459 0.0635 7.22 < .001
# Speed 0.471 0.0862 5.46 < .001
# Textual Textual 1.000 a
# Speed 0.283 0.0715 3.96 < .001
# Speed Speed 1.000 a
# --------------------------------------------------------------
# a fixed parameter
#
#
# MODEL FIT
#
# Test for Exact Fit
# ------------------------
# X² df p
# ------------------------
# 85.3 24 < .001
# ------------------------
#
#
# Fit Measures
# -----------------------------------------------
# CFI TLI RMSEA Lower Upper
# -----------------------------------------------
# 0.931 0.896 0.0921 0.0714 0.114
# -----------------------------------------------
#
Contingency Tables
Description
The X² test of association (not to be confused with the X² goodness of fit) is used to test whether two categorical variables are independent or associated. If the p-value is low, it suggests the variables are not independent, and that there is a relationship between the two variables.
Usage
contTables(data, rows, cols, counts = NULL, layers = NULL,
chiSq = TRUE, chiSqCorr = FALSE, zProp = FALSE, likeRat = FALSE,
fisher = FALSE, contCoef = FALSE, phiCra = FALSE,
diffProp = FALSE, logOdds = FALSE, odds = FALSE, relRisk = FALSE,
ci = TRUE, ciWidth = 95, compare = "rows",
hypothesis = "different", gamma = FALSE, taub = FALSE,
mh = FALSE, obs = TRUE, exp = FALSE, pcRow = FALSE,
pcCol = FALSE, pcTot = FALSE, barplot = FALSE, yaxis = "ycounts",
yaxisPc = "total_pc", xaxis = "xrows", bartype = "dodge",
resU = FALSE, resP = FALSE, hlresP = 2, resS = FALSE,
hlresS = 2, resA = FALSE, hlresA = 2, formula)
Arguments
data |
the data as a data frame |
rows |
the variable to use as the rows in the contingency table (not necessary when providing a formula, see the examples) |
cols |
the variable to use as the columns in the contingency table (not necessary when providing a formula, see the examples) |
counts |
the variable to use as the counts in the contingency table (not necessary when providing a formula, see the examples) |
layers |
the variables to use to split the contingency table (not necessary when providing a formula, see the examples) |
chiSq |
|
chiSqCorr |
|
zProp |
|
likeRat |
|
fisher |
|
contCoef |
|
phiCra |
|
diffProp |
|
logOdds |
|
odds |
|
relRisk |
|
ci |
|
ciWidth |
a number between 50 and 99.9 (default: 95), width of the confidence intervals to provide |
compare |
|
hypothesis |
|
gamma |
|
taub |
|
mh |
|
obs |
|
exp |
|
pcRow |
|
pcCol |
|
pcTot |
|
barplot |
|
yaxis |
ycounts (default) or ypc. Use respectively |
yaxisPc |
total_pc (default), column_pc, or row_pc. Use respectively
percentages |
xaxis |
rows (default), or columns in bar plot X axis |
bartype |
stack or side by side (default), barplot type |
resU |
|
resP |
|
hlresP |
A numeric value (default: 2.0), highlight Pearson residuals above this threshold in the post hoc tests table. |
resS |
|
hlresS |
A numeric value (default: 2.0), highlight standardized residuals above this threshold in the post hoc tests table. |
resA |
|
hlresA |
A numeric value (default: 2.0), highlight deviance residuals above this threshold in the post hoc tests table. |
formula |
(optional) the formula to use, see the examples |
Value
A results object containing:
results$freqs | a table of proportions | ||||
results$chiSq | a table of X² test results | ||||
results$odds | a table of comparative measures | ||||
results$nom | a table of the 'nominal' test results | ||||
results$gamma | a table of the gamma test results | ||||
results$taub | a table of the Kendall's tau-b test results | ||||
results$mh | a table of the Mantel-Haenszel test for trend | ||||
results$postHoc | a table of post-hoc residuals | ||||
results$barplot | an image | ||||
Tables can be converted to data frames with asDF or as.data.frame. For example:
results$freqs$asDF
as.data.frame(results$freqs)
Examples
data('HairEyeColor')
dat <- as.data.frame(HairEyeColor)
contTables(formula = Freq ~ Hair:Eye, dat)
#
# CONTINGENCY TABLES
#
# Contingency Tables
# -----------------------------------------------------
# Hair Brown Blue Hazel Green Total
# -----------------------------------------------------
# Black 68 20 15 5 108
# Brown 119 84 54 29 286
# Red 26 17 14 14 71
# Blond 7 94 10 16 127
# Total 220 215 93 64 592
# -----------------------------------------------------
#
#
# X² Tests
# -------------------------------
# Value df p
# -------------------------------
# X² 138 9 < .001
# N 592
# -------------------------------
#
# Alternatively, omit the left of the formula (`Freq`) if each row
# represents a single observation:
contTables(formula = ~ Hair:Eye, dat)
Paired Samples Contingency Tables
Description
McNemar test
Usage
contTablesPaired(data, rows, cols, counts = NULL, chiSq = TRUE,
chiSqCorr = FALSE, exact = FALSE, pcRow = FALSE, pcCol = FALSE,
formula)
Arguments
data |
the data as a data frame |
rows |
the variable to use as the rows in the contingency table (not necessary when providing a formula, see the examples) |
cols |
the variable to use as the columns in the contingency table (not necessary when providing a formula, see the examples) |
counts |
the variable to use as the counts in the contingency table (not necessary when providing a formula, see the examples) |
chiSq |
|
chiSqCorr |
|
exact |
|
pcRow |
|
pcCol |
|
formula |
(optional) the formula to use, see the examples |
Value
A results object containing:
results$freqs | a proportions table | ||||
results$test | a table of test results | ||||
Tables can be converted to data frames with asDF or as.data.frame. For example:
results$freqs$asDF
as.data.frame(results$freqs)
Examples
dat <- data.frame(
`1st survey` = c('Approve', 'Approve', 'Disapprove', 'Disapprove'),
`2nd survey` = c('Approve', 'Disapprove', 'Approve', 'Disapprove'),
`Counts` = c(794, 150, 86, 570),
check.names=FALSE)
contTablesPaired(formula = Counts ~ `1st survey`:`2nd survey`, data = dat)
#
# PAIRED SAMPLES CONTINGENCY TABLES
#
# Contingency Tables
# ------------------------------------------------
# 1st survey Approve Disapprove Total
# ------------------------------------------------
# Approve 794 150 944
# Disapprove 86 570 656
# Total 880 720 1600
# ------------------------------------------------
#
#
# McNemar Test
# -----------------------------------------------------
# Value df p
# -----------------------------------------------------
# X² 17.4 1 < .001
# X² continuity correction 16.8 1 < .001
# -----------------------------------------------------
#
# Alternatively, omit the left of the formula (`Counts`) from the
# formula if each row represents a single observation:
contTablesPaired(formula = ~ `1st survey`:`2nd survey`, data = dat)
Correlation Matrix
Description
Correlation matrices are a way to examine linear relationships between two or more continuous variables.
