Tables

Introduction

First, load {tidyverse} and {papaja} and import the data.

library(tidyverse)
── Attaching core tidyverse packages ──────────────────────── tidyverse 2.0.0 ──
✔ dplyr     1.1.4     ✔ readr     2.1.5
✔ forcats   1.0.0     ✔ stringr   1.5.1
✔ ggplot2   3.5.0     ✔ tibble    3.2.1
✔ lubridate 1.9.3     ✔ tidyr     1.3.1
✔ purrr     1.0.2     
── Conflicts ────────────────────────────────────────── tidyverse_conflicts() ──
✖ dplyr::filter() masks stats::filter()
✖ dplyr::lag()    masks stats::lag()
ℹ Use the conflicted package (<http://conflicted.r-lib.org/>) to force all conflicts to become errors
library(papaja)
Loading required package: tinylabels
hai <- read_csv("https://decisionslab.unl.edu/data/thayer_stevens_2020_data1.csv") |>
  select(experiment, participant, condition, gender,
         panas_pre = panas_pre_pos, panas_post = panas_post_pos)
Rows: 157 Columns: 38
── Column specification ────────────────────────────────────────────────────────
Delimiter: ","
chr   (4): condition, gender, race, parent_income
dbl  (33): experiment, participant, age_num, pas, pets_now, pets_child, dog_...
dttm  (1): date

ℹ Use `spec()` to retrieve the full column specification for this data.
ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.
hai_long <- hai |>
  pivot_longer(contains("panas"), names_to = "prepost", values_to = "panas") |>
  mutate(condition = fct_recode(condition, "Control" = "control", "HAI" = "hai"),
         prepost = fct_recode(prepost, "Pre" = "panas_pre", "Post" = "panas_post"),
         prepost = fct_relevel(prepost, c("Pre", "Post")))

Now let’s build a data frame that will be our table.

(condition_prepost_means <- hai_long |>
  group_by(condition, prepost) |>
  summarise(across(starts_with("panas"), ~ mean(.x, na.rm = TRUE))))
`summarise()` has grouped output by 'condition'. You can override using the
`.groups` argument.
# A tibble: 4 × 3
# Groups:   condition [2]
  condition prepost panas
  <fct>     <fct>   <dbl>
1 Control   Pre      2.99
2 Control   Post     2.76
3 HAI       Pre      2.97
4 HAI       Post     3.23

Tables by {knitr}

The {knitr} package uses the kable() function to format tables.

library(knitr)
kable(condition_prepost_means)
condition prepost panas
Control Pre 2.989873
Control Post 2.760759
HAI Pre 2.971795
HAI Post 3.232051

Column and row names

You can control column names and row names with col.names and row.names.

kable(condition_prepost_means, 
      col.names = c("Condition", "Prepost", "Mean PANAS"))
Condition Prepost Mean PANAS
Control Pre 2.989873
Control Post 2.760759
HAI Pre 2.971795
HAI Post 3.232051

Column alignment

By default, character columns are left aligned and numeric columns are right aligned. You can set alignment manually with the align argument with l = left, c = center, and r = right. You can just pass a character string with a series of those letters.

kable(condition_prepost_means,
      col.names = c("Condition", "Prepost", "Mean PANAS"), 
      align = "rcl")
Condition Prepost Mean PANAS
Control Pre 2.989873
Control Post 2.760759
HAI Pre 2.971795
HAI Post 3.232051

Digit rounding

Round the digits for all numeric data columns with digits argument.

kable(condition_prepost_means,
      col.names = c("Condition", "Prepost", "Mean PANAS"), 
      digits = 2)
Condition Prepost Mean PANAS
Control Pre 2.99
Control Post 2.76
HAI Pre 2.97
HAI Post 3.23

If you want different digits for different columns, you can pass a vector to the digits argument.

condition_prepost_means |> 
  pivot_wider(id_cols = condition, names_from = prepost, values_from = panas) |> 
  kable(digits = c(0, 2, 3))
condition Pre Post
Control 2.99 2.761
HAI 2.97 3.232

Table titles

Add a title to the table with the caption argument. The good news is that we can cross-reference easily (Table @ref(tab:title-table)). The bad news is that with captions, tables in PDFs are automatically placed at the top of the page. We’ll see how to fix this later.

