Generate Statistical Tables

Generate Statistical Tables

Create publication-ready statistical tables for reports and manuscripts.

When to Use

• Creating descriptive statistics tables
• Formatting regression or ANOVA output
• Building correlation matrices
• Producing APA-style tables for academic papers
• Generating tables for Quarto/R Markdown documents

Inputs

• Required: Statistical analysis results (model objects, summary data)
• Required: Output format (HTML, PDF, Word)
• Optional: Style guide (APA, journal-specific)
• Optional: Table numbering scheme

Procedure

Step 1: Choose Table Package

Expected: A table package selected based on the output format and use case. The chosen package is installed and loadable.

On failure: If the required package is not installed, run install.packages("gt") (or the appropriate package). For gtsummary, both gt and gtsummary must be installed.

Step 2: Descriptive Statistics Table

library(gt)

descriptives <- data |>
  group_by(group) |>
  summarise(
    n = n(),
    M = mean(score, na.rm = TRUE),
    SD = sd(score, na.rm = TRUE),
    Min = min(score, na.rm = TRUE),
    Max = max(score, na.rm = TRUE)
  )

gt(descriptives) |>
  tab_header(
    title = "Table 1",
    subtitle = "Descriptive Statistics by Group"
  ) |>
  fmt_number(columns = c(M, SD), decimals = 2) |>
  fmt_number(columns = c(Min, Max), decimals = 1) |>
  cols_label(
    group = "Group",
    n = md("*n*"),
    M = md("*M*"),
    SD = md("*SD*")
  )

Expected: A gt table object with formatted means, SDs, and counts grouped by category. Column headers use proper statistical notation (italicized M, SD, n).

On failure: If group_by() produces unexpected results, verify the grouping variable exists and has the expected levels. If fmt_number() throws an error, ensure the target columns contain numeric data.

Step 3: Regression Results Table

model <- lm(outcome ~ predictor1 + predictor2 + predictor3, data = data)

library(gtsummary)

tbl_regression(model) |>
  bold_p() |>
  add_glance_source_note(
    include = c(r.squared, adj.r.squared, nobs)
  ) |>
  modify_header(label = "**Predictor**") |>
  modify_caption("Table 2: Regression Results")

Expected: A gtsummary regression table with bolded p-values, model fit statistics (R-squared, N) in a source note, and a descriptive caption.

On failure: If tbl_regression() fails, verify the input is a model object (e.g., lm, glm). If add_glance_source_note() errors, check that broom can tidy the model: broom::glance(model).

Step 4: Correlation Matrix

library(gt)

cor_matrix <- cor(data[, c("var1", "var2", "var3", "var4")],
                  use = "pairwise.complete.obs")

# Format lower triangle
cor_matrix[upper.tri(cor_matrix)] <- NA

as.data.frame(cor_matrix) |>
  tibble::rownames_to_column("Variable") |>
  gt() |>
  fmt_number(decimals = 2) |>
  sub_missing(missing_text = "") |>
  tab_header(title = "Table 3", subtitle = "Correlation Matrix")

Expected: A lower-triangle correlation matrix rendered as a gt table with blanked upper triangle, two decimal places, and a clear caption.

On failure: If sub_missing() does not blank the upper triangle, verify that NA values were set correctly with cor_matrix[upper.tri(cor_matrix)] <- NA. If variables are non-numeric, cor() will fail; filter to numeric columns first.

Step 5: ANOVA Table

aov_result <- aov(score ~ group * condition, data = data)

library(gtsummary)

tbl_anova <- broom::tidy(aov_result) |>
  gt() |>
  fmt_number(columns = c(sumsq, meansq, statistic), decimals = 2) |>
  fmt_number(columns = p.value, decimals = 3) |>
  cols_label(
    term = "Source",
    df = md("*df*"),
    sumsq = md("*SS*"),
    meansq = md("*MS*"),
    statistic = md("*F*"),
    p.value = md("*p*")
  ) |>
  tab_header(title = "Table 4", subtitle = "ANOVA Results")

Expected: A formatted ANOVA table with Source, df, SS, MS, F, and p columns. Interaction terms are clearly labeled and p-values are formatted to three decimal places.

On failure: If broom::tidy(aov_result) produces unexpected columns, verify the model is an aov object. For Type III sums of squares, use car::Anova(model, type = 3) instead of base aov().

Step 6: Save Tables

# Save as HTML
gtsave(my_table, "table1.html")

# Save as Word
gtsave(my_table, "table1.docx")

# Save as PNG image
gtsave(my_table, "table1.png")

# For LaTeX/PDF (kableExtra)
kableExtra::save_kable(kable_table, "table1.pdf")

Expected: Table saved to the specified file format (HTML, Word, PNG, or PDF). The output file opens correctly in the appropriate application.

On failure: If gtsave() fails for Word format, ensure the webshot2 package is installed. For PDF output via kableExtra, ensure a LaTeX distribution (TinyTeX or MiKTeX) is installed.

Step 7: Embed in Quarto Document

```{r}
#| label: tbl-descriptives
#| tbl-cap: "Descriptive Statistics by Group"

gt(descriptives) |>
  fmt_number(columns = c(M, SD), decimals = 2)
```

See @tbl-descriptives for summary statistics.

Expected: The table renders inline in the Quarto document with a cross-referenceable label (@tbl-*) and a proper caption. The table adapts to the document's output format automatically.

On failure: If the table does not render, verify the chunk label starts with tbl- for Quarto cross-referencing. If formatting is lost in PDF, switch from gt to kableExtra for LaTeX-based output.

Validation

• Table renders correctly in target format (HTML, PDF, Word)
• Numbers are formatted consistently (decimal places, alignment)
• Statistical notation follows the style guide (italicized, proper symbols)
• Table has a clear caption and numbering
• Column headers are meaningful
• Notes/footnotes explain abbreviations or significance markers

Common Pitfalls

• gt in PDF: gt has limited PDF support. Use kableExtra for LaTeX-heavy documents.
• Rounding inconsistency: Always use fmt_number() (gt) or format() rather than round() for display
• Missing values display: Configure with sub_missing() in gt or options(knitr.kable.NA = "")
• Wide tables in PDF: Tables exceeding page width need landscape() or font size reduction
• APA number formatting: No leading zero for values bounded by 1 (p-values, correlations): ".03" not "0.03"

Related Skills

• format-apa-report - tables within APA manuscripts
• create-quarto-report - embedding tables in reports
• build-parameterized-report - tables that adapt to parameters