Usage
corrMatrix(data, vars, pearson = TRUE, spearman = FALSE,
kendall = FALSE, sig = TRUE, flag = FALSE, n = FALSE,
ci = FALSE, ciWidth = 95, plots = FALSE, plotDens = FALSE,
plotStats = FALSE, hypothesis = "corr")
Arguments
data |
the data as a data frame |
vars |
a vector of strings naming the variables to correlate in
|
pearson |
|
spearman |
|
kendall |
|
sig |
|
flag |
|
n |
|
ci |
|
ciWidth |
a number between 50 and 99.9 (default: 95), the width of confidence intervals to provide |
plots |
|
plotDens |
|
plotStats |
|
hypothesis |
one of |
Details
For each pair of variables, a Pearson's r value indicates the strength and direction of the relationship between those two variables. A positive value indicates a positive relationship (higher values of one variable predict higher values of the other variable). A negative Pearson's r indicates a negative relationship (higher values of one variable predict lower values of the other variable, and vice-versa). A value of zero indicates no relationship (whether a variable is high or low, does not tell us anything about the value of the other variable).
More formally, it is possible to test the null hypothesis that the correlation is zero and calculate a p-value. If the p-value is low, it suggests the correlation co-efficient is not zero, and there is a linear (or more complex) relationship between the two variables.
Value
A results object containing:
results$matrix | a correlation matrix table | ||||
results$plot | a correlation matrix plot | ||||
Tables can be converted to data frames with asDF or as.data.frame. For example:
results$matrix$asDF
as.data.frame(results$matrix)
Examples
data('mtcars')
corrMatrix(mtcars, vars = vars(mpg, cyl, disp, hp))
#
# CORRELATION MATRIX
#
# Correlation Matrix
# --------------------------------------------------------------
# mpg cyl disp hp
# --------------------------------------------------------------
# mpg Pearson's r — -0.852 -0.848 -0.776
# p-value — < .001 < .001 < .001
#
# cyl Pearson's r — 0.902 0.832
# p-value — < .001 < .001
#
# disp Pearson's r — 0.791
# p-value — < .001
#
# hp Pearson's r —
# p-value —
# --------------------------------------------------------------
#
Partial Correlation
Description
Partial correlation matrices are a way to examine linear relationships between two or more continuous variables while controlling for other variables
Usage
corrPart(data, vars, controls, pearson = TRUE, spearman = FALSE,
kendall = FALSE, type = "part", sig = TRUE, flag = FALSE,
n = FALSE, hypothesis = "corr")
Arguments
data |
the data as a data frame |
vars |
a vector of strings naming the variables to correlate in
|
controls |
a vector of strings naming the control variables in
|
pearson |
|
spearman |
|
kendall |
|
type |
one of |
sig |
|
flag |
|
n |
|
hypothesis |
one of |
Details
For each pair of variables, a Pearson's r value indicates the strength and direction of the relationship between those two variables. A positive value indicates a positive relationship (higher values of one variable predict higher values of the other variable). A negative Pearson's r indicates a negative relationship (higher values of one variable predict lower values of the other variable, and vice-versa). A value of zero indicates no relationship (whether a variable is high or low, does not tell us anything about the value of the other variable).
More formally, it is possible to test the null hypothesis that the correlation is zero and calculate a p-value. If the p-value is low, it suggests the correlation co-efficient is not zero, and there is a linear (or more complex) relationship between the two variables.
Value
A results object containing:
results$matrix | a (semi)partial correlation matrix table | ||||
Tables can be converted to data frames with asDF or as.data.frame. For example:
results$matrix$asDF
as.data.frame(results$matrix)
Examples
data('mtcars')
corrPart(mtcars, vars = vars(mpg, cyl, disp), controls = vars(hp))
#
# PARTIAL CORRELATION
#
# Partial Correlation
# ----------------------------------------------------
# mpg cyl disp
# ----------------------------------------------------
# mpg Pearson's r —
# p-value —
#
# cyl Pearson's r -0.590 —
# p-value < .001 —
#
# disp Pearson's r -0.606 0.719 —
# p-value < .001 < .001 —
# ----------------------------------------------------
# Note. controlling for 'hp'
#
Descriptives
Description
Descriptives are an assortment of summarising statistics, and visualizations which allow exploring the shape and distribution of data. It is good practice to explore your data with descriptives before proceeding to more formal tests.
