Package 'statsExpressions'

Template for expressions with statistical details

Description

Creates an expression from a data frame containing statistical details. Ideally, this data frame would come from having run tidy_model_parameters() on your model object.

This function is currently not stable and should not be used outside of this package context.

Usage

add_expression_col(
  data,
  paired = FALSE,
  statistic.text = NULL,
  effsize.text = NULL,
  prior.type = NULL,
  n = NULL,
  n.text = list(if (paired) quote(italic("n")["pairs"]) else quote(italic("n")["obs"])),
  digits = 2L,
  digits.df = 0L,
  digits.df.error = digits.df,
  ...
)

Arguments

data	A data frame containing details from the statistical analysis and should contain some or all of the following columns: statistic: the numeric value of a statistic. df.error: the numeric value of a parameter being modeled (often degrees of freedom for the test); irrelevant if there are no degrees of freedom. df: relevant if the statistic in question has two degrees of freedom. p.value: the p-value associated with the observed statistic. method: method describing the test carried out. effectsize: name of the effect size (if not present, same as method). estimate: estimated value of the effect size. conf.level: width for the confidence intervals. conf.low: lower bound for effect size estimate. conf.high: upper bound for effect size estimate. bf10: Bayes Factor value. If this column is present, the analysis is treated as Bayesian and a Bayesian expression template is used, which additionally needs the conf.method and prior.scale columns.
paired	Logical that decides whether the experimental design is repeated measures/within-subjects or between-subjects. The default is FALSE.
statistic.text	A list containing a language object that specifies the relevant test statistic. For example, for tests with t-statistic, statistic.text = list(quote(italic("t"))). If NULL (default), it is inferred from the method column.
effsize.text	A list containing a language object that specifies the relevant effect size or posterior estimate. For example, effsize.text = list(quote(italic("d"))). If NULL (default), it is inferred from the effectsize column.
prior.type	Currently ignored. The prior label is inferred from the method column.
n	An integer specifying the sample size used for the test. Ignored if data already contains an n.obs column.
n.text	A list containing a language object that specifies the design, which will determine what the n stands for. It defaults to quote(italic("n")["pairs"]) if paired = TRUE, and to quote(italic("n")["obs"]) if paired = FALSE.
digits, digits.df, digits.df.error	Number of decimal places to display for the parameters (digits; default: 2L), and for the degrees of freedom in the df (digits.df; default: 0L) and df.error (digits.df.error; default: same as digits.df) columns.
...	Currently ignored.

Citation

Patil, I., (2021). statsExpressions: R Package for Tidy Dataframes and Expressions with Statistical Details. Journal of Open Source Software, 6(61), 3236, https://doi.org/10.21105/joss.03236

Examples

set.seed(123)

# creating a data frame with stats results
stats_df <- cbind.data.frame(
  statistic  = 5.494,
  df         = 29.234,
  p.value    = 0.00001,
  estimate   = -1.980,
  conf.level = 0.95,
  conf.low   = -2.873,
  conf.high  = -1.088,
  method     = "Student's t-test"
)

# expression for *t*-statistic with Cohen's *d* as effect size
# note that the plotmath expressions need to be quoted
add_expression_col(
  data           = stats_df,
  statistic.text = list(quote(italic("t"))),
  effsize.text   = list(quote(italic("d"))),
  n              = 32L,
  n.text         = list(quote(italic("n")["no.obs"])),
  digits         = 3L,
  digits.df      = 3L
)

---

Tidy version of the "Bugs" dataset.

Description

Tidy version of the "Bugs" dataset.

Usage

bugs_long

Format

A data frame with 372 rows and 6 variables

• subject. Dummy identity number for each participant.

• gender. Participant's gender (Female, Male).

• region. Region of the world the participant was from.

• education. Level of education.

• condition. Condition of the experiment the participant gave rating for (LDLF: low disgustingness and low frighteningness; LDHF: low disgustingness and high frighteningness; HDLF: high disgustingness and low frighteningness; HDHF: high disgustingness and high frighteningness).

• desire. The desire to kill an arthropod was indicated on a scale from 0 to 10.

Details

This data set, "Bugs", provides the extent to which men and women want to kill arthropods that vary in frighteningness (low, high) and disgustingness (low, high). Each participant rates their attitudes towards all arthropods. Subset of the data reported by Ryan et al. (2013).

References

Ryan, R. S., Wilde, M., & Crist, S. (2013). Compared to a small, supervised lab experiment, a large, unsupervised web-based experiment on a previously unknown effect has benefits that outweigh its potential costs. Computers in Human Behavior, 29(4), 1295-1301.

Examples

dplyr::glimpse(bugs_long)

---

Data frame and expression for distribution properties

Description

Parametric, non-parametric, robust, and Bayesian measures of centrality.

Usage

centrality_description(
  data,
  x,
  y,
  type = "parametric",
  conf.level = 0.95,
  tr = 0.2,
  digits = 2L,
  ...
)

Arguments

data	A data frame (or a tibble) from which variables specified are to be taken. Other data types (e.g., matrix, table, array, etc.) will not be accepted. Additionally, grouped data frames from {dplyr} should be ungrouped before they are entered as data.
x	The grouping (or independent) variable in data.
y	The response (or outcome or dependent) variable from data.
type	A character specifying the type of statistical approach: "parametric" "nonparametric" "robust" "bayes" You can specify just the initial letter (e.g. "np" or "bf"). Matching is on the initial lowercase letter only, so any other value (including capitalized values such as "Bayes") falls back to "parametric" without a warning.
conf.level	Scalar between 0 and 1 (default: ⁠95%⁠ confidence/credible intervals, 0.95).
tr	Trim level for the mean when carrying out robust tests. In case of an error, try reducing the value of tr, which is by default set to 0.2. Lowering the value might help.
digits	Number of digits for rounding or significant figures. May also be "signif" to return significant figures or "scientific" to return scientific notation. Control the number of digits by adding the value as suffix, e.g. digits = "scientific4" to have scientific notation with 4 decimal places, or digits = "signif5" for 5 significant figures (see also signif()).
...	Currently ignored.

Details

This function describes a distribution for y variable for each level of the grouping variable in x by a set of indices (e.g., measures of centrality, dispersion, range, skewness, kurtosis, etc.). It additionally returns an expression containing a specified centrality measure. The function internally relies on datawizard::describe_distribution() function.

Value

The returned object is a tibble data frame with the additional class "statsExpressions". Unlike the hypothesis-testing functions, this one returns one row per level of the grouping variable x, describing the distribution of y within that level. The columns are:

• the grouping variable (named after x, e.g. Species)

• the requested centrality estimate (named after y, e.g. Sepal.Length): the mean (type = "parametric"), median ("nonparametric"), trimmed mean ("robust"), or MAP estimate ("bayes")

• std.dev: standard deviation (for "parametric" and "robust")

• mad: median absolute deviation (for "nonparametric"); neither std.dev nor mad is returned for "bayes"

• iqr: interquartile range

• conf.low, conf.high: bounds of the confidence/credible interval for the centrality estimate

• min, max: minimum and maximum values

• skewness, kurtosis: shape statistics of the distribution

• n.obs: number of non-missing observations

• missing.obs: number of missing observations after filtering (always 0, since rows with missing x or y are removed before summarizing)

• expression: a list-column of pre-formatted plotmath expressions (each element is a language object, not a character string) containing the centrality estimate, ready to be used in {ggplot2}

• n.expression: a label combining the group name and its sample size (e.g. "setosa\n(n = 50)"), useful as an axis or facet label

For more examples, see the Data frame outputs and the Return value schema articles.

Centrality measures

Type	Measure	Function used
Parametric	mean	datawizard::describe_distribution()
Non-parametric	median	datawizard::describe_distribution()
Robust	trimmed mean	datawizard::describe_distribution()
Bayesian	MAP	datawizard::describe_distribution()

Citation

Patil, I., (2021). statsExpressions: R Package for Tidy Dataframes and Expressions with Statistical Details. Journal of Open Source Software, 6(61), 3236, https://doi.org/10.21105/joss.03236

Examples

# for reproducibility
set.seed(123)

# ----------------------- parametric -----------------------

centrality_description(iris, Species, Sepal.Length, type = "parametric")

# ----------------------- non-parametric -------------------

centrality_description(mtcars, am, wt, type = "nonparametric")

# ----------------------- robust ---------------------------

centrality_description(ToothGrowth, supp, len, type = "robust")

# ----------------------- Bayesian -------------------------

centrality_description(sleep, group, extra, type = "bayes")

---

Keep complete subjects in repeated measures data

Description

Prepares long/tidy repeated measures data for analysis by keeping only the subjects observed in every level of x. This is the listwise deletion step that long_to_wide_converter() applies when paired = TRUE, and thus the data that two_sample_test(), oneway_anova(), and pairwise_comparisons() analyze for within-subjects designs.

The following steps are carried out:

• Rows with a missing x or subject.id are dropped, since they cannot be paired.

• An error is thrown if any subject has more than one row in a level of x.

• Subjects with a missing y value, or without a row, in any level of x are removed entirely.

Usage

complete_subjects(data, x, y, subject.id = NULL)

Arguments

data	A data frame (or a tibble) from which variables specified are to be taken. Other data types (e.g., matrix, table, array, etc.) will not be accepted. Additionally, grouped data frames from {dplyr} should be ungrouped before they are entered as data.
x	The grouping (or independent) variable from data. For repeated measures designs, see the note on ordering in subject.id.
y	The response (or outcome or dependent) variable from data.
subject.id	Relevant in case of a repeated measures or within-subjects design (i.e., paired = TRUE), it specifies the subject or repeated measures identifier. Important: If this argument is NULL (which is the default), observations are paired by their row order within each level of x (i.e., the data is assumed to be sorted in a subject-1, subject-2, ... pattern within every level). If the data is not sorted this way, the paired results will be silently incorrect, so it is safest to always specify subject.id. For paired designs, each subject must have at most one row per level of x (an error is thrown otherwise), and rows with a missing subject.id are dropped because they cannot be paired (see complete_subjects()).

