Creating comparison tables with MakeComparisonTable()

1 Overview

MakeComparisonTable() creates publication-ready summary tables for comparing groups across continuous and categorical variables.

This vignette demonstrates:

Basic group comparisons
Adding effect sizes
Parametric versus nonparametric testing
Covariate adjustment
Pairwise comparisons
Reference-group comparisons
Automatic use of variable labels

2 Load packages

library(SciDataReportR)
library(dplyr)

3 Load example data

SciDataReportR includes an example dataset and codebook.

data("SampleData")
data("SampleVariableTypes")

RevaluedObj <- RevalueData(
  SampleData,
  SampleVariableTypes
) 
df_Revalued <- RevaluedObj$RevaluedData

4 Basic comparison table

The most common use case is comparing groups across a set of variables.

Here we compare diagnostic groups.

tbl_basic <- MakeComparisonTable(
  df_Revalued,
  CompVariable = "Diagnosis",
  Variables = c(
    "age",
    "sex",
    "Genotype",
    "AXL",
    "Calbindin",
    "Ferritin",
    "MMP7"
  )
)

tbl_basic

Comparison by Diagnosis (values: mean (SD)).
Characteristic	Control N = 242¹	Impaired N = 91¹	p-value²	Test²
Age	72.75 (13.26)	71.78 (13.12)	0.553	Welch t-test
Sex			0.0272	Pearson chi-squared
Female	157 (65%)	47 (52%)
Male	85 (35%)	44 (48%)
Genotype			<0.001	Fisher (sim.)
E2E2	2 (0.8%)	0 (0%)
E2E3	30 (12%)	7 (7.7%)
E2E4	7 (2.9%)	1 (1.1%)
E3E3	133 (55%)	34 (37%)
E3E4	65 (27%)	41 (45%)
E4E4	5 (2.1%)	8 (8.8%)
AXL receptor tyrosine kinase	0.28 (0.46)	0.34 (0.41)	0.238	Welch t-test
Calbindin	21.99 (3.90)	22.93 (4.85)	0.101	Welch t-test
Ferritin	2.70 (0.76)	2.90 (0.83)	0.0426	Welch t-test
Matrix metalloproteinase 7	-4.07 (1.58)	-3.21 (1.28)	<0.001	Welch t-test
¹ Mean (SD); n (%)
² p-values use Welch t-test (two groups) or ANOVA (more than two groups) for continuous outcomes and chi-square or Fisher’s exact test, selected from expected counts for categorical outcomes.

The resulting table contains:

Descriptive statistics
Group comparison tests
P-values
Automatically applied variable labels

5 Adding effect sizes

Effect sizes provide information about the magnitude of group differences.

tbl_effectsize <- MakeComparisonTable(
  df_Revalued,
  CompVariable = "Diagnosis",
  Variables = c(
    "age",
    "sex",
    "Genotype",
    "AXL",
    "Calbindin",
    "Ferritin",
    "MMP7"
  ),
  AddEffectSize = TRUE
)

tbl_effectsize

Comparison by Diagnosis (values: mean (SD)).
Characteristic	Control N = 242¹	Impaired N = 91¹	p-value²	Test²	Effect size	ES method
Age	72.75 (13.26)	71.78 (13.12)	0.553	Welch t-test	0.07	\|d\|
Sex			0.0272	Pearson chi-squared	0.12	Cramer's V
Female	157 (65%)	47 (52%)
Male	85 (35%)	44 (48%)
Genotype			0.0011	Fisher (sim.)	0.25	Cramer's V
E2E2	2 (0.8%)	0 (0%)
E2E3	30 (12%)	7 (7.7%)
E2E4	7 (2.9%)	1 (1.1%)
E3E3	133 (55%)	34 (37%)
E3E4	65 (27%)	41 (45%)
E4E4	5 (2.1%)	8 (8.8%)
AXL receptor tyrosine kinase	0.28 (0.46)	0.34 (0.41)	0.238	Welch t-test	0.14	\|d\|
Calbindin	21.99 (3.90)	22.93 (4.85)	0.101	Welch t-test	0.22	\|d\|
Ferritin	2.70 (0.76)	2.90 (0.83)	0.0426	Welch t-test	0.26	\|d\|
Matrix metalloproteinase 7	-4.07 (1.58)	-3.21 (1.28)	<0.001	Welch t-test	0.57	\|d\|
¹ Mean (SD); n (%)
² p-values use Welch t-test (two groups) or ANOVA (more than two groups) for continuous outcomes and chi-square or Fisher’s exact test, selected from expected counts for categorical outcomes.