Usage
descriptives(data, vars, splitBy = NULL, freq = FALSE,
desc = "columns", hist = FALSE, dens = FALSE, bar = FALSE,
barCounts = FALSE, box = FALSE, violin = FALSE, dot = FALSE,
dotType = "jitter", boxMean = FALSE, boxLabelOutliers = TRUE,
qq = FALSE, n = TRUE, missing = TRUE, mean = TRUE,
median = TRUE, mode = FALSE, sum = FALSE, sd = TRUE,
variance = FALSE, range = FALSE, min = TRUE, max = TRUE,
se = FALSE, ci = FALSE, ciWidth = 95, iqr = FALSE,
skew = FALSE, kurt = FALSE, sw = FALSE, pcEqGr = FALSE,
pcNEqGr = 4, pc = FALSE, pcValues = "25,50,75", extreme = FALSE,
extremeN = 5, formula)
Arguments
data |
the data as a data frame |
vars |
a vector of strings naming the variables of interest in
|
splitBy |
a vector of strings naming the variables used to split
|
freq |
|
desc |
|
hist |
|
dens |
|
bar |
|
barCounts |
|
box |
|
violin |
|
dot |
|
dotType |
. |
boxMean |
|
boxLabelOutliers |
|
qq |
|
n |
|
missing |
|
mean |
|
median |
|
mode |
|
sum |
|
sd |
|
variance |
|
range |
|
min |
|
max |
|
se |
|
ci |
|
ciWidth |
a number between 50 and 99.9 (default: 95), the width of confidence intervals |
iqr |
|
skew |
|
kurt |
|
sw |
|
pcEqGr |
|
pcNEqGr |
an integer (default: 4) specifying the number of equal groups |
pc |
|
pcValues |
a comma-sepated list (default: 25,50,75) specifying the percentiles |
extreme |
|
extremeN |
an integer (default: 5) specifying the number of extreme values |
formula |
(optional) the formula to use, see the examples |
Value
A results object containing:
results$descriptives | a table of the descriptive statistics | ||||
results$descriptivesT | a table of the descriptive statistics | ||||
results$frequencies | an array of frequency tables | ||||
results$extremeValues | an array of extreme values tables | ||||
results$plots | an array of descriptive plots | ||||
Tables can be converted to data frames with asDF or as.data.frame. For example:
results$descriptives$asDF
as.data.frame(results$descriptives)
Examples
data('mtcars')
dat <- mtcars
# frequency tables can be provided for factors
dat$gear <- as.factor(dat$gear)
descriptives(dat, vars = vars(mpg, cyl, disp, gear), freq = TRUE)
#
# DESCRIPTIVES
#
# Descriptives
# -------------------------------------------
# mpg cyl disp gear
# -------------------------------------------
# N 32 32 32 32
# Missing 0 0 0 0
# Mean 20.1 6.19 231 3.69
# Median 19.2 6.00 196 4.00
# Minimum 10.4 4.00 71.1 3
# Maximum 33.9 8.00 472 5
# -------------------------------------------
#
#
# FREQUENCIES
#
# Frequencies of gear
# --------------------
# Levels Counts
# --------------------
# 3 15
# 4 12
# 5 5
# --------------------
#
# spliting by a variable
descriptives(formula = disp + mpg ~ cyl, dat,
median=FALSE, min=FALSE, max=FALSE, n=FALSE,
missing=FALSE)
# providing histograms
descriptives(formula = mpg ~ cyl, dat, hist=TRUE,
median=FALSE, min=FALSE, max=FALSE, n=FALSE,
missing=FALSE)
# splitting by multiple variables
descriptives(formula = mpg ~ cyl:gear, dat,
median=FALSE, min=FALSE, max=FALSE,
missing=FALSE)
Exploratory Factor Analysis
Description
Exploratory Factor Analysis
Usage
efa(data, vars, nFactorMethod = "parallel", nFactors = 1,
minEigen = 0, extraction = "minres", rotation = "oblimin",
hideLoadings = 0.3, sortLoadings = FALSE, screePlot = FALSE,
eigen = FALSE, factorCor = FALSE, factorSummary = FALSE,
modelFit = FALSE, kmo = FALSE, bartlett = FALSE,
factorScoreMethod = "Thurstone")
Arguments
data |
the data as a data frame |
vars |
a vector of strings naming the variables of interest in
|
nFactorMethod |
|
nFactors |
an integer (default: 1), the number of factors in the model |
minEigen |
a number (default: 0), the minimal eigenvalue for a factor to be included in the model |
extraction |
|
rotation |
|
hideLoadings |
a number (default: 0.3), hide factor loadings below this value |
sortLoadings |
|
screePlot |
|
eigen |
|
factorCor |
|
factorSummary |
|
modelFit |
|
kmo |
|
bartlett |
|
factorScoreMethod |
|
Value
A results object containing:
results$text | a preformatted | ||||
Examples
data('iris')
efa(iris, vars = vars(Sepal.Length, Sepal.Width, Petal.Length, Petal.Width))
#
# EXPLORATORY FACTOR ANALYSIS
#
# Factor Loadings
# ------------------------------------------------
# 1 2 Uniqueness
# ------------------------------------------------
# Sepal.Length 0.993 0.10181
# Sepal.Width 0.725 0.42199
# Petal.Length 0.933 0.00483
# Petal.Width 0.897 0.07088
# ------------------------------------------------
# Note. 'oblimin' rotation was used
#
Update reference levels for a set of variables
Description
Update reference levels for a set of variables
Usage
getReferenceLevels(data, vars, refLevels)
Arguments
data |
The data frame containing the variables |
vars |
The names of the variables to update |
refLevels |
A list of reference levels to use for each variable |
Value
A list of updated reference levels and a list of variables that had their reference levels changed
iris
Description
iris
Linear Regression
Description
Linear regression is used to explore the relationship between a continuous dependent variable, and one or more continuous and/or categorical explanatory variables. Other statistical methods, such as ANOVA and ANCOVA, are in reality just forms of linear regression.