Value

A tibble in long format with columns .rowid (the subject identifier, or an internal row identifier when subject.id is NULL), x (as a factor with unused levels dropped), and y, sorted by .rowid.

Citation

Patil, I., (2021). statsExpressions: R Package for Tidy Dataframes and Expressions with Statistical Details. Journal of Open Source Software, 6(61), 3236, https://doi.org/10.21105/joss.03236

Examples

# subjects with a missing rating in any condition are removed
complete_subjects(bugs_long, condition, desire, subject.id = subject)

---

Contingency table analyses

Description

Parametric and Bayesian one-way and two-way contingency table analyses.

Usage

contingency_table(
  data,
  x,
  y = NULL,
  paired = FALSE,
  type = "parametric",
  counts = NULL,
  ratio = NULL,
  alternative = "two.sided",
  digits = 2L,
  conf.level = 0.95,
  sampling.plan = "indepMulti",
  fixed.margin = "rows",
  prior.concentration = 1,
  ...
)

Arguments

data	A data frame (or a tibble) from which variables specified are to be taken. Other data types (e.g., matrix, table, array, etc.) will not be accepted. Additionally, grouped data frames from {dplyr} should be ungrouped before they are entered as data.
x	The variable to use as the rows in the contingency table.
y	The variable to use as the columns in the contingency table. Default is NULL. If NULL, one-sample proportion test (a goodness of fit test) will be run for the x variable.
paired	Logical indicating whether data came from a within-subjects or repeated measures design study (Default: FALSE). Only relevant for two-way tables (i.e., when y is supplied), and paired designs are only supported for frequentist tests (McNemar's test). For a two-way table with type = "bayes", paired is ignored and the Bayesian test of independence for unpaired data is run instead. For one-way tables (y = NULL), paired has no effect.
type	A character specifying the type of statistical approach: "parametric" "nonparametric" "robust" "bayes" You can specify just the initial letter (e.g. "np" or "bf"). Matching is on the initial lowercase letter only, so any other value (including capitalized values such as "Bayes") falls back to "parametric" without a warning.
counts	The variable in data containing counts, or NULL if each row represents a single observation.
ratio	A vector of proportions: the expected proportions for the proportion test (should sum to 1). Default is NULL, which means the null is equal theoretical proportions across the levels of the nominal variable. E.g., ratio = c(0.5, 0.5) for two levels, ratio = c(0.25, 0.25, 0.25, 0.25) for four levels, etc.
alternative	A character string specifying the alternative hypothesis; Controls the type of CI returned: "two.sided" (default, two-sided CI), "greater" or "less" (one-sided CI). Partial matching is allowed (e.g., "g", "l", "two"...). See section One-Sided CIs in the effectsize_CIs vignette.
digits	Number of digits for rounding or significant figures. May also be "signif" to return significant figures or "scientific" to return scientific notation. Control the number of digits by adding the value as suffix, e.g. digits = "scientific4" to have scientific notation with 4 decimal places, or digits = "signif5" for 5 significant figures (see also signif()).
conf.level	Scalar between 0 and 1 (default: ⁠95%⁠ confidence/credible intervals, 0.95).
sampling.plan	Character describing the sampling plan. Possible options: "indepMulti" (independent multinomial; default) "poisson" "jointMulti" (joint multinomial) "hypergeom" (hypergeometric). Only used for Bayesian two-way tables. For more, see BayesFactor::contingencyTableBF().
fixed.margin	For the independent multinomial sampling plan, which margin is fixed ("rows" or "cols"). Defaults to "rows". Only used for Bayesian two-way tables.
prior.concentration	Specifies the prior concentration parameter, set to 1 by default. It indexes the expected deviation from the null hypothesis under the alternative, and corresponds to Gunel and Dickey's (1974) "a" parameter. Only used for Bayesian analyses.
...	Additional arguments (currently ignored).

Value

The returned object is a tibble data frame with the additional class "statsExpressions". The exact set of columns depends on the test and, for functions that accept a type argument, on the chosen analysis (parametric, non-parametric, robust, or Bayesian). Any given call therefore returns some (not all) of the columns below.

Hypothesis testing

• statistic: the numeric value of a statistic

• df and/or df.error: degrees of freedom for the test. A t-test uses df.error, a chi-squared test uses df, and a parametric ANOVA uses both (numerator df and denominator df.error)

• p.value: the p-value associated with the observed statistic (two-sided unless a one-sided alternative is requested, where supported)

• method: the name of the inferential statistical test

Effect size estimation

• effectsize: the name of the effect size

• estimate: estimated value of the effect size

• conf.level: the coverage level of the confidence/credible interval (e.g. 0.95); the interval itself spans conf.low to conf.high

• conf.low: lower bound for the effect size estimate

• conf.high: upper bound for the effect size estimate

• conf.method: method used to compute the confidence/credible interval

• conf.distribution: statistical distribution for the effect

Bayesian analysis (only when type = "bayes")

• bf10: Bayes factor for the alternative hypothesis relative to the null

• log_e_bf10: natural logarithm of the Bayes factor (present for most, but not all, Bayesian analyses)

• prior.distribution, prior.scale, prior.location: prior specification used to compute the Bayes factor and posterior estimates

Pairwise comparisons (for pairwise_comparisons() and pairwise_contingency_table())

• group1, group2: the two levels being compared

• p.adjust.method: the adjustment method used for multiple comparisons

• p.value.adj: the adjusted p-value; returned by pairwise_contingency_table(). Note that pairwise_comparisons() instead folds the adjusted value into p.value (and does not return a separate p.value.adj column)

Common columns

• n.obs: sample size used for the analysis: observations for unpaired tests, complete subjects/pairs for paired two-sample tests and ANOVAs, or studies for meta-analysis

• expression: a list-column of pre-formatted plotmath expressions; each element is a language object (not a character string) containing the statistical details, ready to be used in {ggplot2} (e.g. in labs() or annotate())

For a per-function, column-by-column breakdown of the output (and an explanation of the internal add_expression_col() engine that builds the expression column), see the Return value schema article. For more examples, see the Data frame outputs article.

Contingency table analyses

Below is a summary of:

• the statistical test carried out for inferential statistics

• the type of effect size estimate and a measure of uncertainty for this estimate

• the functions used internally to compute these details

There are no dedicated non-parametric or robust contingency table analyses. type = "nonparametric" and type = "robust" are accepted, but run the same frequentist tests as type = "parametric" (the "Parametric/Non-parametric" rows below).

Two-way table

Hypothesis testing

Type	Design	Test	Function used
Parametric/Non-parametric	Unpaired	Pearson's chi-squared test	stats::chisq.test()
Bayesian	Unpaired	Bayesian Pearson's chi-squared test	BayesFactor::contingencyTableBF()
Parametric/Non-parametric	Paired	McNemar's chi-squared test	stats::mcnemar.test()
Bayesian	Paired	No	No

Effect size estimation

Type	Design	Effect size	CI available?	Function used
Parametric/Non-parametric	Unpaired	Cramer's V	Yes	effectsize::cramers_v()
Bayesian	Unpaired	Cramer's V	Yes	effectsize::cramers_v()
Parametric/Non-parametric	Paired	Cohen's g	Yes	effectsize::cohens_g()
Bayesian	Paired	No	No	No

Paired Bayesian analysis is not supported: for a two-way table with type = "bayes", the paired argument is ignored and the unpaired Bayesian test is run instead.

One-way table

Hypothesis testing

Type	Test	Function used
Parametric/Non-parametric	Goodness of fit chi-squared test	stats::chisq.test()
Bayesian	Bayesian Goodness of fit chi-squared test	(custom)

Effect size estimation

Type	Effect size	CI available?	Function used
Parametric/Non-parametric	Pearson's C	Yes	effectsize::pearsons_c()
Bayesian	No	No	No

Citation

Patil, I., (2021). statsExpressions: R Package for Tidy Dataframes and Expressions with Statistical Details. Journal of Open Source Software, 6(61), 3236, https://doi.org/10.21105/joss.03236

Examples

#### -------------------- association test ------------------------ ####

# ------------------------ frequentist ---------------------------------

# unpaired

set.seed(123)
contingency_table(
  data = mtcars,
  x = am,
  y = vs,
  paired = FALSE
)

# paired

paired_data <- dplyr::tibble(
  response_before = factor(c("no", "yes", "no", "yes")),
  response_after = factor(c("no", "no", "yes", "yes")),
  Freq = c(65L, 25L, 5L, 5L)
)

set.seed(123)
contingency_table(
  data = paired_data,
  x = response_before,
  y = response_after,
  paired = TRUE,
  counts = Freq
)

# ------------------------ Bayesian -------------------------------------

# unpaired

set.seed(123)
contingency_table(
  data = mtcars,
  x = am,
  y = vs,
  paired = FALSE,
  type = "bayes"
)

# paired

set.seed(123)
contingency_table(
  data = paired_data,
  x = response_before,
  y = response_after,
  paired = TRUE,
  counts = Freq,
  type = "bayes"
)

#### -------------------- goodness-of-fit test -------------------- ####

# ------------------------ frequentist ---------------------------------

set.seed(123)
contingency_table(
  data = as.data.frame(HairEyeColor),
  x = Eye,
  counts = Freq
)

# ------------------------ Bayesian -------------------------------------

set.seed(123)
contingency_table(
  data = as.data.frame(HairEyeColor),
  x = Eye,
  counts = Freq,
  ratio = c(0.2, 0.2, 0.3, 0.3),
  type = "bayes"
)

---

Correlation analyses

Description

Parametric, non-parametric, robust, and Bayesian correlation test.