kable(condition_prepost_means,
      col.names = c("Condition", "Prepost", "Mean PANAS"), 
      caption = "PANAS scores by condition and prepost")
PANAS scores by condition and prepost
Condition Prepost Mean PANAS
Control Pre 2.989873
Control Post 2.760759
HAI Pre 2.971795
HAI Post 3.232051

Supplementing kable with {kableExtra}

The kable() function is intentionally simple to use and therefore does not have a lot of additional functionality. The {kableExtra} package supplements the kable() functionality with additional formatting options by adding additional functions after the kable() function call using the |> pipe (a bit like how ggplot() works). Check out Create Awesome LaTeX Table with knitr::kable and kableExtra.

# install.packages("kableExtra")
library(kableExtra)

Attaching package: 'kableExtra'
The following object is masked from 'package:dplyr':

    group_rows

General styling

The kable_styling() function formats a number of things such as font size, table width, and table alignment. I’ll also add latex_options = "hold_position" to keep the table in the text. Otherwise, it puts it at the top of the page.

kable(condition_prepost_means,
      caption = "PANAS scores by condition and prepost",
      col.names = c("Condition", "Prepost", "Mean PANAS"), 
      booktabs = TRUE) |>
  kable_styling(font_size = 18, latex_options = "hold_position")
PANAS scores by condition and prepost
Condition Prepost Mean PANAS
Control Pre 2.989873
Control Post 2.760759
HAI Pre 2.971795
HAI Post 3.232051

Labels spanning rows

If you want to label groups of rows, use pack_rows(). Let’s get rid of the condition column and label the conditions explicitly.

condition_prepost_means2 <- condition_prepost_means |>
  ungroup(condition) |>
  select(-condition)
kable(condition_prepost_means2, 
      booktabs = TRUE)
prepost panas
Pre 2.989873
Post 2.760759
Pre 2.971795
Post 3.232051
kable(condition_prepost_means2, 
      booktabs = TRUE,
      col.names = c("", "Mean PANAS")) |>
  pack_rows("Control", start_row = 1, end_row = 2) |>
  pack_rows("HAI", start_row = 3, end_row = 4)
Mean PANAS
Control
Pre 2.989873
Post 2.760759
HAI
Pre 2.971795
Post 3.232051

Notice that we removed the first column name with "".

Labels spanning columns

You can labels groups of columns with the add_header_above() function. Let’s rearrange the data into wide format to illustrate this.

(wide_means <- condition_prepost_means |>
   unite(cond_prepost, condition:prepost) |>
   pivot_wider(names_from = cond_prepost, values_from = panas))
# A tibble: 1 × 4
  Control_Pre Control_Post HAI_Pre HAI_Post
        <dbl>        <dbl>   <dbl>    <dbl>
1        2.99         2.76    2.97     3.23
kable(wide_means, booktabs = TRUE)
Control_Pre Control_Post HAI_Pre HAI_Post
2.989873 2.760759 2.971795 3.232051

Now that the data are in wide format, we can add the column names by repeating Pre and Post then add the headers.

kable(wide_means, 
      booktabs = TRUE, 
      col.names = rep(c("Pre", "Post"), 2),
      digits = 2) |>
  add_header_above(c("Control" = 2, "HAI" = 2))
Control
HAI
Pre Post Pre Post
2.99 2.76 2.97 3.23

Maybe we need a column stating this is Mean PANAS.

wide_means2 <- wide_means |> 
  mutate(label = "Mean PANAS", .before = 1)
kable(wide_means2, 
      booktabs = TRUE, 
      col.names = c("", rep(c("Pre", "Post"), 2)),
      digits = 2) |>
  add_header_above(c("", "Control" = 2, "HAI" = 2))
Control
HAI
Pre Post Pre Post
Mean PANAS 2.99 2.76 2.97 3.23

Table footnotes

Add table notes with the footnote() function.

kable(condition_prepost_means, 
      booktabs = TRUE,
      caption = "PANAS scores by condition and prepost",
      col.names = c("Condition*", "Prepost", "Mean PANAS")) |>
  kable_styling(latex_options = "hold_position") |>
  footnote(general = "Source: Thayer & Stevens (2021)",
           symbol = "73 control participants, 72 HAI participants",
           footnote_as_chunk = TRUE)
PANAS scores by condition and prepost
Condition* Prepost Mean PANAS
Control Pre 2.989873
Control Post 2.760759
HAI Pre 2.971795
HAI Post 3.232051
Note: Source: Thayer & Stevens (2021)
* 73 control participants, 72 HAI participants

