table_stack() sets regression tables one under the other. Each table
becomes a panel, opened by a row with its label, and keeps its own
columns, rows, fit statistics and number formats: the cells of a panel
have the content of the table passed, in every engine. The panels
share one column header when they have the same columns, and each opens
with its own header otherwise. Their notes are merged into one.
Build each table with its own arguments (vcov, cluster,
exponentiate, p_adjust, show_columns, ...), then stack them: a
grid of models is table_stack(table_regression(list(m1, m2)), table_regression(list(m3, m4))). For one panel per model of a list,
table_regression(models, layout = "stacked") is the shortcut.
Usage
table_stack(
...,
panel_labels = NULL,
title = NULL,
note = NULL,
note_add = NULL,
output = c("default", "data.frame", "long", "tinytable", "gt", "flextable", "excel",
"clipboard", "word"),
excel_path = NULL,
excel_sheet = NULL,
clipboard_delim = "\t",
word_path = NULL,
word_template = NULL
)Arguments
- ...
Tables returned by
table_regression()ortable_regression_uv()withoutput = "default", one per panel, in the order of the panels. A name given here labels its panel:table_stack(Women = t1, Men = t2). A list of tables is spliced with!!!:table_stack(!!!tables).- panel_labels
Character vector, one non-empty label per table, in the order of the tables, unnamed: the labels of the panels not named in
...(a name given there takes precedence, with a warning).NULL(default) labels such a panel by the title you typed for its table, else by its outcome when the panels have different outcomes, else "Panel k". Two panels cannot share a label.- title
Title of the stacked table.
NULL(default) keeps the title of the tables when they all have the same one, and gives none otherwise;FALSEgives none; a string sets it.- note
NULL(default) merges the notes of the panels;FALSEremoves the note; a string replaces the general note, printed after the lead-in ("Note."), with the legend of the stars kept below it.- note_add
Sentences added at the end of the general note.
- output
Output format, as in
table_regression():"default"(a printable data frame),"data.frame","long","tinytable","gt","flextable","excel","clipboard"or"word".- excel_path
File path for
output = "excel", ending in".xlsx", in a directory that exists. DefaultNULL(required whenoutput = "excel").- excel_sheet
Sheet name when writing to Excel.
NULL(the default) uses"Regression".- clipboard_delim
Field delimiter for
output = "clipboard". Default"\t"(tab-separated, pastes cleanly into Excel / Google Sheets / Word). The clipboard payload mirrors the Excel layout (title row, spanner row, header, body, footer note) but is plain text – horizontal rules, cell merging, decimal alignment and the monospace font cannot be encoded in delimited text and are therefore absent from the paste. A cell holding the delimiter itself (a label with a comma underclipboard_delim = ",", or any number underdecimal_mark = ","), a double quote or a line break is quoted RFC 4180-style, so the grid survives whatever delimiter you choose.Paste behavior by target:
Excel / Google Sheets: numerics are auto-detected and right-aligned; text cells stay left-aligned. (P-values such as
.005get re-parsed as0.005by Excel's auto-format – to preserve the APA leading-zero-dropped display, preferoutput = "excel".)Word: the paste is converted to a Word table; all cells start left-aligned. Apply a Table Style (Insert > Table > Design) for APA-style borders, and set right-alignment on numeric columns (Layout > Align Right). For a self-contained Word file with borders and alignment pre-applied, use
output = "word"instead.
- word_path
File path for
output = "word", ending in".docx", in a directory that exists. DefaultNULL(required whenoutput = "word"). The Word table inherits the flextable styling (Calibri font, APA borders, decimal-aligned numerics) and adds Word-specific features: an auto-numbered caption ("Table 1: ...", "Table 2: ...") via Word'sSEQfield so multipletable_regression()calls in one document number consecutively; a re-printed header row on each page break; row split prevention so a single coefficient row never wraps across two pages; and an APA-styled note line (*Note.*italic prefix per APA Manual 7 §7.14).R Markdown / Quarto: for embedded use, prefer
output = "flextable"(returns the flextable object that knits to docx/HTML/PDF natively).output = "word"writes a standalone .docx file, suited to scripted exports rather than chunk-level rendering.- word_template
Optional path to a custom .docx file used as the template for
output = "word". The template's header, footer, page size, margins, and named styles ("Table Caption" in particular) are honored; the table is appended to the template body. Useful for institutional templates with pre-set headers ("APA Style", "Manuscript Submission Template", etc.). DefaultNULL(uses flextable's stock template).Customizing the caption appearance: the table caption is tagged with the Word named style
"Table Caption". The visual rendering (italic / bold / color / font) follows whatever that style is set to in the docx template. The stock Word template renders"Table Caption"in italic — the APA Manual 7 §7.10 condensed convention. For a different appearance (Nature-style bold non-italic, APA-strict 2-line bold-number / italic-title, etc.), edit the"Table Caption"style in a docx template and pass it viaword_template = "your_template.docx". Style-based delegation keeps the rendered caption consistent with the surrounding document and lets editorial conventions (Nature, APA-strict, journal-specific) be applied without modifying the call site.