When reporting group differences, effect sizes should generally be interpreted alongside p-values.

For parametric continuous outcomes, the reported effect size depends on the number of groups:

Two groups use absolute Cohen’s (d). With covariates, this is the adjusted estimated marginal mean difference divided by the ANCOVA residual standard deviation.
More than two groups use Cohen’s (f). With covariates, partial Cohen’s (f) is calculated from the Type II ANCOVA group effect.

Cohen’s (d) and Cohen’s (f) are not numerically equivalent. In a balanced two-group design only, (d 2f). Interpret each effect-size scale in its methodological context; conventional magnitude guides are heuristics, not thresholds for clinical importance. The generated table caption and methods footnote remain focused on the summaries and hypothesis tests used.

6 Nonparametric analyses

Nonparametric testing uses rank-based methods that are more robust to non-normal distributions.

tbl_nonparametric <- MakeComparisonTable(
  df_Revalued,
  CompVariable = "Diagnosis",
  Variables = c(
    "age",
    "sex",
    "Genotype",
    "AXL",
    "Calbindin",
    "Ferritin",
    "MMP7"
  ),
  Parametric = FALSE,
  AddEffectSize = TRUE
)

tbl_nonparametric

Comparison by Diagnosis (values: median [IQR]).
Characteristic	Control N = 242¹	Impaired N = 91¹	p-value²	Test²	Effect size	ES method
Age	74.00 [64.00, 83.00]	73.00 [63.00, 82.00]	0.707	Wilcoxon rank-sum	-0.00	epsilon-squared
Sex			0.0272	Pearson chi-squared	0.12	Cramer's V
Female	157 (65%)	47 (52%)
Male	85 (35%)	44 (48%)
Genotype			<0.001	Fisher (sim.)	0.25	Cramer's V
E2E2	2 (0.8%)	0 (0%)
E2E3	30 (12%)	7 (7.7%)
E2E4	7 (2.9%)	1 (1.1%)
E3E3	133 (55%)	34 (37%)
E3E4	65 (27%)	41 (45%)
E4E4	5 (2.1%)	8 (8.8%)
AXL receptor tyrosine kinase	0.28 [-0.04, 0.61]	0.28 [0.10, 0.61]	0.481	Wilcoxon rank-sum	-0.00	epsilon-squared
Calbindin	21.92 [19.63, 24.46]	22.66 [20.00, 26.91]	0.119	Wilcoxon rank-sum	0.00	epsilon-squared
Ferritin	2.71 [2.20, 3.22]	2.90 [2.29, 3.33]	0.127	Wilcoxon rank-sum	0.00	epsilon-squared
Matrix metalloproteinase 7	-4.03 [-5.12, -3.16]	-3.35 [-4.03, -2.26]	<0.001	Wilcoxon rank-sum	0.07	epsilon-squared
¹ Median [Q1, Q3]; n (%)
² p-values use Wilcoxon rank-sum test (two groups) or Kruskal-Wallis test (more than two groups) for continuous outcomes and chi-square or Fisher’s exact test, selected from expected counts for categorical outcomes.

7 Including Covariates

Age is frequently included as a covariate in biomedical analyses.

The example below evaluates group differences after accounting for age.

tbl_covariate <- MakeComparisonTable(
  df_Revalued,
  CompVariable = "Diagnosis",
  Variables = c(
    "sex",
    "Genotype",
    "AXL",
    "Calbindin",
    "Ferritin",
    "MMP7"
  ),
  Covariates = "age",
  AddEffectSize = TRUE
)