Usage
linReg(data, dep, covs = NULL, factors = NULL, weights = NULL,
blocks = list(list()), refLevels = NULL, intercept = "refLevel",
r = TRUE, r2 = TRUE, r2Adj = FALSE, aic = FALSE, bic = FALSE,
rmse = FALSE, modelTest = FALSE, anova = FALSE, ci = FALSE,
ciWidth = 95, stdEst = FALSE, ciStdEst = FALSE,
ciWidthStdEst = 95, norm = FALSE, qqPlot = FALSE,
resPlots = FALSE, durbin = FALSE, collin = FALSE, cooks = FALSE,
mahal = FALSE, mahalp = "0.001", emMeans = list(list()),
ciEmm = TRUE, ciWidthEmm = 95, emmPlots = TRUE,
emmTables = FALSE, emmWeights = TRUE)
Arguments
data |
the data as a data frame |
dep |
the dependent variable from |
covs |
the covariates from |
factors |
the fixed factors from |
weights |
the (optional) weights from |
blocks |
a list containing vectors of strings that name the predictors that are added to the model. The elements are added to the model according to their order in the list |
refLevels |
a list of lists specifying reference levels of the dependent variable and all the factors |
intercept |
|
r |
|
r2 |
|
r2Adj |
|
aic |
|
bic |
|
rmse |
|
modelTest |
|
anova |
|
ci |
|
ciWidth |
a number between 50 and 99.9 (default: 95) specifying the confidence interval width |
stdEst |
|
ciStdEst |
|
ciWidthStdEst |
a number between 50 and 99.9 (default: 95) specifying the confidence interval width |
norm |
|
qqPlot |
|
resPlots |
|
durbin |
|
collin |
|
cooks |
|
mahal |
|
mahalp |
|
emMeans |
a formula containing the terms to estimate marginal means for, supports up to three variables per term |
ciEmm |
|
ciWidthEmm |
a number between 50 and 99.9 (default: 95) specifying the confidence interval width for the estimated marginal means |
emmPlots |
|
emmTables |
|
emmWeights |
|
Value
A results object containing:
results$modelFit | a table | ||||
results$modelComp | a table | ||||
results$models | an array of model specific results | ||||
results$predictOV | an output | ||||
results$residsOV | an output | ||||
results$cooksOV | an output | ||||
results$mahalOV | an output | ||||
Tables can be converted to data frames with asDF or as.data.frame. For example:
results$modelFit$asDF
as.data.frame(results$modelFit)
Examples
data('Prestige', package='carData')
linReg(data = Prestige, dep = income,
covs = vars(education, prestige, women),
blocks = list(list('education', 'prestige', 'women')))
#
# LINEAR REGRESSION
#
# Model Fit Measures
# ---------------------------
# Model R R²
# ---------------------------
# 1 0.802 0.643
# ---------------------------
#
#
# MODEL SPECIFIC RESULTS
#
# MODEL 1
#
#
# Model Coefficients
# --------------------------------------------------------
# Predictor Estimate SE t p
# --------------------------------------------------------
# Intercept -253.8 1086.16 -0.234 0.816
# women -50.9 8.56 -5.948 < .001
# prestige 141.4 29.91 4.729 < .001
# education 177.2 187.63 0.944 0.347
# --------------------------------------------------------
#
Log-Linear Regression
Description
Log-Linear Regression
Usage
logLinear(data, factors = NULL, counts = NULL, blocks = list(list()),
refLevels = NULL, modelTest = FALSE, dev = TRUE, aic = TRUE,
bic = FALSE, pseudoR2 = list("r2mf"), omni = FALSE, ci = FALSE,
ciWidth = 95, RR = FALSE, ciRR = FALSE, ciWidthRR = 95,
emMeans = list(list()), ciEmm = TRUE, ciWidthEmm = 95,
emmPlots = TRUE, emmTables = FALSE, emmWeights = TRUE)
Arguments
data |
the data as a data frame |
factors |
a vector of strings naming the factors from |
counts |
a string naming a variable in |
blocks |
a list containing vectors of strings that name the predictors that are added to the model. The elements are added to the model according to their order in the list |
refLevels |
a list of lists specifying reference levels of the dependent variable and all the factors |
modelTest |
|
dev |
|
aic |
|
bic |
|
pseudoR2 |
one or more of |
omni |
|
ci |
|
ciWidth |
a number between 50 and 99.9 (default: 95) specifying the confidence interval width |
RR |
|
ciRR |
|
ciWidthRR |
a number between 50 and 99.9 (default: 95) specifying the confidence interval width |
emMeans |
a list of lists specifying the variables for which the estimated marginal means need to be calculate. Supports up to three variables per term. |
ciEmm |
|
ciWidthEmm |
a number between 50 and 99.9 (default: 95) specifying the confidence interval width for the estimated marginal means |
emmPlots |
|
emmTables |
|
emmWeights |
|
Value
A results object containing:
results$modelFit | a table | ||||
results$modelComp | a table | ||||
results$models | an array of model specific results | ||||
Tables can be converted to data frames with asDF or as.data.frame. For example:
results$modelFit$asDF
as.data.frame(results$modelFit)
Examples
data('mtcars')
tab <- table('gear'=mtcars$gear, 'cyl'=mtcars$cyl)
dat <- as.data.frame(tab)
logLinear(data = dat, factors = vars(gear, cyl), counts = Freq,
blocks = list(list("gear", "cyl", c("gear", "cyl"))),
refLevels = list(
list(var="gear", ref="3"),
list(var="cyl", ref="4")))
#
# LOG-LINEAR REGRESSION
#
# Model Fit Measures
# ---------------------------------------
# Model Deviance AIC R²-McF
# ---------------------------------------
# 1 4.12e-10 41.4 1.000
# ---------------------------------------
#
#
# MODEL SPECIFIC RESULTS
#
# MODEL 1
#
# Model Coefficients
# ------------------------------------------------------------------
# Predictor Estimate SE Z p
# ------------------------------------------------------------------
# Intercept -4.71e-16 1.00 -4.71e-16 1.000
# gear:
# 4 – 3 2.079 1.06 1.961 0.050
# 5 – 3 0.693 1.22 0.566 0.571
# cyl:
# 6 – 4 0.693 1.22 0.566 0.571
# 8 – 4 2.485 1.04 2.387 0.017
# gear:cyl:
# (4 – 3):(6 – 4) -1.386 1.37 -1.012 0.311
# (5 – 3):(6 – 4) -1.386 1.73 -0.800 0.423
# (4 – 3):(8 – 4) -26.867 42247.17 -6.36e -4 0.999
# (5 – 3):(8 – 4) -2.485 1.44 -1.722 0.085
# ------------------------------------------------------------------
#
#
Binomial Logistic Regression
Description
Binomial Logistic Regression
Usage
logRegBin(data, dep, covs = NULL, factors = NULL,
blocks = list(list()), refLevels = NULL, modelTest = FALSE,
dev = TRUE, aic = TRUE, bic = FALSE, pseudoR2 = list("r2mf"),
omni = FALSE, ci = FALSE, ciWidth = 95, OR = FALSE,
ciOR = FALSE, ciWidthOR = 95, emMeans = list(list()),
ciEmm = TRUE, ciWidthEmm = 95, emmPlots = TRUE,
emmTables = FALSE, emmWeights = TRUE, class = FALSE, acc = FALSE,
spec = FALSE, sens = FALSE, auc = FALSE, rocPlot = FALSE,
cutOff = 0.5, cutOffPlot = FALSE, collin = FALSE,
boxTidwell = FALSE, cooks = FALSE)
Arguments
data |
the data as a data frame |
dep |
a string naming the dependent variable from |
covs |
a vector of strings naming the covariates from |
factors |
a vector of strings naming the fixed factors from
|
blocks |
a list containing vectors of strings that name the predictors that are added to the model. The elements are added to the model according to their order in the list |
refLevels |
a list of lists specifying reference levels of the dependent variable and all the factors |
modelTest |
|
dev |
|
aic |
|
bic |
|
pseudoR2 |
one or more of |
omni |
|
ci |
|
ciWidth |
a number between 50 and 99.9 (default: 95) specifying the confidence interval width |
OR |
|
ciOR |
|
ciWidthOR |