Value
With output = "default", a spicy_stack_table (classes
c("spicy_stack_table", "spicy_table", "data.frame")): the cells as
printed, one row per printed row, plus a blank row for the bounds of
an interval in a panel's header. Its attributes panels (the tables
stacked, named by their labels), panel_rows (the rows of each panel
in it, its label and header rows excluded), title and note are
documented; the others are internal. as_structured() returns the
typed views of the panels, as a named list.
With output = "data.frame" (and from as.data.frame()), the data
frames of the panels bound by rows; with output = "long" (and from
broom::tidy() and broom::glance()), the outputs of their models.
Each opens with a panel column holding the label of the panel; the
columns are the union of the panels' and the note, as printed, is the
note attribute. In the output of table_stack(), a row is keyed by
panel, model_id (the model inside its panel), term and
estimate_type. With table_regression(layout = "stacked"), these
three outputs are those of the same models side by side, panel
aside: one row per (model_id, term, estimate_type), model_id
being the identifier of the model in the list. With "tinytable",
"gt" or "flextable", the object of that package; with "excel"
or "word", the path of the file, invisibly; with "clipboard", the
stacked table, invisibly.
Panels
A row with the label of the panel opens each panel ("Women:"), in the
typography of the block headers. The label of a panel is its name in
..., else its entry in panel_labels, else the title you typed for
its table (title = "..."), else its outcome when every panel has an
outcome of its own, else "Panel k". A title spicy built is not a label:
it describes the model ("Linear regression: bmi"), not the panel. The
rows of a panel follow, in the order of its table: coefficients, blocks
(thresholds, random effects, components), fit statistics.
When every panel has the same column headers, one header heads the
whole table. Otherwise the header of the first panel heads it and each
later panel opens with its own header rows, below its label row:
an exponentiate = TRUE glm panel reads "OR" where an lm panel reads
"B", and a multinomial panel keeps its outcome categories in columns.
A panel narrower than the widest one leaves the extra columns blank.
The header of a panel has the rows of the header of its table: its
column groups, ruled under, its labels, and the bounds of its interval
("LL", "UL"), which the console leaves out since it prints an interval
in one column.
In the console, the columns are matched by position and every panel shares their widths: a panel with other columns can widen a column of the first. When the panels have the same columns and decimals, the decimals line up from one panel to the next, only the padding of a cell changing; otherwise each panel keeps the padding of its table. When the console splits a wide stack by width, a column group cut by the split keeps its label over each part, and a continuation leaves out the panels with no cell in it. The other engines write each cell as its table does.
A table of table_regression_uv() is a panel like any other, its
univariable models and their own sample sizes included.
Notes
The notes are merged sentence by sentence. The type of each model and
its standard errors are worded as in a table of the models side by
side: "Linear regression models.", or "Women: linear regression; Men:
logistic regression." (in English; another language gives each panel
its own sentence, as below), and "Std. errors: classical (OLS)." or
one estimator per panel. So is an adjustment of the p-values ("P-values
adjusted within each model ..."). Any other sentence that every panel
carries is printed once, first; one that only some panels carry follows,
opening with their labels ("Women, Men: ..."), panel after panel. The
abbreviations are defined once. The legend of the stars is printed
once, and when panels have different legends, each opens with the
labels of its panels. A panel built with note = FALSE contributes no
sentence, and one built with a typed note contributes that note.
Compatibility
The panels must be built in the same language
(options(spicy.language)), with the same decimal_mark and align,
and under styles that format a cell alike: the stacked table has one of
each. Two styles that print the same table are compatible (no style
and style = "apa"); a lever that differs (the decimals of a p-value,
the separator or brackets of an interval, the minus sign, the lead-in
of the note, the abbreviations) is not, nor is p_digits typed for one
table of a theme that sets its own p-values. Otherwise table_stack()
stops with spicy_stack_mismatch, naming what differs.
Only the tables of table_regression() and table_regression_uv()
can be panels for now; the descriptive tables will follow.
Classed conditions
Errors carry spicy_invalid_input (no table, an object that is not a
table built with output = "default", a stack passed as a panel, a
table whose rows were changed after it was built, invalid or
duplicated labels), spicy_empty_table (a table without rows),
spicy_not_implemented with spicy_unsupported (a descriptive table)
and spicy_stack_mismatch with spicy_invalid_input (panels built in
different languages, with different decimal marks or alignments, or
under styles that format a cell differently). Warnings carry
spicy_ignored_arg (an entry of panel_labels overridden by a name in
...).
See also
table_regression(), whose layout = "stacked" stacks one
panel per model; as_structured().