Landscape

Rotate wide tables with landscape() function.

kable(condition_prepost_means, 
      booktabs = TRUE,
      caption = "PANAS scores by condition and prepost",
      col.names = c("Condition", "Prepost", "Mean PANAS")) |>
  kable_styling(latex_options = "hold_position") |>
  footnote(general = "Source: Thayer & Stevens (2021)",
           footnote_as_chunk = TRUE) |>
  landscape()
PANAS scores by condition and prepost
Condition Prepost Mean PANAS
Control Pre 2.989873
Control Post 2.760759
HAI Pre 2.971795
HAI Post 3.232051
Note: Source: Thayer & Stevens (2021)

Tables by {papaja}

The {papaja} package uses the kable() function to format tables in APA format with the apa_table() function. You can use many of the same arguments that are available in the kable() function. You can control where the table is placed (here, top, bottom) with the placement argument. You can add a general footnote with the note argument.

apa_table(condition_prepost_means,
          caption = "PANAS scores by condition and prepost",
          col.names = c("Condition", "Prepost", "Mean PANAS"),
          placement = "h",
          note = "Source: Thayer & Stevens (2021)")
(#tab:unnamed-chunk-19) PANAS scores by condition and prepost
Condition Prepost Mean PANAS
Control Pre 2.99
Control Post 2.76
HAI Pre 2.97
HAI Post 3.23

Note. Source: Thayer & Stevens (2021)

 

Notice the alignment is different, with everything left aligned. Let’s right align the means.

apa_table(condition_prepost_means, 
          caption = "PANAS scores by condition and prepost",
          col.names = c("Condition", "Prepost", "Mean PANAS"),
          align = c("l", "l", "r"),
          placement = "h",
          note = "Source: Thayer & Stevens (2021)")
(#tab:unnamed-chunk-20) PANAS scores by condition and prepost
Condition Prepost Mean PANAS
Control Pre 2.99
Control Post 2.76
HAI Pre 2.97
HAI Post 3.23

Note. Source: Thayer & Stevens (2021)

 

You can rotate to landscape orientation with the landscape = TRUE argument.

apa_table(condition_prepost_means, 
          caption = "PANAS scores by condition and prepost",
          col.names = c("Condition", "Prepost", "Mean PANAS"),
          align = c("l", "l", "r"),
          placement = "h",
          note = "Source: Thayer & Stevens (2021)",
          landscape = TRUE)
(#tab:unnamed-chunk-21) PANAS scores by condition and prepost
Condition Prepost Mean PANAS
Control Pre 2.99
Control Post 2.76
HAI Pre 2.97
HAI Post 3.23

Note. Source: Thayer & Stevens (2021)

 

APA-formatted statistics by {papaja}

{papaja} also includes apa_print(), which extracts statistical values in APA format.

Linear regression

hai_lm <- lm(panas_post ~ condition * gender, data = hai)
summary(hai_lm)

Call:
lm(formula = panas_post ~ condition * gender, data = hai)

Residuals:
     Min       1Q   Median       3Q      Max 
-1.80909 -0.58525  0.01475  0.48333  1.88333 

Coefficients:
                        Estimate Std. Error t value Pr(>|t|)    
(Intercept)              2.68525    0.09578  28.035  < 2e-16 ***
conditionhai             0.52385    0.13287   3.943 0.000122 ***
genderMale               0.33142    0.20066   1.652 0.100662    
conditionhai:genderMale -0.18218    0.30884  -0.590 0.556142    
---
Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1

Residual standard error: 0.7481 on 153 degrees of freedom
Multiple R-squared:  0.109, Adjusted R-squared:  0.09149 
F-statistic: 6.236 on 3 and 153 DF,  p-value: 0.0005053
apa_print(hai_lm)
$estimate
$estimate$Intercept
[1] "$b = 2.69$, 95\\% CI $[2.50, 2.87]$"

$estimate$conditionhai
[1] "$b = 0.52$, 95\\% CI $[0.26, 0.79]$"