a number between 50 and 99.9 (default: 95) specifying the confidence interval width |
emMeans |
a list of lists specifying the variables for which the estimated marginal means need to be calculate. Supports up to three variables per term. |
ciEmm |
|
ciWidthEmm |
a number between 50 and 99.9 (default: 95) specifying the confidence interval width for the estimated marginal means |
emmPlots |
|
emmTables |
|
emmWeights |
|
class |
|
acc |
|
spec |
|
sens |
|
auc |
|
rocPlot |
|
cutOff |
|
cutOffPlot |
|
collin |
|
boxTidwell |
|
cooks |
|
Value
A results object containing:
results$modelFit | a table | ||||
results$modelComp | a table | ||||
results$models | an array of model specific results | ||||
results$predictOV | an output | ||||
results$residsOV | an output | ||||
results$cooksOV | an output | ||||
Tables can be converted to data frames with asDF or as.data.frame. For example:
results$modelFit$asDF
as.data.frame(results$modelFit)
Examples
data('birthwt', package='MASS')
dat <- data.frame(
low = factor(birthwt$low),
age = birthwt$age,
bwt = birthwt$bwt)
logRegBin(data = dat, dep = low,
covs = vars(age, bwt),
blocks = list(list("age", "bwt")),
refLevels = list(list(var="low", ref="0")))
#
# BINOMIAL LOGISTIC REGRESSION
#
# Model Fit Measures
# ---------------------------------------
# Model Deviance AIC R²-McF
# ---------------------------------------
# 1 4.97e-7 6.00 1.000
# ---------------------------------------
#
#
# MODEL SPECIFIC RESULTS
#
# MODEL 1
#
# Model Coefficients
# ------------------------------------------------------------
# Predictor Estimate SE Z p
# ------------------------------------------------------------
# Intercept 2974.73225 218237.2 0.0136 0.989
# age -0.00653 482.7 -1.35e-5 1.000
# bwt -1.18532 87.0 -0.0136 0.989
# ------------------------------------------------------------
# Note. Estimates represent the log odds of "low = 1"
# vs. "low = 0"
#
#
Multinomial Logistic Regression
Description
Multinomial Logistic Regression
Usage
logRegMulti(data, dep, covs = NULL, factors = NULL,
blocks = list(list()), refLevels = NULL, modelTest = FALSE,
dev = TRUE, aic = TRUE, bic = FALSE, pseudoR2 = list("r2mf"),
omni = FALSE, ci = FALSE, ciWidth = 95, OR = FALSE,
ciOR = FALSE, ciWidthOR = 95, emMeans = list(list()),
ciEmm = TRUE, ciWidthEmm = 95, emmPlots = TRUE,
emmTables = FALSE, emmWeights = TRUE)
Arguments
data |
the data as a data frame |
dep |
a string naming the dependent variable from |
covs |
a vector of strings naming the covariates from |
factors |
a vector of strings naming the fixed factors from
|
blocks |
a list containing vectors of strings that name the predictors that are added to the model. The elements are added to the model according to their order in the list |
refLevels |
a list of lists specifying reference levels of the dependent variable and all the factors |
modelTest |
|
dev |
|
aic |
|
bic |
|
pseudoR2 |
one or more of |
omni |
|
ci |
|
ciWidth |
a number between 50 and 99.9 (default: 95) specifying the confidence interval width |
OR |
|
ciOR |
|
ciWidthOR |
a number between 50 and 99.9 (default: 95) specifying the confidence interval width |
emMeans |
a list of lists specifying the variables for which the estimated marginal means need to be calculate. Supports up to three variables per term. |
ciEmm |
|
ciWidthEmm |
a number between 50 and 99.9 (default: 95) specifying the confidence interval width for the estimated marginal means |
emmPlots |
|
emmTables |
|
emmWeights |
|
Value
A results object containing:
results$modelFit | a table | ||||
results$modelComp | a table | ||||
results$models | an array of model specific results | ||||
Tables can be converted to data frames with asDF or as.data.frame. For example:
results$modelFit$asDF
as.data.frame(results$modelFit)
Examples
data('birthwt', package='MASS')
dat <- data.frame(
race = factor(birthwt$race),
age = birthwt$age,
low = factor(birthwt$low))
logRegMulti(data = dat, dep = race,
covs = age, factors = low,
blocks = list(list("age", "low")),
refLevels = list(
list(var="race", ref="1"),
list(var="low", ref="0")))
#
# MULTINOMIAL LOGISTIC REGRESSION
#
# Model Fit Measures
# --------------------------------------
# Model Deviance AIC R²-McF
# --------------------------------------
# 1 360 372 0.0333
# --------------------------------------
#
#
# MODEL SPECIFIC RESULTS
#
# MODEL 1
#
# Model Coefficients
# ---------------------------------------------------------------
# race Predictor Estimate SE Z p
# ---------------------------------------------------------------
# 2 - 1 Intercept 0.8155 1.1186 0.729 0.466
# age -0.1038 0.0487 -2.131 0.033
# low:
# 1 – 0 0.7527 0.4700 1.601 0.109
# 3 - 1 Intercept 1.0123 0.7798 1.298 0.194
# age -0.0663 0.0324 -2.047 0.041
# low:
# 1 – 0 0.5677 0.3522 1.612 0.107
# ---------------------------------------------------------------
#
#
Ordinal Logistic Regression
Description
Ordinal Logistic Regression
Usage
logRegOrd(data, dep, covs = NULL, factors = NULL,
blocks = list(list()), refLevels = NULL, modelTest = FALSE,
dev = TRUE, aic = TRUE, bic = FALSE, pseudoR2 = list("r2mf"),
omni = FALSE, thres = FALSE, ci = FALSE, ciWidth = 95,
OR = FALSE, ciOR = FALSE, ciWidthOR = 95)
Arguments
data |
the data as a data frame |
dep |
a string naming the dependent variable from |
covs |
a vector of strings naming the covariates from |
factors |
a vector of strings naming the fixed factors from
|
blocks |
a list containing vectors of strings that name the predictors that are added to the model. The elements are added to the model according to their order in the list |
refLevels |
a list of lists specifying reference levels of the dependent variable and all the factors |
modelTest |
|
dev |
|
aic |
|
bic |
|
pseudoR2 |
one or more of |
omni |
|
thres |
|
ci |
|
ciWidth |
a number between 50 and 99.9 (default: 95) specifying the confidence interval width |
OR |
|
ciOR |
|
ciWidthOR |
a number between 50 and 99.9 (default: 95) specifying the confidence interval width |
Value
A results object containing:
results$modelFit | a table | ||||
results$modelComp | a table | ||||
results$models | an array of model specific results | ||||
Tables can be converted to data frames with asDF or as.data.frame. For example:
results$modelFit$asDF
as.data.frame(results$modelFit)
Examples
set.seed(1337)
y <- factor(sample(1:3, 100, replace = TRUE))
x1 <- rnorm(100)
x2 <- rnorm(100)
df <- data.frame(y=y, x1=x1, x2=x2)
logRegOrd(data = df, dep = y,
covs = vars(x1, x2),
blocks = list(list("x1", "x2")))
#
# ORDINAL LOGISTIC REGRESSION
#
# Model Fit Measures
# ---------------------------------------
# Model Deviance AIC R²-McF
# ---------------------------------------
# 1 218 226 5.68e-4
# ---------------------------------------
#
#
# MODEL SPECIFIC RESULTS
#
# MODEL 1
#
# Model Coefficients
# ----------------------------------------------------
# Predictor Estimate SE Z p
# ----------------------------------------------------
# x1 0.0579 0.193 0.300 0.764
# x2 0.0330 0.172 0.192 0.848
# ----------------------------------------------------
#
#
MANCOVA
Description
Multivariate Analysis of (Co)Variance (MANCOVA) is used to explore the relationship between multiple dependent variables, and one or more categorical and/or continuous explanatory variables.