Examples
women <- sochealth[sochealth$sex == "Female", ]
men <- sochealth[sochealth$sex == "Male", ]
table_stack(
Women = table_regression(lm(bmi ~ age + smoking, data = women)),
Men = table_regression(lm(bmi ~ age + smoking, data = men))
)
#> Linear regression: bmi
#>
#> Variable │ B SE 95% CI p
#> ─────────────┼─────────────────────────────────────
#> Women: │
#> (Intercept) │ 24.04 0.53 [23.00, 25.09] <.001
#> age │ 0.03 0.01 [ 0.01, 0.05] <.001
#> smoking: │
#> No (ref.) │ – – – –
#> Yes │ -0.18 0.37 [-0.90, 0.55] .637
#> ╌╌╌╌╌╌╌╌╌╌╌╌╌┼╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌
#> n │ 602
#> R² │ 0.02
#> Adj. R² │ 0.02
#> ╌╌╌╌╌╌╌╌╌╌╌╌╌┼╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌
#> Men: │
#> (Intercept) │ 24.41 0.55 [23.33, 25.48] <.001
#> age │ 0.04 0.01 [ 0.02, 0.06] <.001
#> smoking: │
#> No (ref.) │ – – – –
#> Yes │ 0.07 0.38 [-0.67, 0.81] .850
#> ╌╌╌╌╌╌╌╌╌╌╌╌╌┼╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌
#> n │ 561
#> R² │ 0.02
#> Adj. R² │ 0.02
#>
#> Note. Linear regression models. Std. errors: classical (OLS). CI = confidence interval.
# Panels with their own columns: the second opens with its header.
fit_bmi <- lm(bmi ~ age + sex, data = sochealth)
fit_dentist <- glm(dentist_12m ~ age + sex, data = sochealth,
family = binomial)
table_stack(
BMI = table_regression(fit_bmi),
Dentist = table_regression(fit_dentist, exponentiate = TRUE)
)
#> Variable │ B SE 95% CI p
#> ─────────────────┼──────────────────────────────────────
#> BMI: │
#> (Intercept) │ 23.95 0.39 [23.18, 24.71] <.001
#> age │ 0.04 0.01 [ 0.02, 0.05] <.001
#> sex: │
#> Female (ref.) │ – – – –
#> Male │ 0.52 0.21 [ 0.10, 0.94] .015
#> ╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌┼╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌
#> n │ 1188
#> R² │ 0.02
#> Adj. R² │ 0.02
#> ╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌┼╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌
#> Dentist: │
#> │ OR SE 95% CI p
#> ╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌┼╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌
#> (Intercept) │ 1.56 0.36 [ 1.00, 2.44] .052
#> age │ 1.01 0.00 [ 1.00, 1.02] .044
#> sex: │
#> Female (ref.) │ – – – –
#> Male │ 1.00 0.13 [ 0.78, 1.29] .978
#> ╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌┼╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌
#> n │ 1200
#> R² (McFadden) │ 0.00
#> R² (Nagelkerke) │ 0.00
#> AIC │ 1457.7
#>
#> Note. BMI: linear regression; Dentist: logistic regression. Dentist: coefficients exponentiated and displayed as OR; SE on the OR scale (delta method); CI bounds exponentiated (asymmetric). Dentist: the (Intercept) row estimates the baseline odds. Std. errors: BMI: classical (OLS), Dentist: classical (Fisher information). CI = confidence interval; AIC = Akaike information criterion.
# One panel per model.
table_regression(list(BMI = fit_bmi, Dentist = fit_dentist),
layout = "stacked")
#> Regression comparison
#>
#> Variable │ B SE 95% CI p
#> ─────────────────┼──────────────────────────────────────
#> BMI: │
#> (Intercept) │ 23.95 0.39 [23.18, 24.71] <.001
#> age │ 0.04 0.01 [ 0.02, 0.05] <.001
#> sex: │
#> Female (ref.) │ – – – –
#> Male │ 0.52 0.21 [ 0.10, 0.94] .015
#> ╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌┼╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌
#> n │ 1188
#> R² │ 0.02
#> Adj. R² │ 0.02
#> ╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌┼╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌
#> Dentist: │
#> (Intercept) │ 0.44 0.23 [-0.00, 0.89] .052
#> age │ 0.01 0.00 [ 0.00, 0.02] .044
#> sex: │
#> Female (ref.) │ – – – –
#> Male │ 0.00 0.13 [-0.25, 0.25] .978
#> ╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌┼╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌╌
#> n │ 1200
#> R² (McFadden) │ 0.00
#> R² (Nagelkerke) │ 0.00
#> AIC │ 1457.7
#>
#> Note. BMI: linear regression; Dentist: logistic regression. Std. errors: BMI: classical (OLS), Dentist: classical (Fisher information). CI = confidence interval; AIC = Akaike information criterion.