Usage

corr_test(
  data,
  x,
  y,
  type = "parametric",
  digits = 2L,
  conf.level = 0.95,
  tr = 0.2,
  bf.prior = 0.707,
  ...
)

Arguments

data	A data frame (or a tibble) from which variables specified are to be taken. Other data types (e.g., matrix, table, array, etc.) will not be accepted. Additionally, grouped data frames from {dplyr} should be ungrouped before they are entered as data.
x	The column in data containing the explanatory variable to be plotted on the x-axis.
y	The column in data containing the response (outcome) variable to be plotted on the y-axis.
type	A character specifying the type of statistical approach: "parametric" "nonparametric" "robust" "bayes" You can specify just the initial letter (e.g. "np" or "bf"). Matching is on the initial lowercase letter only, so any other value (including capitalized values such as "Bayes") falls back to "parametric" without a warning.
digits	Number of digits for rounding or significant figures. May also be "signif" to return significant figures or "scientific" to return scientific notation. Control the number of digits by adding the value as suffix, e.g. digits = "scientific4" to have scientific notation with 4 decimal places, or digits = "signif5" for 5 significant figures (see also signif()).
conf.level	Scalar between 0 and 1 (default: ⁠95%⁠ confidence/credible intervals, 0.95).
tr	Trim level for the mean when carrying out robust tests. In case of an error, try reducing the value of tr, which is by default set to 0.2. Lowering the value might help.
bf.prior	A number between 0.5 and 2 (default 0.707), the prior width to use in calculating Bayes factors and posterior estimates. In addition to numeric arguments, several named values are also recognized: "medium", "wide", and "ultrawide", corresponding to r scale values of 1/2, sqrt(2)/2, and 1, respectively. In case of an ANOVA, this value corresponds to scale for fixed effects.
...	Additional arguments (currently ignored).

Value

The returned object is a tibble data frame with the additional class "statsExpressions". The exact set of columns depends on the test and, for functions that accept a type argument, on the chosen analysis (parametric, non-parametric, robust, or Bayesian). Any given call therefore returns some (not all) of the columns below.

Hypothesis testing

• statistic: the numeric value of a statistic

• df and/or df.error: degrees of freedom for the test. A t-test uses df.error, a chi-squared test uses df, and a parametric ANOVA uses both (numerator df and denominator df.error)

• p.value: the p-value associated with the observed statistic (two-sided unless a one-sided alternative is requested, where supported)

• method: the name of the inferential statistical test

Effect size estimation

• effectsize: the name of the effect size

• estimate: estimated value of the effect size

• conf.level: the coverage level of the confidence/credible interval (e.g. 0.95); the interval itself spans conf.low to conf.high

• conf.low: lower bound for the effect size estimate

• conf.high: upper bound for the effect size estimate

• conf.method: method used to compute the confidence/credible interval

• conf.distribution: statistical distribution for the effect

Bayesian analysis (only when type = "bayes")

• bf10: Bayes factor for the alternative hypothesis relative to the null

• log_e_bf10: natural logarithm of the Bayes factor (present for most, but not all, Bayesian analyses)

• prior.distribution, prior.scale, prior.location: prior specification used to compute the Bayes factor and posterior estimates

Pairwise comparisons (for pairwise_comparisons() and pairwise_contingency_table())

• group1, group2: the two levels being compared

• p.adjust.method: the adjustment method used for multiple comparisons

• p.value.adj: the adjusted p-value; returned by pairwise_contingency_table(). Note that pairwise_comparisons() instead folds the adjusted value into p.value (and does not return a separate p.value.adj column)

Common columns

• n.obs: sample size used for the analysis: observations for unpaired tests, complete subjects/pairs for paired two-sample tests and ANOVAs, or studies for meta-analysis

• expression: a list-column of pre-formatted plotmath expressions; each element is a language object (not a character string) containing the statistical details, ready to be used in {ggplot2} (e.g. in labs() or annotate())

For a per-function, column-by-column breakdown of the output (and an explanation of the internal add_expression_col() engine that builds the expression column), see the Return value schema article. For more examples, see the Data frame outputs article.

Correlation analyses

Below is a summary of:

• the statistical test carried out for inferential statistics

• the type of effect size estimate and a measure of uncertainty for this estimate

• the functions used internally to compute these details

Hypothesis testing and effect size estimation

Type	Test	CI available?	Function used
Parametric	Pearson's correlation coefficient	Yes	correlation::correlation()
Non-parametric	Spearman's rank correlation coefficient	Yes	correlation::correlation()
Robust	Winsorized Pearson's correlation coefficient	Yes	correlation::correlation()
Bayesian	Bayesian Pearson's correlation coefficient	Yes	correlation::correlation()

Citation

Patil, I., (2021). statsExpressions: R Package for Tidy Dataframes and Expressions with Statistical Details. Journal of Open Source Software, 6(61), 3236, https://doi.org/10.21105/joss.03236

Examples

# for reproducibility
set.seed(123)

# ----------------------- parametric -----------------------

corr_test(mtcars, wt, mpg, type = "parametric")

# ----------------------- non-parametric -------------------

corr_test(mtcars, wt, mpg, type = "nonparametric")

# ----------------------- robust ---------------------------

corr_test(mtcars, wt, mpg, type = "robust")

# ----------------------- Bayesian -------------------------

corr_test(mtcars, wt, mpg, type = "bayes")

---

Switch the type of statistics.

Description

Relevant mostly for {ggstatsplot} and {statsExpressions} packages, where different statistical approaches are supported via this argument: parametric, non-parametric, robust, and Bayesian. This switch function converts strings entered by users to a common pattern for convenience.

stats_type_switch() is an alias of extract_stats_type().

Usage

extract_stats_type(type)

stats_type_switch(type)

Arguments

type

A character specifying the type of statistical approach: "parametric" "nonparametric" "robust" "bayes" You can specify just the initial letter (e.g. "np" or "bf"). Matching is on the initial lowercase letter only, so any other value (including capitalized values such as "Bayes") falls back to "parametric" without a warning.

Value

A character vector of the same length as type, with values "parametric", "nonparametric", "robust", or "bayes".

Examples

extract_stats_type("p")
extract_stats_type("bf")
extract_stats_type(c("np", "robust", "Bayes"))

---

Edgar Anderson's Iris Data in long format.

Description

Edgar Anderson's Iris Data in long format.

Usage

iris_long

Format

A data frame with 600 rows and 6 variables

• id. Dummy identity number for each flower (150 flowers in total).

• Species. The species are Iris setosa, versicolor, and virginica.

• condition. Factor giving a detailed description of the attribute (Four levels: "Petal.Length", "Petal.Width", "Sepal.Length", "Sepal.Width").

• attribute. What attribute is being measured ("Sepal" or "Petal").

• measure. What aspect of the attribute is being measured ("Length" or "Width").

• value. Value of the measurement.

Details

This famous (Fisher's or Anderson's) iris data set gives the measurements in centimeters of the variables sepal length and width and petal length and width, respectively, for 50 flowers from each of 3 species of iris. The species are Iris setosa, versicolor, and virginica.

This is a modified dataset from {datasets} package.

Examples

dplyr::glimpse(iris_long)

---

Convert long/tidy data frame to wide format

Description

This conversion is helpful mostly for repeated measures design, where removing NAs by participant can be a bit tedious.

Usage

long_to_wide_converter(
  data,
  x,
  y,
  subject.id = NULL,
  paired = TRUE,
  spread = TRUE,
  ...
)

Arguments

data	A data frame (or a tibble) from which variables specified are to be taken. Other data types (e.g., matrix, table, array, etc.) will not be accepted. Additionally, grouped data frames from {dplyr} should be ungrouped before they are entered as data.
x	The grouping (or independent) variable from data. For repeated measures designs, see the note on ordering in subject.id.
y	The response (or outcome or dependent) variable from data.
subject.id	Relevant in case of a repeated measures or within-subjects design (i.e., paired = TRUE), it specifies the subject or repeated measures identifier. Important: If this argument is NULL (which is the default), observations are paired by their row order within each level of x (i.e., the data is assumed to be sorted in a subject-1, subject-2, ... pattern within every level). If the data is not sorted this way, the paired results will be silently incorrect, so it is safest to always specify subject.id. For paired designs, each subject must have at most one row per level of x (an error is thrown otherwise), and rows with a missing subject.id are dropped because they cannot be paired (see complete_subjects()).
paired	Logical that decides whether the experimental design is repeated measures/within-subjects or between-subjects. The default is FALSE.
spread	Logical that decides whether the data frame needs to be converted from long/tidy to wide (default: TRUE).
...	Currently ignored.

Value

A tibble with NAs removed while respecting the between-or-within-subjects nature of the dataset: for paired designs, only subjects observed in every level of x are kept (a subject with a missing value or a missing row in any condition is removed entirely), while for unpaired designs only the rows with missing values are removed. Rows are grouped by subject.id whenever it is supplied, so with paired = FALSE and a subject.id, a missing value still removes every row of that subject. The .rowid column contains the subject identifier (or an internal row identifier).

Citation

Patil, I., (2021). statsExpressions: R Package for Tidy Dataframes and Expressions with Statistical Details. Journal of Open Source Software, 6(61), 3236, https://doi.org/10.21105/joss.03236

Examples

# repeated measures design
long_to_wide_converter(
  bugs_long,
  condition,
  desire,
  subject.id = subject,
  paired = TRUE
)

# independent measures design
long_to_wide_converter(mtcars, cyl, wt, paired = FALSE)

---

Random-effects meta-analysis

Description

Parametric, robust, and Bayesian random-effects meta-analysis.

A non-parametric meta-analysis is not available, so type = "nonparametric" is not supported.