Usage
mancova(data, deps, factors = NULL, covs = NULL,
multivar = list("pillai", "wilks", "hotel", "roy"), boxM = FALSE,
shapiro = FALSE, qqPlot = FALSE)
Arguments
data |
the data as a data frame |
deps |
a string naming the dependent variable from |
factors |
a vector of strings naming the factors from |
covs |
a vector of strings naming the covariates from |
multivar |
one or more of |
boxM |
|
shapiro |
|
qqPlot |
|
Value
A results object containing:
results$multivar | a table | ||||
results$univar | a table | ||||
results$assump$boxM | a table | ||||
results$assump$shapiro | a table | ||||
results$assump$qqPlot | an image | ||||
Tables can be converted to data frames with asDF or as.data.frame. For example:
results$multivar$asDF
as.data.frame(results$multivar)
Examples
data('iris')
mancova(data = iris,
deps = vars(Sepal.Length, Sepal.Width, Petal.Length, Petal.Width),
factors = Species)
#
# MANCOVA
#
# Multivariate Tests
# ---------------------------------------------------------------------------
# value F df1 df2 p
# ---------------------------------------------------------------------------
# Species Pillai's Trace 1.19 53.5 8 290 < .001
# Wilks' Lambda 0.0234 199 8 288 < .001
# Hotelling's Trace 32.5 581 8 286 < .001
# Roy's Largest Root 32.2 1167 4 145 < .001
# ---------------------------------------------------------------------------
#
#
# Univariate Tests
# -----------------------------------------------------------------------------------------------
# Dependent Variable Sum of Squares df Mean Square F p
# -----------------------------------------------------------------------------------------------
# Species Sepal.Length 63.21 2 31.6061 119.3 < .001
# Sepal.Width 11.34 2 5.6725 49.2 < .001
# Petal.Length 437.10 2 218.5514 1180.2 < .001
# Petal.Width 80.41 2 40.2067 960.0 < .001
# Residuals Sepal.Length 38.96 147 0.2650
# Sepal.Width 16.96 147 0.1154
# Petal.Length 27.22 147 0.1852
# Petal.Width 6.16 147 0.0419
# -----------------------------------------------------------------------------------------------
#
Principal Component Analysis
Description
Principal Component Analysis
Usage
pca(data, vars, nFactorMethod = "parallel", nFactors = 1,
minEigen = 1, rotation = "varimax", hideLoadings = 0.3,
sortLoadings = FALSE, screePlot = FALSE, eigen = FALSE,
factorCor = FALSE, factorSummary = FALSE, kmo = FALSE,
bartlett = FALSE)
Arguments
data |
the data as a data frame |
vars |
a vector of strings naming the variables of interest in
|
nFactorMethod |
|
nFactors |
an integer (default: 1), the number of components in the model |
minEigen |
a number (default: 1), the minimal eigenvalue for a component to be included in the model |
rotation |
|
hideLoadings |
a number (default: 0.3), hide loadings below this value |
sortLoadings |
|
screePlot |
|
eigen |
|
factorCor |
|
factorSummary |
|
kmo |
|
bartlett |
|
Value
A results object containing:
results$loadings | a table | ||||
results$factorStats$factorSummary | a table | ||||
results$factorStats$factorCor | a table | ||||
results$modelFit$fit | a table | ||||
results$assump$bartlett | a table | ||||
results$assump$kmo | a table | ||||
results$eigen$initEigen | a table | ||||
results$eigen$screePlot | an image | ||||
results$factorScoresOV | an output | ||||
Tables can be converted to data frames with asDF or as.data.frame. For example:
results$loadings$asDF
as.data.frame(results$loadings)
Examples
data('iris')
pca(iris, vars = vars(Sepal.Length, Sepal.Width, Petal.Length, Petal.Width))
#
# PRINCIPAL COMPONENT ANALYSIS
#
# Component Loadings
# ----------------------------------------
# 1 Uniqueness
# ----------------------------------------
# Sepal.Length 0.890 0.2076
# Sepal.Width -0.460 0.7883
# Petal.Length 0.992 0.0168
# Petal.Width 0.965 0.0688
# ----------------------------------------
# Note. 'varimax' rotation was used
#
Proportion Test (2 Outcomes)
Description
The Binomial test is used to test the Null hypothesis that the proportion of observations match some expected value. If the p-value is low, this suggests that the Null hypothesis is false, and that the true proportion must be some other value.
