`vannstats`: simplifying statistical procedures in the social sciences

Overview

vannstats is an R package designed to simplify the statistical workflows commonly used in undergraduate and graduate-level social science courses — particularly those transitioning from SPSS to R. The package provides intuitive wrappers for normality diagnostics, descriptive statistics, bivariate analyses, and multivariate techniques, producing output that mirrors the look and feel of SPSS results.

Whether you are assessing distributional assumptions, running crosstabs, computing correlations, or fitting ANOVA models, vannstats reduces the complexity of base R syntax into concise, readable function calls.

Installation

Install the released version from CRAN:

install.packages("vannstats")

Then load the package:

library(vannstats)

Functions

Diagnostic Plots

Function	Description
`hst()`	Histogram — can be parsed by up to 2 grouping variables
`box()`	Box-and-Whisker plot — can be parsed by up to 2 grouping variables
`qq()`	Q-Q (Quantile-Quantile) plot — can be parsed by up to 2 grouping variables
`residplot()`	Residuals plot for a regression equation
`scatter()`	Scatterplot with an overlaid fit line
`bar.chart()`	Bar chart for a given variable
`chiplot()`	Standardized residuals plot for chi square differences

Descriptive Statistics

Function	Description
`univ.desc()`	Univariate descriptive statistics, optionally parsed by a grouping variable
`mode()`	Mode of a variable
`z.calc()`	Z-score for a given value relative to a variable’s mean and SD
`se()`	Standard error of a variable
`ci()`	Confidence interval for a variable at a specified confidence level

Bivariate Comparisons & Analyses

Function	Description
`tab()`	Crosstab of observed and expected frequencies for two categorical variables
`chi.sq()`	Pearson’s Chi-Square test (with optional post-hoc comparisons and Cramér’s V)
`p.corr()`	Pearson’s Product-Moment Correlation Coefficient
`cormat()`	Correlation matrix for variables in a formula
`ow.anova()`	One-Way ANOVA (with optional Tukey’s HSD post-hoc test and means plot)
`rm.anova()`	One-Way Repeated Measures ANOVA (with sphericity correction and test)

Data Management

Function	Description
`dummy()`	Creates dummy-coded columns for the values of a categorical variable
`revcode()`	Reverse-codes a scale variable
`stata.plm.margins()`	Predictive margins from a panel linear model, mirroring STATA’s `margins` output

Examples

Histogram

data <- mtcars
hst(data, mpg, cyl)

Box-and-Whisker Plot

data <- mtcars
box(data, mpg, cyl)

Q-Q Plot

data <- mtcars
qq(data, mpg, cyl)

Chi Square Plot

data <- mtcars
x2 <- chi.sq(data, vs, am, post=TRUE)
chiplot(zmat = x2$z_mat, p.mat = x2$p_z_mat, z.crit = round(x2$z_crit, 4), outline.color = "white", lab = TRUE, digits = 3)

Crosstab

data <- mtcars
tab(data, vs, am)

Chi-Square Test

data <- mtcars
x2 <- chi.sq(data, vs, am)
summary(x2)

Pearson’s Correlation

data <- mtcars
p.corr(data, mpg, wt)

Correlation Matrix

data <- mtcars
cormat(data, mpg ~ wt + am)

One-Way ANOVA

data <- mtcars
ow <- ow.anova(data, mpg, cyl)
summary(ow)

Repeated Measures ANOVA

# Wide format
data <- howell_aids_wide
rm.anova(data, student, c("t1", "t2", "t3"))

# Long format
data2 <- howell_aids_long
rm.anova(data2, student, time, scores = knowledge)

Residuals Plot

data <- mtcars
residplot(data, mpg ~ wt + cyl)

Included Datasets

vannstats ships with several built-in datasets useful for teaching and demonstration purposes:

Dataset	Description
`GSS2024`	General Social Survey, 2024 — individual survey responses on a range of social topics (NORC at the University of Chicago)
`GSS2022`	General Social Survey, 2022 — individual survey responses on a range of social topics (NORC at the University of Chicago)
`GSS2014`	General Social Survey, 2014 — individual survey responses on a range of social topics (NORC at the University of Chicago)
`UCR2015`	Uniform Crime Reports, 2015 — county-level crime data
`howell_aids_wide`	Howell Student AIDS Knowledge Data in wide format (3 time points)
`howell_aids_long`	Howell Student AIDS Knowledge Data in long format
`Defendants2025`	Simulated data representing individual defendants at the Richard J. Donovan Correctional Facility, San Diego, CA (1,738 observations, 11 variables)

Dependencies

vannstats builds on a suite of well-established R packages including ggplot2, dplyr, tidyverse, car, rstatix, plm, purrr, gplots, rlang, stringr, MASS, formula.tools, gdata, and stats.

vannstats: simplifying statistical procedures in the social sciences