Usage

meta_analysis(
  data,
  type = "parametric",
  random = "mixture",
  digits = 2L,
  conf.level = 0.95,
  ...
)

Arguments

data	A data frame. It must contain columns named estimate (effect sizes or outcomes) and std.error (corresponding standard errors). These two columns will be used: as yi and sei arguments in metafor::rma() (for parametric test) as yi and sei arguments in metaplus::metaplus() (for robust test) as y and SE arguments in metaBMA::meta_random() (for Bayesian test)
type	A character specifying the type of statistical approach: "parametric" "nonparametric" "robust" "bayes" You can specify just the initial letter (e.g. "np" or "bf"). Matching is on the initial lowercase letter only, so any other value (including capitalized values such as "Bayes") falls back to "parametric" without a warning.
random	The type of random-effects distribution for the robust meta-analysis: "mixture" (default; mixture of normals), "normal", or "t-dist" (t-distribution). Passed to metaplus::metaplus() and only used when type = "robust".
digits	Number of digits for rounding or significant figures. May also be "signif" to return significant figures or "scientific" to return scientific notation. Control the number of digits by adding the value as suffix, e.g. digits = "scientific4" to have scientific notation with 4 decimal places, or digits = "signif5" for 5 significant figures (see also signif()).
conf.level	Scalar between 0 and 1 (default: ⁠95%⁠ confidence/credible intervals, 0.95).
...	Additional arguments passed to the respective meta-analysis function.

Value

The returned object is a tibble data frame with the additional class "statsExpressions". The exact set of columns depends on the test and, for functions that accept a type argument, on the chosen analysis (parametric, non-parametric, robust, or Bayesian). Any given call therefore returns some (not all) of the columns below.

Hypothesis testing

• statistic: the numeric value of a statistic

• df and/or df.error: degrees of freedom for the test. A t-test uses df.error, a chi-squared test uses df, and a parametric ANOVA uses both (numerator df and denominator df.error)

• p.value: the p-value associated with the observed statistic (two-sided unless a one-sided alternative is requested, where supported)

• method: the name of the inferential statistical test

Effect size estimation

• effectsize: the name of the effect size

• estimate: estimated value of the effect size

• conf.level: the coverage level of the confidence/credible interval (e.g. 0.95); the interval itself spans conf.low to conf.high

• conf.low: lower bound for the effect size estimate

• conf.high: upper bound for the effect size estimate

• conf.method: method used to compute the confidence/credible interval

• conf.distribution: statistical distribution for the effect

Bayesian analysis (only when type = "bayes")

• bf10: Bayes factor for the alternative hypothesis relative to the null

• log_e_bf10: natural logarithm of the Bayes factor (present for most, but not all, Bayesian analyses)

• prior.distribution, prior.scale, prior.location: prior specification used to compute the Bayes factor and posterior estimates

Pairwise comparisons (for pairwise_comparisons() and pairwise_contingency_table())

• group1, group2: the two levels being compared

• p.adjust.method: the adjustment method used for multiple comparisons

• p.value.adj: the adjusted p-value; returned by pairwise_contingency_table(). Note that pairwise_comparisons() instead folds the adjusted value into p.value (and does not return a separate p.value.adj column)

Common columns

• n.obs: sample size used for the analysis: observations for unpaired tests, complete subjects/pairs for paired two-sample tests and ANOVAs, or studies for meta-analysis

• expression: a list-column of pre-formatted plotmath expressions; each element is a language object (not a character string) containing the statistical details, ready to be used in {ggplot2} (e.g. in labs() or annotate())

For a per-function, column-by-column breakdown of the output (and an explanation of the internal add_expression_col() engine that builds the expression column), see the Return value schema article. For more examples, see the Data frame outputs article.

Random-effects meta-analysis

Below is a summary of:

• the statistical test carried out for inferential statistics

• the type of effect size estimate and a measure of uncertainty for this estimate

• the functions used internally to compute these details

Hypothesis testing and effect size estimation

Type	Test	Effect size	CI available?	Function used
Parametric	Meta-analysis via random-effects models	beta	Yes	metafor::rma()
Robust	Meta-analysis via robust random-effects models	beta	Yes	metaplus::metaplus()
Bayesian	Meta-analysis via Bayesian random-effects models	beta	Yes	metaBMA::meta_random()

Citation

Patil, I., (2021). statsExpressions: R Package for Tidy Dataframes and Expressions with Statistical Details. Journal of Open Source Software, 6(61), 3236, https://doi.org/10.21105/joss.03236

Note

The package needed for the chosen type ({metafor}, {metaplus}, or {metaBMA}) is only suggested, so you will be asked to install it if it is not available.

Examples

# let's use `mag` dataset from `{metaplus}`
data(mag, package = "metaplus")
dat <- dplyr::rename(mag, estimate = yi, std.error = sei)

# ----------------------- parametric ---------------------------------------

set.seed(123)
meta_analysis(dat)

# ----------------------- robust --------------------------------------------

set.seed(123)
meta_analysis(dat, type = "robust", random = "normal")

# ----------------------- Bayesian ------------------------------------------

# suppress warnings about divergent transitions after warmup
set.seed(123)
suppressWarnings(meta_analysis(dat, type = "bayes"))

---

Movie information and user ratings from IMDB.

Description

Movie information and user ratings from IMDB.

Usage

movies_long

Format

A data frame with 1,579 rows and 8 variables

• title. Title of the movie.

• year. Year of release.

• budget. Total budget (if known) in US dollars.

• length. Length in minutes.

• rating. Average IMDB user rating.

• votes. Number of IMDB users who rated this movie.

• mpaa. MPAA rating.

• genre. Genre of the movie ("Action", "Action Comedy", "Action Drama", "Animated", "Comedy", "Comedy Drama", "Drama", "Romance Drama", "RomCom").

Details

Modified dataset from {ggplot2movies} package.

Source

https://CRAN.R-project.org/package=ggplot2movies

Examples

dplyr::glimpse(movies_long)

---

One-sample tests

Description

Parametric, non-parametric, robust, and Bayesian one-sample tests.

Usage

one_sample_test(
  data,
  x,
  type = "parametric",
  test.value = 0,
  alternative = "two.sided",
  digits = 2L,
  conf.level = 0.95,
  tr = 0.2,
  bf.prior = 0.707,
  effsize.type = "g",
  exact = FALSE,
  ...
)

Arguments

data	A data frame (or a tibble) from which variables specified are to be taken. Other data types (e.g., matrix, table, array, etc.) will not be accepted. Additionally, grouped data frames from {dplyr} should be ungrouped before they are entered as data.
x	A numeric variable from the data frame data.
type	A character specifying the type of statistical approach: "parametric" "nonparametric" "robust" "bayes" You can specify just the initial letter (e.g. "np" or "bf"). Matching is on the initial lowercase letter only, so any other value (including capitalized values such as "Bayes") falls back to "parametric" without a warning.
test.value	A number indicating the true value of the mean (Default: 0).
alternative	a character string specifying the alternative hypothesis, must be one of "two.sided" (default), "greater" or "less". You can specify just the initial letter.
digits	Number of digits for rounding or significant figures. May also be "signif" to return significant figures or "scientific" to return scientific notation. Control the number of digits by adding the value as suffix, e.g. digits = "scientific4" to have scientific notation with 4 decimal places, or digits = "signif5" for 5 significant figures (see also signif()).
conf.level	Scalar between 0 and 1 (default: ⁠95%⁠ confidence/credible intervals, 0.95).
tr	Trim level for the mean when carrying out robust tests. In case of an error, try reducing the value of tr, which is by default set to 0.2. Lowering the value might help.
bf.prior	A number between 0.5 and 2 (default 0.707), the prior width to use in calculating Bayes factors and posterior estimates. In addition to numeric arguments, several named values are also recognized: "medium", "wide", and "ultrawide", corresponding to r scale values of 1/2, sqrt(2)/2, and 1, respectively. In case of an ANOVA, this value corresponds to scale for fixed effects.
effsize.type	Type of effect size needed for parametric tests. The argument can be "d" (for Cohen's d) or "g" (for Hedges' g).
exact	A logical indicating whether exact p-values should be computed. Relevant only when type = "nonparametric" (Default: FALSE).
...	Currently ignored.

Value

The returned object is a tibble data frame with the additional class "statsExpressions". The exact set of columns depends on the test and, for functions that accept a type argument, on the chosen analysis (parametric, non-parametric, robust, or Bayesian). Any given call therefore returns some (not all) of the columns below.

Hypothesis testing

• statistic: the numeric value of a statistic

• df and/or df.error: degrees of freedom for the test. A t-test uses df.error, a chi-squared test uses df, and a parametric ANOVA uses both (numerator df and denominator df.error)

• p.value: the p-value associated with the observed statistic (two-sided unless a one-sided alternative is requested, where supported)

• method: the name of the inferential statistical test

Effect size estimation

• effectsize: the name of the effect size

• estimate: estimated value of the effect size

• conf.level: the coverage level of the confidence/credible interval (e.g. 0.95); the interval itself spans conf.low to conf.high

• conf.low: lower bound for the effect size estimate

• conf.high: upper bound for the effect size estimate

• conf.method: method used to compute the confidence/credible interval

• conf.distribution: statistical distribution for the effect

Bayesian analysis (only when type = "bayes")

• bf10: Bayes factor for the alternative hypothesis relative to the null

• log_e_bf10: natural logarithm of the Bayes factor (present for most, but not all, Bayesian analyses)

• prior.distribution, prior.scale, prior.location: prior specification used to compute the Bayes factor and posterior estimates

Pairwise comparisons (for pairwise_comparisons() and pairwise_contingency_table())

• group1, group2: the two levels being compared

• p.adjust.method: the adjustment method used for multiple comparisons

• p.value.adj: the adjusted p-value; returned by pairwise_contingency_table(). Note that pairwise_comparisons() instead folds the adjusted value into p.value (and does not return a separate p.value.adj column)

Common columns

• n.obs: sample size used for the analysis: observations for unpaired tests, complete subjects/pairs for paired two-sample tests and ANOVAs, or studies for meta-analysis

• expression: a list-column of pre-formatted plotmath expressions; each element is a language object (not a character string) containing the statistical details, ready to be used in {ggplot2} (e.g. in labs() or annotate())

For a per-function, column-by-column breakdown of the output (and an explanation of the internal add_expression_col() engine that builds the expression column), see the Return value schema article. For more examples, see the Data frame outputs article.