Usage
propTest2(data, vars, areCounts = FALSE, testValue = 0.5,
hypothesis = "notequal", ci = FALSE, ciWidth = 95, bf = FALSE,
priorA = 1, priorB = 1, ciBayes = FALSE, ciBayesWidth = 95,
postPlots = FALSE)
Arguments
data |
the data as a data frame |
vars |
a vector of strings naming the variables of interest in
|
areCounts |
|
testValue |
a number (default: 0.5), the value for the null hypothesis |
hypothesis |
|
ci |
|
ciWidth |
a number between 50 and 99.9 (default: 95), the confidence interval width |
bf |
|
priorA |
a number (default: 1), the beta prior 'a' parameter |
priorB |
a number (default: 1), the beta prior 'b' parameter |
ciBayes |
|
ciBayesWidth |
a number between 50 and 99.9 (default: 95), the credible interval width |
postPlots |
|
Value
A results object containing:
results$table | a table of the proportions and test results | ||||
results$postPlots | an array of the posterior plots | ||||
Tables can be converted to data frames with asDF or as.data.frame. For example:
results$table$asDF
as.data.frame(results$table)
Examples
dat <- data.frame(x=c(8, 15))
propTest2(dat, vars = x, areCounts = TRUE)
#
# PROPORTION TEST (2 OUTCOMES)
#
# Binomial Test
# -------------------------------------------------------
# Level Count Total Proportion p
# -------------------------------------------------------
# x 1 8 23 0.348 0.210
# 2 15 23 0.652 0.210
# -------------------------------------------------------
# Note. Ha is proportion != 0.5
#
Proportion Test (N Outcomes)
Description
The X² Goodness of fit test (not to be confused with the X² test of independence), tests the Null hypothesis that the proportions of observations match some expected proportions. If the p-value is low, this suggests that the Null hypothesis is false, and that the true proportions are different to those tested.
Usage
propTestN(data, var, counts = NULL, expected = FALSE, ratio = NULL,
formula)
Arguments
data |
the data as a data frame |
var |
the variable of interest in |
counts |
the counts in |
expected |
|
ratio |
a vector of numbers: the expected proportions |
formula |
(optional) the formula to use, see the examples |
Value
A results object containing:
results$props | a table of the proportions | ||||
results$tests | a table of the test results | ||||
Tables can be converted to data frames with asDF or as.data.frame. For example:
results$props$asDF
as.data.frame(results$props)
Examples
data('HairEyeColor')
dat <- as.data.frame(HairEyeColor)
propTestN(formula = Freq ~ Eye, data = dat, ratio = c(1,1,1,1))
#
# PROPORTION TEST (N OUTCOMES)
#
# Proportions
# --------------------------------
# Level Count Proportion
# --------------------------------
# Brown 220 0.372
# Blue 215 0.363
# Hazel 93 0.157
# Green 64 0.108
# --------------------------------
#
#
# X² Goodness of Fit
# -----------------------
# X² df p
# -----------------------
# 133 3 < .001
# -----------------------
#
Reliability Analysis
Description
Reliability Analysis
Usage
reliability(data, vars, alphaScale = TRUE, omegaScale = FALSE,
meanScale = FALSE, sdScale = FALSE, corPlot = FALSE,
alphaItems = FALSE, omegaItems = FALSE, meanItems = FALSE,
sdItems = FALSE, itemRestCor = FALSE, revItems = NULL)
Arguments
data |
the data as a data frame |
vars |
a vector of strings naming the variables of interest in
|
alphaScale |
|
omegaScale |
|
meanScale |
|
sdScale |
|
corPlot |
|
alphaItems |
|
omegaItems |
|
meanItems |
|
sdItems |
|
itemRestCor |
|
revItems |
a vector containing strings naming the varibales that are reverse scaled |
Value
A results object containing:
results$scale | a table | ||||
results$items | a table | ||||
results$corPlot | an image | ||||
results$meanScoreOV | an output | ||||
results$sumScoreOV | an output | ||||
Tables can be converted to data frames with asDF or as.data.frame. For example:
results$scale$asDF
as.data.frame(results$scale)
Examples
data('iris')
reliability(iris, vars = c('Sepal.Length', 'Sepal.Width', 'Petal.Length', 'Petal.Width'),
omegaScale = TRUE)
#
# RELIABILITY ANALYSIS
#
# Scale Reliability Statistics
# -----------------------------------------
# Cronbach's alpha McDonald's omega
# -----------------------------------------
# scale 0.708 0.848
# -----------------------------------------
#
Set a global analysis notice
Description
Set a global analysis notice
Usage
setAnalysisNotice(self, message, name, type)
Arguments
self |
The analysis object |
message |
The message to set |
name |
The name of the notice |
type |
The type of the notice |
Set a warning notice for reference level changes
Description
Set a warning notice for reference level changes
Usage
setRefLevelWarning(self, changedVars)
Arguments
self |
The analysis object |
changedVars |
The variables that had their reference levels changed |
Set a warning notice for singularity in the design matrix
Description
Set a warning notice for singularity in the design matrix
Usage
setSingularityWarning(self)
Arguments
self |
The analysis object |
Summarize an Anova object from a repeated measures model into a single table
Description
Summarize an Anova object from a repeated measures model into a single table
Usage
summarizeAnovaModel(object, aov_table)
Arguments
object |
An Anova object from a repeated measures model |
aov_table |
The ANOVA table from the model containing the generalized eta squared |
Value
A data frame containing all the relevant ANOVA statistics
Independent Samples T-Test
Description
The Student's Independent samples t-test (sometimes called a two-samples t-test) is used to test the null hypothesis that two groups have the same mean. A low p-value suggests that the null hypothesis is not true, and therefore the group means are different.
Usage
ttestIS(data, vars, group, students = TRUE, bf = FALSE,
bfPrior = 0.707, welchs = FALSE, mann = FALSE,
hypothesis = "different", norm = FALSE, qq = FALSE, eqv = FALSE,
meanDiff = FALSE, ci = FALSE, ciWidth = 95, effectSize = FALSE,
ciES = FALSE, ciWidthES = 95, desc = FALSE, plots = FALSE,
miss = "perAnalysis", formula)
Arguments
data |
the data as a data frame |
vars |
the dependent variables (not necessary when using a formula, see the examples) |
group |
the grouping variable with two levels (not necessary when using a formula, see the examples) |
students |
|
bf |
|
bfPrior |
a number between 0.01 and 2 (default 0.707), the prior width to use in calculating Bayes factors |
welchs |
|
mann |
|
hypothesis |
|
norm |
|
qq |
|
eqv |
|
meanDiff |
|
ci |
|
ciWidth |
a number between 50 and 99.9 (default: 95), the width of confidence intervals |
effectSize |
|
ciES |
|
ciWidthES |
a number between 50 and 99.9 (default: 95), the width of confidence intervals for the effect sizes |
desc |
|
plots |
|
miss |
|
formula |
(optional) the formula to use, see the examples |
Details
The Student's independent t-test assumes that the data from each group are from a normal distribution, and that the variances of these groups are equal. If unwilling to assume the groups have equal variances, the Welch's t-test can be used in it's place. If one is additionally unwilling to assume the data from each group are from a normal distribution, the non-parametric Mann-Whitney U test can be used instead (However, note that the Mann-Whitney U test has a slightly different null hypothesis; that the distributions of each group is equal).