$estimate$genderMale
[1] "$b = 0.33$, 95\\% CI $[-0.07, 0.73]$"

$estimate$conditionhai_genderMale
[1] "$b = -0.18$, 95\\% CI $[-0.79, 0.43]$"

$estimate$modelfit
$estimate$modelfit$r2
[1] "$R^2 = .11$"

$estimate$modelfit$r2_adj
[1] "$R^2_{adj} = .09$"

$estimate$modelfit$aic
[1] "$\\mathrm{AIC} = 360.36$"

$estimate$modelfit$bic
[1] "$\\mathrm{BIC} = 375.64$"



$statistic
$statistic$Intercept
[1] "$t(153) = 28.03$, $p < .001$"

$statistic$conditionhai
[1] "$t(153) = 3.94$, $p < .001$"

$statistic$genderMale
[1] "$t(153) = 1.65$, $p = .101$"

$statistic$conditionhai_genderMale
[1] "$t(153) = -0.59$, $p = .556$"

$statistic$modelfit
$statistic$modelfit$r2
[1] "$F(3, 153) = 6.24$, $p < .001$"



$full_result
$full_result$Intercept
[1] "$b = 2.69$, 95\\% CI $[2.50, 2.87]$, $t(153) = 28.03$, $p < .001$"

$full_result$conditionhai
[1] "$b = 0.52$, 95\\% CI $[0.26, 0.79]$, $t(153) = 3.94$, $p < .001$"

$full_result$genderMale
[1] "$b = 0.33$, 95\\% CI $[-0.07, 0.73]$, $t(153) = 1.65$, $p = .101$"

$full_result$conditionhai_genderMale
[1] "$b = -0.18$, 95\\% CI $[-0.79, 0.43]$, $t(153) = -0.59$, $p = .556$"

$full_result$modelfit
$full_result$modelfit$r2
[1] "$R^2 = .11$, $F(3, 153) = 6.24$, $p < .001$"



$table
A data.frame with 6 labelled columns:

                               term estimate      conf.int statistic  df
1                         Intercept     2.69  [2.50, 2.87]     28.03 153
2                      Conditionhai     0.52  [0.26, 0.79]      3.94 153
3                        GenderMale     0.33 [-0.07, 0.73]      1.65 153
4 Conditionhai $\\times$ GenderMale    -0.18 [-0.79, 0.43]     -0.59 153
  p.value
1  < .001
2  < .001
3    .101
4    .556

term     : Predictor 
estimate : $b$ 
conf.int : 95\\% CI 
statistic: $t$ 
df       : $\\mathit{df}$ 
p.value  : $p$ 
attr(,"class")
[1] "apa_results" "list"       
apa_table(apa_print(hai_lm)$table,
          caption = "Linear regression results",
          placement = "h")
(#tab:lm-table) Linear regression results
Predictor \(b\) 95% CI \(t\) \(\mathit{df}\) \(p\)
Intercept 2.69 [2.50, 2.87] 28.03 153 < .001
Conditionhai 0.52 [0.26, 0.79] 3.94 153 < .001
GenderMale 0.33 [-0.07, 0.73] 1.65 153 .101
Conditionhai \(\times\) GenderMale -0.18 [-0.79, 0.43] -0.59 153 .556

Let’s clean up those predictor names.

hai_lm_table <- apa_print(hai_lm)$table
hai_lm_table <- hai_lm_table |>
  mutate(term = str_replace(term, "Conditionhai", "Condition"),
         term = str_replace(term, "GenderMale", "Gender"))
apa_table(hai_lm_table,
          caption = "Linear regression results",
          placement = "h")
(#tab:unnamed-chunk-23) Linear regression results
term \(b\) 95% CI \(t\) \(\mathit{df}\) \(p\)
Intercept 2.69 [2.50, 2.87] 28.03 153 < .001
Condition 0.52 [0.26, 0.79] 3.94 153 < .001
Gender 0.33 [-0.07, 0.73] 1.65 153 .101
Condition \(\times\) Gender -0.18 [-0.79, 0.43] -0.59 153 .556

How could we name the first column Predictor instead of term?

Other table packages

  • {gt} RStudio’s grammar of tables (logically like ggplot2)
  • {flextable} Good Word output but a bit tricky to work with
  • {huxtable} Very flexible but tricky to work with