One-sample tests

Below is a summary of:

• the statistical test carried out for inferential statistics

• the type of effect size estimate and a measure of uncertainty for this estimate

• the functions used internally to compute these details

Hypothesis testing

Type	Test	Function used
Parametric	One-sample Student's t-test	stats::t.test()
Non-parametric	One-sample Wilcoxon test	stats::wilcox.test()
Robust	Bootstrap-t method for one-sample test	WRS2::trimcibt()
Bayesian	One-sample Student's t-test	BayesFactor::ttestBF()

Effect size estimation

Type	Effect size	CI available?	Function used
Parametric	Cohen's d, Hedges' g	Yes	effectsize::cohens_d(), effectsize::hedges_g()
Non-parametric	r (rank-biserial correlation)	Yes	effectsize::rank_biserial()
Robust	trimmed mean	Yes	WRS2::trimcibt()
Bayesian	difference	Yes	bayestestR::describe_posterior()

Citation

Patil, I., (2021). statsExpressions: R Package for Tidy Dataframes and Expressions with Statistical Details. Journal of Open Source Software, 6(61), 3236, https://doi.org/10.21105/joss.03236

Examples

# for reproducibility
set.seed(123)

# ----------------------- parametric -----------------------

one_sample_test(mtcars, wt, test.value = 3)

# biased (Cohen's d) effect size
one_sample_test(mtcars, wt, test.value = 3, effsize.type = "d")

# ----------------------- non-parametric -------------------

one_sample_test(mtcars, wt, test.value = 3, type = "nonparametric")

# ----------------------- robust ---------------------------

one_sample_test(mtcars, wt, test.value = 3, type = "robust")

# ----------------------- Bayesian -------------------------

one_sample_test(mtcars, wt, test.value = 3, type = "bayes")

---

One-way analysis of variance (ANOVA)

Description

Parametric, non-parametric, robust, and Bayesian one-way ANOVA.

Usage

oneway_anova(
  data,
  x,
  y,
  subject.id = NULL,
  type = "parametric",
  paired = FALSE,
  digits = 2L,
  conf.level = 0.95,
  effsize.type = "omega",
  var.equal = FALSE,
  bf.prior = 0.707,
  tr = 0.2,
  nboot = 100L,
  ...
)

Arguments

data	A data frame (or a tibble) from which variables specified are to be taken. Other data types (e.g., matrix, table, array, etc.) will not be accepted. Additionally, grouped data frames from {dplyr} should be ungrouped before they are entered as data.
x	The grouping (or independent) variable from data. For repeated measures designs, see the note on ordering in subject.id.
y	The response (or outcome or dependent) variable from data.
subject.id	Relevant in case of a repeated measures or within-subjects design (i.e., paired = TRUE), it specifies the subject or repeated measures identifier. Important: If this argument is NULL (which is the default), observations are paired by their row order within each level of x (i.e., the data is assumed to be sorted in a subject-1, subject-2, ... pattern within every level). If the data is not sorted this way, the paired results will be silently incorrect, so it is safest to always specify subject.id. For paired designs, each subject must have at most one row per level of x (an error is thrown otherwise), and rows with a missing subject.id are dropped because they cannot be paired (see complete_subjects()).
type	A character specifying the type of statistical approach: "parametric" "nonparametric" "robust" "bayes" You can specify just the initial letter (e.g. "np" or "bf"). Matching is on the initial lowercase letter only, so any other value (including capitalized values such as "Bayes") falls back to "parametric" without a warning.
paired	Logical that decides whether the experimental design is repeated measures/within-subjects or between-subjects. The default is FALSE.
digits	Number of digits for rounding or significant figures. May also be "signif" to return significant figures or "scientific" to return scientific notation. Control the number of digits by adding the value as suffix, e.g. digits = "scientific4" to have scientific notation with 4 decimal places, or digits = "signif5" for 5 significant figures (see also signif()).
conf.level	Scalar between 0 and 1 (default: ⁠95%⁠ confidence/credible intervals, 0.95).
effsize.type	Type of effect size needed for parametric tests. The argument can be "eta" (partial eta-squared) or "omega" (partial omega-squared).
var.equal	a logical variable indicating whether to treat the two variances as being equal. If TRUE then the pooled variance is used to estimate the variance otherwise the Welch (or Satterthwaite) approximation to the degrees of freedom is used.
bf.prior	A number between 0.5 and 2 (default 0.707), the prior width to use in calculating Bayes factors and posterior estimates. In addition to numeric arguments, several named values are also recognized: "medium", "wide", and "ultrawide", corresponding to r scale values of 1/2, sqrt(2)/2, and 1, respectively. In case of an ANOVA, this value corresponds to scale for fixed effects.
tr	Trim level for the mean when carrying out robust tests. In case of an error, try reducing the value of tr, which is by default set to 0.2. Lowering the value might help.
nboot	Number of bootstrap samples for computing confidence interval for the effect size (Default: 100L).
...	Additional arguments (currently ignored).

Value

The returned object is a tibble data frame with the additional class "statsExpressions". The exact set of columns depends on the test and, for functions that accept a type argument, on the chosen analysis (parametric, non-parametric, robust, or Bayesian). Any given call therefore returns some (not all) of the columns below.

Hypothesis testing

• statistic: the numeric value of a statistic

• df and/or df.error: degrees of freedom for the test. A t-test uses df.error, a chi-squared test uses df, and a parametric ANOVA uses both (numerator df and denominator df.error)

• p.value: the p-value associated with the observed statistic (two-sided unless a one-sided alternative is requested, where supported)

• method: the name of the inferential statistical test

Effect size estimation

• effectsize: the name of the effect size

• estimate: estimated value of the effect size

• conf.level: the coverage level of the confidence/credible interval (e.g. 0.95); the interval itself spans conf.low to conf.high

• conf.low: lower bound for the effect size estimate

• conf.high: upper bound for the effect size estimate

• conf.method: method used to compute the confidence/credible interval

• conf.distribution: statistical distribution for the effect

Bayesian analysis (only when type = "bayes")

• bf10: Bayes factor for the alternative hypothesis relative to the null

• log_e_bf10: natural logarithm of the Bayes factor (present for most, but not all, Bayesian analyses)

• prior.distribution, prior.scale, prior.location: prior specification used to compute the Bayes factor and posterior estimates

Pairwise comparisons (for pairwise_comparisons() and pairwise_contingency_table())

• group1, group2: the two levels being compared

• p.adjust.method: the adjustment method used for multiple comparisons

• p.value.adj: the adjusted p-value; returned by pairwise_contingency_table(). Note that pairwise_comparisons() instead folds the adjusted value into p.value (and does not return a separate p.value.adj column)

Common columns

• n.obs: sample size used for the analysis: observations for unpaired tests, complete subjects/pairs for paired two-sample tests and ANOVAs, or studies for meta-analysis

• expression: a list-column of pre-formatted plotmath expressions; each element is a language object (not a character string) containing the statistical details, ready to be used in {ggplot2} (e.g. in labs() or annotate())

For a per-function, column-by-column breakdown of the output (and an explanation of the internal add_expression_col() engine that builds the expression column), see the Return value schema article. For more examples, see the Data frame outputs article.

One-way ANOVA

Below is a summary of:

• the statistical test carried out for inferential statistics

• the type of effect size estimate and a measure of uncertainty for this estimate

• the functions used internally to compute these details

Between-subjects

Hypothesis testing

Type	No. of groups	Test	Function used
Parametric	> 2	Fisher's or Welch's one-way ANOVA	stats::oneway.test()
Non-parametric	> 2	Kruskal-Wallis one-way ANOVA	stats::kruskal.test()
Robust	> 2	Heteroscedastic one-way ANOVA for trimmed means	WRS2::t1way()
Bayesian	> 2	Fisher's ANOVA	BayesFactor::anovaBF()

Effect size estimation

Type	No. of groups	Effect size	CI available?	Function used
Parametric	> 2	partial eta-squared, partial omega-squared	Yes	effectsize::omega_squared(), effectsize::eta_squared()
Non-parametric	> 2	rank epsilon squared	Yes	effectsize::rank_epsilon_squared()
Robust	> 2	Explanatory measure of effect size	Yes	WRS2::t1way()
Bayesian	> 2	Bayesian R-squared	Yes	performance::r2_bayes()

Within-subjects

Paired data: See complete_subjects() for subject matching, duplicate observations, and removal of incomplete subjects.