Value
A results object containing:
results$ttest | a table containing the t-test results | ||||
results$assum$norm | a table containing the normality tests | ||||
results$assum$eqv | a table containing the homogeneity of variances tests | ||||
results$desc | a table containing the group descriptives | ||||
results$plots | an array of groups of plots | ||||
Tables can be converted to data frames with asDF or as.data.frame. For example:
results$ttest$asDF
as.data.frame(results$ttest)
Examples
data('ToothGrowth')
ttestIS(formula = len ~ supp, data = ToothGrowth)
#
# INDEPENDENT SAMPLES T-TEST
#
# Independent Samples T-Test
# ----------------------------------------------------
# statistic df p
# ----------------------------------------------------
# len Student's t 1.92 58.0 0.060
# ----------------------------------------------------
#
One Sample T-Test
Description
The Student's One-sample t-test is used to test the null hypothesis that the true mean is equal to a particular value (typically zero). A low p-value suggests that the null hypothesis is not true, and therefore the true mean must be different from the test value.
Usage
ttestOneS(data, vars, students = TRUE, bf = FALSE, bfPrior = 0.707,
wilcoxon = FALSE, testValue = 0, hypothesis = "dt", norm = FALSE,
qq = FALSE, meanDiff = FALSE, ci = FALSE, ciWidth = 95,
effectSize = FALSE, ciES = FALSE, ciWidthES = 95, desc = FALSE,
plots = FALSE, miss = "perAnalysis", mann = FALSE)
Arguments
data |
the data as a data frame |
vars |
a vector of strings naming the variables of interest in
|
students |
|
bf |
|
bfPrior |
a number between 0.5 and 2.0 (default 0.707), the prior width to use in calculating Bayes factors |
wilcoxon |
|
testValue |
a number specifying the value of the null hypothesis |
hypothesis |
|
norm |
|
qq |
|
meanDiff |
|
ci |
|
ciWidth |
a number between 50 and 99.9 (default: 95), the width of confidence intervals |
effectSize |
|
ciES |
|
ciWidthES |
a number between 50 and 99.9 (default: 95), the width of confidence intervals for the effect sizes |
desc |
|
plots |
|
miss |
|
mann |
deprecated |
Details
The Student's One-sample t-test assumes that the data are from a normal distribution – in the case that one is unwilling to assume this, the non-parametric Wilcoxon signed-rank can be used in it's place (However, note that the Wilcoxon signed-rank has a slightly different null hypothesis; that the *median* is equal to the test value).
Value
A results object containing:
results$ttest | a table containing the t-test results | ||||
results$normality | a table containing the normality test results | ||||
results$descriptives | a table containing the descriptives | ||||
results$plots | an image of the descriptive plots | ||||
results$qq | an array of Q-Q plots | ||||
Tables can be converted to data frames with asDF or as.data.frame. For example:
results$ttest$asDF
as.data.frame(results$ttest)
Examples
data('ToothGrowth')
ttestOneS(ToothGrowth, vars = vars(len, dose))
#
# ONE SAMPLE T-TEST
#
# One Sample T-Test
# ------------------------------------------------------
# statistic df p
# ------------------------------------------------------
# len Student's t 19.1 59.0 < .001
# dose Student's t 14.4 59.0 < .001
# ------------------------------------------------------
#
Paired Samples T-Test
Description
The Student's paired samples t-test (sometimes called a dependent-samples t-test) is used to test the null hypothesis that the difference between pairs of measurements is equal to zero. A low p-value suggests that the null hypothesis is not true, and that the difference between the measurement pairs is not zero.
Usage
ttestPS(data, pairs, students = TRUE, bf = FALSE, bfPrior = 0.707,
wilcoxon = FALSE, hypothesis = "different", norm = FALSE,
qq = FALSE, meanDiff = FALSE, ci = FALSE, ciWidth = 95,
effectSize = FALSE, ciES = FALSE, ciWidthES = 95, desc = FALSE,
plots = FALSE, miss = "perAnalysis")
Arguments
data |
the data as a data frame |
pairs |
a list of lists specifying the pairs of measurement in
|
students |
|
bf |
|
bfPrior |
a number between 0.5 and 2 (default 0.707), the prior width to use in calculating Bayes factors |
wilcoxon |
|
hypothesis |
|
norm |
|
qq |
|
meanDiff |
|
ci |
|
ciWidth |
a number between 50 and 99.9 (default: 95), the width of confidence intervals |
effectSize |
|
ciES |
|
ciWidthES |
a number between 50 and 99.9 (default: 95), the width of confidence intervals for the effect sizes |
desc |
|
plots |
|
miss |
|
Details
The Student's paired samples t-test assumes that pair differences follow a normal distribution – in the case that one is unwilling to assume this, the non-parametric Wilcoxon signed-rank can be used in it's place (However, note that the Wilcoxon signed-rank has a slightly different null hypothesis; that the two groups of measurements follow the same distribution).
Value
A results object containing:
results$ttest | a table containing the t-test results | ||||
results$norm | a table containing the normality test results | ||||
results$desc | a table containing the descriptives | ||||
results$plots | an array of the descriptive plots | ||||
Tables can be converted to data frames with asDF or as.data.frame. For example:
results$ttest$asDF
as.data.frame(results$ttest)
Examples
data('bugs', package = 'jmv')
ttestPS(bugs, pairs = list(
list(i1 = 'LDLF', i2 = 'LDHF')))
#
# PAIRED SAMPLES T-TEST
#
# Paired Samples T-Test
# --------------------------------------------------------------
# statistic df p
# --------------------------------------------------------------
# LDLF LDHF Student's t -6.65 90.0 < .001
# --------------------------------------------------------------
#