Hypothesis testing

Type	No. of groups	Test	Function used
Parametric	> 2	One-way repeated measures ANOVA	afex::aov_ez()
Non-parametric	> 2	Friedman rank sum test	stats::friedman.test()
Robust	> 2	Heteroscedastic one-way repeated measures ANOVA for trimmed means	WRS2::rmanova()
Bayesian	> 2	One-way repeated measures ANOVA	BayesFactor::anovaBF()

Effect size estimation

Type	No. of groups	Effect size	CI available?	Function used
Parametric	> 2	partial eta-squared, partial omega-squared	Yes	effectsize::omega_squared(), effectsize::eta_squared()
Non-parametric	> 2	Kendall's coefficient of concordance	Yes	effectsize::kendalls_w()
Robust	> 2	Algina-Keselman-Penfield robust standardized difference average	Yes	WRS2::wmcpAKP()
Bayesian	> 2	Bayesian R-squared	Yes	performance::r2_bayes()

Citation

Patil, I., (2021). statsExpressions: R Package for Tidy Dataframes and Expressions with Statistical Details. Journal of Open Source Software, 6(61), 3236, https://doi.org/10.21105/joss.03236

Examples

# for reproducibility
set.seed(123)

# ----------------------- parametric -------------------------------------

# between-subjects
oneway_anova(
  data = mtcars,
  x = cyl,
  y = wt
)

# biased (partial eta-squared) effect size
oneway_anova(
  data = mtcars,
  x = cyl,
  y = wt,
  effsize.type = "eta"
)

# within-subjects design
oneway_anova(
  data = bugs_long,
  x = condition,
  y = desire,
  subject.id = subject,
  paired = TRUE
)

# ----------------------- non-parametric ----------------------------------

# between-subjects
oneway_anova(
  data = mtcars,
  x = cyl,
  y = wt,
  type = "np"
)

# within-subjects design
oneway_anova(
  data = bugs_long,
  x = condition,
  y = desire,
  subject.id = subject,
  paired = TRUE,
  type = "np"
)

# ----------------------- robust -------------------------------------

# between-subjects
oneway_anova(
  data = mtcars,
  x = cyl,
  y = wt,
  type = "r"
)

# within-subjects design
oneway_anova(
  data = bugs_long,
  x = condition,
  y = desire,
  subject.id = subject,
  paired = TRUE,
  type = "r"
)

# ----------------------- Bayesian -------------------------------------

# between-subjects
oneway_anova(
  data = mtcars,
  x = cyl,
  y = wt,
  type = "bayes"
)

# within-subjects design
oneway_anova(
  data = bugs_long,
  x = condition,
  y = desire,
  subject.id = subject,
  paired = TRUE,
  type = "bayes"
)

---

Multiple pairwise comparison for one-way design

Description

Calculate parametric, non-parametric, robust, and Bayes Factor pairwise comparisons between group levels with corrections for multiple testing.

Usage

pairwise_comparisons(
  data,
  x,
  y,
  subject.id = NULL,
  type = "parametric",
  paired = FALSE,
  var.equal = FALSE,
  tr = 0.2,
  bf.prior = 0.707,
  p.adjust.method = "holm",
  digits = 2L,
  exact = FALSE,
  ...
)

Arguments

data	A data frame (or a tibble) from which variables specified are to be taken. Other data types (e.g., matrix, table, array, etc.) will not be accepted. Additionally, grouped data frames from {dplyr} should be ungrouped before they are entered as data.
x	The grouping (or independent) variable from data. For repeated measures designs, see the note on ordering in subject.id.
y	The response (or outcome or dependent) variable from data.
subject.id	Relevant in case of a repeated measures or within-subjects design (i.e., paired = TRUE), it specifies the subject or repeated measures identifier. Important: If this argument is NULL (which is the default), observations are paired by their row order within each level of x (i.e., the data is assumed to be sorted in a subject-1, subject-2, ... pattern within every level). If the data is not sorted this way, the paired results will be silently incorrect, so it is safest to always specify subject.id. For paired designs, each subject must have at most one row per level of x (an error is thrown otherwise), and rows with a missing subject.id are dropped because they cannot be paired (see complete_subjects()).
type	A character specifying the type of statistical approach: "parametric" "nonparametric" "robust" "bayes" You can specify just the initial letter (e.g. "np" or "bf"). Matching is on the initial lowercase letter only, so any other value (including capitalized values such as "Bayes") falls back to "parametric" without a warning.
paired	Logical that decides whether the experimental design is repeated measures/within-subjects or between-subjects. The default is FALSE.
var.equal	a logical variable indicating whether to treat the two variances as being equal. If TRUE then the pooled variance is used to estimate the variance otherwise the Welch (or Satterthwaite) approximation to the degrees of freedom is used.
tr	Trim level for the mean when carrying out robust tests. In case of an error, try reducing the value of tr, which is by default set to 0.2. Lowering the value might help.
bf.prior	A number between 0.5 and 2 (default 0.707), the prior width to use in calculating Bayes factors and posterior estimates. In addition to numeric arguments, several named values are also recognized: "medium", "wide", and "ultrawide", corresponding to r scale values of 1/2, sqrt(2)/2, and 1, respectively. In case of an ANOVA, this value corresponds to scale for fixed effects.
p.adjust.method	Adjustment method for p-values for multiple comparisons. Possible methods are: "holm" (default), "hochberg", "hommel", "bonferroni", "BH", "BY", "fdr", "none".
digits	Number of digits for rounding or significant figures. May also be "signif" to return significant figures or "scientific" to return scientific notation. Control the number of digits by adding the value as suffix, e.g. digits = "scientific4" to have scientific notation with 4 decimal places, or digits = "signif5" for 5 significant figures (see also signif()).
exact	A logical indicating whether exact p-values should be computed. Relevant only when type = "nonparametric" (Default: FALSE).
...	Additional arguments passed to the underlying pairwise test function for the parametric and non-parametric tests (see the table below). Ignored for robust tests. For Bayesian tests, they are passed to the frequentist test (stats::pairwise.t.test() or PMCMRplus::gamesHowellTest()) that is run first to enumerate the pairs, so they do not change the Bayes factors, but unsupported arguments can still cause an error.

Value

The returned object is a tibble data frame with the additional class "statsExpressions". The exact set of columns depends on the test and, for functions that accept a type argument, on the chosen analysis (parametric, non-parametric, robust, or Bayesian). Any given call therefore returns some (not all) of the columns below.

Hypothesis testing

• statistic: the numeric value of a statistic

• df and/or df.error: degrees of freedom for the test. A t-test uses df.error, a chi-squared test uses df, and a parametric ANOVA uses both (numerator df and denominator df.error)

• p.value: the p-value associated with the observed statistic (two-sided unless a one-sided alternative is requested, where supported)

• method: the name of the inferential statistical test

Effect size estimation

• effectsize: the name of the effect size

• estimate: estimated value of the effect size

• conf.level: the coverage level of the confidence/credible interval (e.g. 0.95); the interval itself spans conf.low to conf.high

• conf.low: lower bound for the effect size estimate

• conf.high: upper bound for the effect size estimate

• conf.method: method used to compute the confidence/credible interval

• conf.distribution: statistical distribution for the effect

Bayesian analysis (only when type = "bayes")

• bf10: Bayes factor for the alternative hypothesis relative to the null

• log_e_bf10: natural logarithm of the Bayes factor (present for most, but not all, Bayesian analyses)

• prior.distribution, prior.scale, prior.location: prior specification used to compute the Bayes factor and posterior estimates

Pairwise comparisons (for pairwise_comparisons() and pairwise_contingency_table())

• group1, group2: the two levels being compared

• p.adjust.method: the adjustment method used for multiple comparisons

• p.value.adj: the adjusted p-value; returned by pairwise_contingency_table(). Note that pairwise_comparisons() instead folds the adjusted value into p.value (and does not return a separate p.value.adj column)

Common columns

• n.obs: sample size used for the analysis: observations for unpaired tests, complete subjects/pairs for paired two-sample tests and ANOVAs, or studies for meta-analysis

• expression: a list-column of pre-formatted plotmath expressions; each element is a language object (not a character string) containing the statistical details, ready to be used in {ggplot2} (e.g. in labs() or annotate())

For a per-function, column-by-column breakdown of the output (and an explanation of the internal add_expression_col() engine that builds the expression column), see the Return value schema article. For more examples, see the Data frame outputs article.

Pairwise comparison tests

Below is a summary of:

• the statistical test carried out for inferential statistics

• the type of effect size estimate and a measure of uncertainty for this estimate

• the functions used internally to compute these details

Between-subjects

Hypothesis testing

Type	Equal variance?	Test	p-value adjustment?	Function used
Parametric	No	Games-Howell test	Yes	PMCMRplus::gamesHowellTest()
Parametric	Yes	Student's t-test	Yes	stats::pairwise.t.test()
Non-parametric	No	Dunn test	Yes	PMCMRplus::kwAllPairsDunnTest()
Robust	No	Yuen's trimmed means test	Yes	WRS2::lincon()
Bayesian	NA	Student's t-test	NA	BayesFactor::ttestBF()

Effect size estimation

Not supported.

Within-subjects

Paired data: See complete_subjects() for subject matching, duplicate observations, and removal of incomplete subjects.

Hypothesis testing

Type	Test	p-value adjustment?	Function used
Parametric	Student's t-test	Yes	stats::pairwise.t.test()
Non-parametric	Durbin-Conover test	Yes	PMCMRplus::durbinAllPairsTest()
Robust	Yuen's trimmed means test	Yes	WRS2::rmmcp()
Bayesian	Student's t-test	NA	BayesFactor::ttestBF()

Effect size estimation

Not supported.

Citation

Patil, I., (2021). statsExpressions: R Package for Tidy Dataframes and Expressions with Statistical Details. Journal of Open Source Software, 6(61), 3236, https://doi.org/10.21105/joss.03236

References

For more, see: https://www.indrapatil.com/ggstatsplot/articles/web_only/pairwise.html

Examples

# for reproducibility
set.seed(123)

#------------------- between-subjects design ----------------------------

# parametric
# if `var.equal = TRUE`, then Student's t-test will be run
pairwise_comparisons(
  data = mtcars,
  x = cyl,
  y = wt,
  type = "parametric",
  var.equal = TRUE,
  paired = FALSE,
  p.adjust.method = "none"
)

# if `var.equal = FALSE`, then Games-Howell test will be run
pairwise_comparisons(
  data = mtcars,
  x = cyl,
  y = wt,
  type = "parametric",
  var.equal = FALSE,
  paired = FALSE,
  p.adjust.method = "bonferroni"
)

# non-parametric (Dunn test)
pairwise_comparisons(
  data = mtcars,
  x = cyl,
  y = wt,
  type = "nonparametric",
  paired = FALSE,
  p.adjust.method = "none"
)

# robust (Yuen's trimmed means t-test)
pairwise_comparisons(
  data = mtcars,
  x = cyl,
  y = wt,
  type = "robust",
  paired = FALSE,
  p.adjust.method = "fdr"
)

# Bayes Factor (Student's t-test)
pairwise_comparisons(
  data = mtcars,
  x = cyl,
  y = wt,
  type = "bayes",
  paired = FALSE
)

#------------------- within-subjects design ----------------------------

# parametric (Student's t-test)
pairwise_comparisons(
  data = bugs_long,
  x = condition,
  y = desire,
  subject.id = subject,
  type = "parametric",
  paired = TRUE,
  p.adjust.method = "BH"
)

# non-parametric (Durbin-Conover test)
pairwise_comparisons(
  data = bugs_long,
  x = condition,
  y = desire,
  subject.id = subject,
  type = "nonparametric",
  paired = TRUE,
  p.adjust.method = "BY"
)

# robust (Yuen's trimmed means t-test)
pairwise_comparisons(
  data = bugs_long,
  x = condition,
  y = desire,
  subject.id = subject,
  type = "robust",
  paired = TRUE,
  p.adjust.method = "hommel"
)

# Bayes Factor (Student's t-test)
pairwise_comparisons(
  data = bugs_long,
  x = condition,
  y = desire,
  subject.id = subject,
  type = "bayes",
  paired = TRUE
)

---

Pairwise contingency table analyses

Description

Pairwise Fisher's exact tests as post hoc tests for contingency table analyses, with effect sizes (Cramer's V) and p-value adjustment for multiple comparisons.

Usage

pairwise_contingency_table(
  data,
  x,
  y,
  counts = NULL,
  p.adjust.method = "holm",
  digits = 2L,
  conf.level = 0.95,
  alternative = "two.sided",
  ...
)

Arguments

data	A data frame (or a tibble) from which variables specified are to be taken. Other data types (e.g., matrix, table, array, etc.) will not be accepted. Additionally, grouped data frames from {dplyr} should be ungrouped before they are entered as data.
x	The variable to use as the rows in the contingency table.
y	The variable to use as the columns in the contingency table. Default is NULL. If NULL, one-sample proportion test (a goodness of fit test) will be run for the x variable.
counts	The variable in data containing counts, or NULL if each row represents a single observation.
p.adjust.method	Adjustment method for p-values for multiple comparisons. Possible methods are: "holm" (default), "hochberg", "hommel", "bonferroni", "BH", "BY", "fdr", "none".
digits	Number of digits for rounding or significant figures. May also be "signif" to return significant figures or "scientific" to return scientific notation. Control the number of digits by adding the value as suffix, e.g. digits = "scientific4" to have scientific notation with 4 decimal places, or digits = "signif5" for 5 significant figures (see also signif()).
conf.level	Scalar between 0 and 1 (default: ⁠95%⁠ confidence/credible intervals, 0.95).
alternative	A character string specifying the alternative hypothesis; Controls the type of CI returned: "two.sided" (default, two-sided CI), "greater" or "less" (one-sided CI). Partial matching is allowed (e.g., "g", "l", "two"...). See section One-Sided CIs in the effectsize_CIs vignette.
...	Additional arguments passed to stats::fisher.test().

Value

The returned object is a tibble data frame with the additional class "statsExpressions". The exact set of columns depends on the test and, for functions that accept a type argument, on the chosen analysis (parametric, non-parametric, robust, or Bayesian). Any given call therefore returns some (not all) of the columns below.

Hypothesis testing

• statistic: the numeric value of a statistic

• df and/or df.error: degrees of freedom for the test. A t-test uses df.error, a chi-squared test uses df, and a parametric ANOVA uses both (numerator df and denominator df.error)

• p.value: the p-value associated with the observed statistic (two-sided unless a one-sided alternative is requested, where supported)

• method: the name of the inferential statistical test

Effect size estimation

• effectsize: the name of the effect size

• estimate: estimated value of the effect size

• conf.level: the coverage level of the confidence/credible interval (e.g. 0.95); the interval itself spans conf.low to conf.high

• conf.low: lower bound for the effect size estimate

• conf.high: upper bound for the effect size estimate

• conf.method: method used to compute the confidence/credible interval

• conf.distribution: statistical distribution for the effect

Bayesian analysis (only when type = "bayes")

• bf10: Bayes factor for the alternative hypothesis relative to the null

• log_e_bf10: natural logarithm of the Bayes factor (present for most, but not all, Bayesian analyses)

• prior.distribution, prior.scale, prior.location: prior specification used to compute the Bayes factor and posterior estimates

Pairwise comparisons (for pairwise_comparisons() and pairwise_contingency_table())

• group1, group2: the two levels being compared

• p.adjust.method: the adjustment method used for multiple comparisons

• p.value.adj: the adjusted p-value; returned by pairwise_contingency_table(). Note that pairwise_comparisons() instead folds the adjusted value into p.value (and does not return a separate p.value.adj column)

Common columns

• n.obs: sample size used for the analysis: observations for unpaired tests, complete subjects/pairs for paired two-sample tests and ANOVAs, or studies for meta-analysis

• expression: a list-column of pre-formatted plotmath expressions; each element is a language object (not a character string) containing the statistical details, ready to be used in {ggplot2} (e.g. in labs() or annotate())

For a per-function, column-by-column breakdown of the output (and an explanation of the internal add_expression_col() engine that builds the expression column), see the Return value schema article. For more examples, see the Data frame outputs article.

Pairwise contingency table tests

Below is a summary of:

• the statistical test carried out for inferential statistics

• the type of effect size estimate and a measure of uncertainty for this estimate

• the functions used internally to compute these details

Hypothesis testing

Test	p-value adjustment?	Function used
Fisher's exact test	Yes	stats::fisher.test()

Effect size estimation

Effect size	CI available?	Function used
Cramer's V	Yes	effectsize::cramers_v()

Citation

Patil, I., (2021). statsExpressions: R Package for Tidy Dataframes and Expressions with Statistical Details. Journal of Open Source Software, 6(61), 3236, https://doi.org/10.21105/joss.03236

Examples

# for reproducibility
set.seed(123)

# pairwise Fisher's exact tests with Holm adjustment
pairwise_contingency_table(
  data = mtcars,
  x = cyl,
  y = am,
  p.adjust.method = "holm"
)

# with counts data and Bonferroni adjustment
pairwise_contingency_table(
  data = as.data.frame(Titanic),
  x = Class,
  y = Survived,
  counts = Freq,
  p.adjust.method = "bonferroni"
)

---

Expressions with statistics for tidy regression data frames

Description

Expressions with statistics for tidy regression data frames

Usage

tidy_model_expressions(
  data,
  statistic = NULL,
  digits = 2L,
  effsize.type = "omega",
  ...
)

Arguments

data	A tidy data frame from regression model object (see tidy_model_parameters()). It must contain a term column identifying each row.
statistic	Which statistic is to be displayed (either "t", "f", "z", or "chi") in the expression. This argument must be specified.
digits	Number of digits for rounding or significant figures. May also be "signif" to return significant figures or "scientific" to return scientific notation. Control the number of digits by adding the value as suffix, e.g. digits = "scientific4" to have scientific notation with 4 decimal places, or digits = "signif5" for 5 significant figures (see also signif()).
effsize.type	Type of effect size needed for parametric tests. The argument can be "eta" (partial eta-squared) or "omega" (partial omega-squared).
...	Currently ignored.

Details

When any of the necessary numeric column values (estimate, statistic, p.value) are missing, for these rows, a NULL is returned instead of an expression with empty strings.

Value

The input data frame (as a tibble) with an additional expression list-column.

Citation

Patil, I., (2021). statsExpressions: R Package for Tidy Dataframes and Expressions with Statistical Details. Journal of Open Source Software, 6(61), 3236, https://doi.org/10.21105/joss.03236

Examples

set.seed(123)

# t-statistic from a tidy linear model data frame
df <- tidy_model_parameters(lm(wt ~ am * cyl, mtcars))
tidy_model_expressions(df, statistic = "t")

# f-statistic (requires a data frame with `df`, `df.error`, and effect-size columns)
df_f <- tidy_model_parameters(
  aov(wt ~ cyl, mtcars),
  es_type = "omega",
  table_wide = TRUE
)
tidy_model_expressions(df_f, statistic = "f")
df_f_eta <- tidy_model_parameters(
  aov(wt ~ cyl, mtcars),
  es_type = "eta",
  table_wide = TRUE
)
tidy_model_expressions(df_f_eta, statistic = "f", effsize.type = "eta")

---

Convert {parameters} package output to {tidyverse} conventions

Description

Runs parameters::model_parameters() on a model object and standardizes the result to {broom}-style column names (e.g. estimate, std.error, conf.low, conf.high, statistic, df.error, p.value). Bayes factors are returned in a bf10 column, along with their natural logarithm in log_e_bf10. For Bayesian ANOVA designs ({BayesFactor} linear models), the estimate is replaced by the Bayesian R-squared.

Usage

tidy_model_parameters(model, ...)

Arguments

model

Statistical Model.

...

Arguments passed to or from other methods. Non-documented arguments are digits, p_digits, ci_digits and footer_digits to set the number of digits for the output. groups can be used to group coefficients. These arguments will be passed to the print-method, or can directly be used in print(), see documentation in print.parameters_model(). If s_value = TRUE, the p-value will be replaced by the S-value in the output (cf. Rafi and Greenland 2020). pd adds an additional column with the probability of direction (see bayestestR::p_direction() for details). Furthermore, see 'Examples' in model_parameters.default(). For developers, whose interest mainly is to get a "tidy" data frame of model summaries, it is recommended to set pretty_names = FALSE to speed up computation of the summary table.

Value

A tibble with one row per parameter.

Citation

Patil, I., (2021). statsExpressions: R Package for Tidy Dataframes and Expressions with Statistical Details. Journal of Open Source Software, 6(61), 3236, https://doi.org/10.21105/joss.03236

Examples

model <- lm(mpg ~ wt + cyl, data = mtcars)
tidy_model_parameters(model)

---

Two-sample tests

Description

Parametric, non-parametric, robust, and Bayesian two-sample tests.

Usage

two_sample_test(
  data,
  x,
  y,
  subject.id = NULL,
  type = "parametric",
  paired = FALSE,
  alternative = "two.sided",
  digits = 2L,
  conf.level = 0.95,
  effsize.type = "g",
  var.equal = FALSE,
  bf.prior = 0.707,
  tr = 0.2,
  nboot = 100L,
  exact = FALSE,
  ...
)

Arguments

data	A data frame (or a tibble) from which variables specified are to be taken. Other data types (e.g., matrix, table, array, etc.) will not be accepted. Additionally, grouped data frames from {dplyr} should be ungrouped before they are entered as data.
x	The grouping (or independent) variable from data. For repeated measures designs, see the note on ordering in subject.id.
y	The response (or outcome or dependent) variable from data.
subject.id	Relevant in case of a repeated measures or within-subjects design (i.e., paired = TRUE), it specifies the subject or repeated measures identifier. Important: If this argument is NULL (which is the default), observations are paired by their row order within each level of x (i.e., the data is assumed to be sorted in a subject-1, subject-2, ... pattern within every level). If the data is not sorted this way, the paired results will be silently incorrect, so it is safest to always specify subject.id. For paired designs, each subject must have at most one row per level of x (an error is thrown otherwise), and rows with a missing subject.id are dropped because they cannot be paired (see complete_subjects()).
type	A character specifying the type of statistical approach: "parametric" "nonparametric" "robust" "bayes" You can specify just the initial letter (e.g. "np" or "bf"). Matching is on the initial lowercase letter only, so any other value (including capitalized values such as "Bayes") falls back to "parametric" without a warning.
paired	Logical that decides whether the experimental design is repeated measures/within-subjects or between-subjects. The default is FALSE.
alternative	a character string specifying the alternative hypothesis, must be one of "two.sided" (default), "greater" or "less". You can specify just the initial letter.
digits	Number of digits for rounding or significant figures. May also be "signif" to return significant figures or "scientific" to return scientific notation. Control the number of digits by adding the value as suffix, e.g. digits = "scientific4" to have scientific notation with 4 decimal places, or digits = "signif5" for 5 significant figures (see also signif()).
conf.level	Scalar between 0 and 1 (default: ⁠95%⁠ confidence/credible intervals, 0.95).
effsize.type	Type of effect size needed for parametric tests. The argument can be "d" (for Cohen's d) or "g" (for Hedges' g).
var.equal	a logical variable indicating whether to treat the two variances as being equal. If TRUE then the pooled variance is used to estimate the variance otherwise the Welch (or Satterthwaite) approximation to the degrees of freedom is used.
bf.prior	A number between 0.5 and 2 (default 0.707), the prior width to use in calculating Bayes factors and posterior estimates. In addition to numeric arguments, several named values are also recognized: "medium", "wide", and "ultrawide", corresponding to r scale values of 1/2, sqrt(2)/2, and 1, respectively. In case of an ANOVA, this value corresponds to scale for fixed effects.
tr	Trim level for the mean when carrying out robust tests. In case of an error, try reducing the value of tr, which is by default set to 0.2. Lowering the value might help.
nboot	Number of bootstrap samples for computing confidence interval for the effect size (Default: 100L).
exact	A logical indicating whether exact p-values should be computed. Relevant only when type = "nonparametric" (Default: FALSE).
...	Currently ignored.

Value

The returned object is a tibble data frame with the additional class "statsExpressions". The exact set of columns depends on the test and, for functions that accept a type argument, on the chosen analysis (parametric, non-parametric, robust, or Bayesian). Any given call therefore returns some (not all) of the columns below.

Hypothesis testing

• statistic: the numeric value of a statistic

• df and/or df.error: degrees of freedom for the test. A t-test uses df.error, a chi-squared test uses df, and a parametric ANOVA uses both (numerator df and denominator df.error)

• p.value: the p-value associated with the observed statistic (two-sided unless a one-sided alternative is requested, where supported)

• method: the name of the inferential statistical test

Effect size estimation

• effectsize: the name of the effect size

• estimate: estimated value of the effect size

• conf.level: the coverage level of the confidence/credible interval (e.g. 0.95); the interval itself spans conf.low to conf.high

• conf.low: lower bound for the effect size estimate

• conf.high: upper bound for the effect size estimate

• conf.method: method used to compute the confidence/credible interval

• conf.distribution: statistical distribution for the effect

Bayesian analysis (only when type = "bayes")

• bf10: Bayes factor for the alternative hypothesis relative to the null

• log_e_bf10: natural logarithm of the Bayes factor (present for most, but not all, Bayesian analyses)

• prior.distribution, prior.scale, prior.location: prior specification used to compute the Bayes factor and posterior estimates

Pairwise comparisons (for pairwise_comparisons() and pairwise_contingency_table())

• group1, group2: the two levels being compared

• p.adjust.method: the adjustment method used for multiple comparisons

• p.value.adj: the adjusted p-value; returned by pairwise_contingency_table(). Note that pairwise_comparisons() instead folds the adjusted value into p.value (and does not return a separate p.value.adj column)

Common columns

• n.obs: sample size used for the analysis: observations for unpaired tests, complete subjects/pairs for paired two-sample tests and ANOVAs, or studies for meta-analysis

• expression: a list-column of pre-formatted plotmath expressions; each element is a language object (not a character string) containing the statistical details, ready to be used in {ggplot2} (e.g. in labs() or annotate())

For a per-function, column-by-column breakdown of the output (and an explanation of the internal add_expression_col() engine that builds the expression column), see the Return value schema article. For more examples, see the Data frame outputs article.

Two-sample tests

Below is a summary of:

• the statistical test carried out for inferential statistics

• the type of effect size estimate and a measure of uncertainty for this estimate

• the functions used internally to compute these details

Between-subjects

Hypothesis testing

Type	No. of groups	Test	Function used
Parametric	2	Student's or Welch's t-test	stats::t.test()
Non-parametric	2	Mann-Whitney U test	stats::wilcox.test()
Robust	2	Yuen's test for trimmed means	WRS2::yuen()
Bayesian	2	Student's t-test	BayesFactor::ttestBF()

Effect size estimation

Type	No. of groups	Effect size	CI available?	Function used
Parametric	2	Cohen's d, Hedges' g	Yes	effectsize::cohens_d(), effectsize::hedges_g()
Non-parametric	2	r (rank-biserial correlation)	Yes	effectsize::rank_biserial()
Robust	2	Algina-Keselman-Penfield robust standardized difference	Yes	WRS2::akp.effect()
Bayesian	2	difference	Yes	bayestestR::describe_posterior()

Within-subjects

Paired data: See complete_subjects() for subject matching, duplicate observations, and removal of incomplete subjects.

Hypothesis testing

Type	No. of groups	Test	Function used
Parametric	2	Student's t-test	stats::t.test()
Non-parametric	2	Wilcoxon signed-rank test	stats::wilcox.test()
Robust	2	Yuen's test on trimmed means for dependent samples	WRS2::yuend()
Bayesian	2	Student's t-test	BayesFactor::ttestBF()

Effect size estimation

Type	No. of groups	Effect size	CI available?	Function used
Parametric	2	Cohen's d, Hedges' g	Yes	effectsize::cohens_d(), effectsize::hedges_g()
Non-parametric	2	r (rank-biserial correlation)	Yes	effectsize::rank_biserial()
Robust	2	Algina-Keselman-Penfield robust standardized difference	Yes	WRS2::wmcpAKP()
Bayesian	2	difference	Yes	bayestestR::describe_posterior()

Citation

Patil, I., (2021). statsExpressions: R Package for Tidy Dataframes and Expressions with Statistical Details. Journal of Open Source Software, 6(61), 3236, https://doi.org/10.21105/joss.03236

Examples

# ----------------------- within-subjects -------------------------------------

# data
df <- dplyr::filter(bugs_long, condition %in% c("LDLF", "LDHF"))

# for reproducibility
set.seed(123)

# ----------------------- parametric ---------------------------------------

two_sample_test(
  df,
  condition,
  desire,
  subject.id = subject,
  paired = TRUE,
  type = "parametric"
)

# ----------------------- non-parametric -----------------------------------

two_sample_test(
  df,
  condition,
  desire,
  subject.id = subject,
  paired = TRUE,
  type = "nonparametric"
)

# ----------------------- robust --------------------------------------------

two_sample_test(
  df,
  condition,
  desire,
  subject.id = subject,
  paired = TRUE,
  type = "robust"
)

# ----------------------- Bayesian ---------------------------------------

two_sample_test(
  df,
  condition,
  desire,
  subject.id = subject,
  paired = TRUE,
  type = "bayes"
)

# ----------------------- between-subjects -------------------------------------

# for reproducibility
set.seed(123)

# ----------------------- parametric ---------------------------------------

# unequal variance
two_sample_test(ToothGrowth, supp, len, type = "parametric")

# equal variance
two_sample_test(ToothGrowth, supp, len, type = "parametric", var.equal = TRUE)

# biased (Cohen's d) effect size
two_sample_test(ToothGrowth, supp, len, type = "parametric", effsize.type = "d")

# ----------------------- non-parametric -----------------------------------

two_sample_test(ToothGrowth, supp, len, type = "nonparametric")

# ----------------------- robust --------------------------------------------

two_sample_test(ToothGrowth, supp, len, type = "robust")

# ----------------------- Bayesian ---------------------------------------

two_sample_test(ToothGrowth, supp, len, type = "bayes")

---