tbl_covariate

Comparison by Diagnosis (values: mean (SD)). p-values adjusted for Age.
Characteristic	Control N = 242¹	Impaired N = 91¹	p-value²	Test²	Effect size	ES method
Sex			0.0323	Logistic regression (LR)	0.12	Cramer's V
Female	157 (65%)	47 (52%)
Male	85 (35%)	44 (48%)
Genotype			<0.001	Multinomial LR	0.25	Cramer's V
E2E2	2 (0.8%)	0 (0%)
E2E3	30 (12%)	7 (7.7%)
E2E4	7 (2.9%)	1 (1.1%)
E3E3	133 (55%)	34 (37%)
E3E4	65 (27%)	41 (45%)
E4E4	5 (2.1%)	8 (8.8%)
AXL receptor tyrosine kinase	0.28 (0.46)	0.34 (0.41)	0.257	ANCOVA (Type II)	0.14	adjusted \|d\| (EMM / residual SD)
Calbindin	21.99 (3.90)	22.93 (4.85)	0.0722	ANCOVA (Type II)	0.22	adjusted \|d\| (EMM / residual SD)
Ferritin	2.70 (0.76)	2.90 (0.83)	0.0454	ANCOVA (Type II)	0.25	adjusted \|d\| (EMM / residual SD)
Matrix metalloproteinase 7	-4.07 (1.58)	-3.21 (1.28)	<0.001	ANCOVA (Type II)	0.55	adjusted \|d\| (EMM / residual SD)
¹ n (%); Mean (SD)
² p-values use Type II ANCOVA for continuous outcomes and logistic likelihood-ratio test for binary outcomes or multinomial likelihood-ratio test for multicategory outcomes for categorical outcomes.

Covariate adjustment can help determine whether observed group differences persist after controlling for potential confounding factors.

8 Pairwise comparisons

When comparing more than two groups, pairwise comparisons can identify which groups differ from one another.

tbl_pairwise <- MakeComparisonTable(
  df_Revalued,
  CompVariable = "Genotype",
  Variables = c(
    "Diagnosis",
    "age",
    "sex",
    "AXL",
    "Calbindin",
    "Ferritin",
    "MMP7"
  ),
  AddPairwise = TRUE,
  AddEffectSize = TRUE
)

tbl_pairwise

Comparison by Genotype (values: mean (SD)). Pairwise p-values are bonferroni-adjusted within outcome.
Characteristic	E2E2 N = 2¹	E2E3 N = 37¹	E2E4 N = 8¹	E3E3 N = 167¹	E3E4 N = 106¹	E4E4 N = 13¹	p-value²	Test²	Effect size	ES method	E2E2 - E2E3	E2E2 - E2E4	E2E2 - E3E3	E2E2 - E3E4	E2E2 - E4E4	E2E3 - E2E4	E2E3 - E3E3	E2E3 - E3E4	E2E3 - E4E4	E2E4 - E3E3	E2E4 - E3E4	E2E4 - E4E4	E3E3 - E3E4	E3E3 - E4E4	E3E4 - E4E4
Diagnosis							<0.001	Fisher (sim.)	0.25	Cramer's V	1	1	1	1	1	1	1	0.426	0.163	1	1	1	0.0143	0.037	1
Control	2 (100%)	30 (81%)	7 (88%)	133 (80%)	65 (61%)	5 (38%)
Impaired	0 (0%)	7 (19%)	1 (13%)	34 (20%)	41 (39%)	8 (62%)
Age	54.00 (15.56)	72.30 (12.78)	78.13 (19.47)	72.04 (13.87)	72.85 (11.37)	74.83 (15.19)	0.305	ANOVA	0.14	Cohen's f	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1
Sex							0.475	Fisher (sim.)	0.12	Cramer's V	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1
Female	0 (0%)	22 (59%)	4 (50%)	100 (60%)	70 (66%)	8 (62%)
Male	2 (100%)	15 (41%)	4 (50%)	67 (40%)	36 (34%)	5 (38%)
AXL receptor tyrosine kinase	0.08 (0.64)	0.32 (0.42)	0.38 (0.66)	0.31 (0.47)	0.29 (0.41)	0.13 (0.33)	0.727	ANOVA	0.09	Cohen's f	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1
Calbindin	20.24 (5.22)	21.98 (4.04)	23.36 (3.89)	22.04 (4.31)	22.70 (3.88)	21.55 (5.81)	0.676	ANOVA	0.10	Cohen's f	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1
Ferritin	2.54 (1.06)	2.76 (0.80)	2.77 (0.91)	2.74 (0.81)	2.81 (0.75)	2.48 (0.62)	0.807	ANOVA	0.08	Cohen's f	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1
Matrix metalloproteinase 7	-3.49 (3.68)	-3.80 (1.46)	-4.10 (1.54)	-3.82 (1.53)	-3.83 (1.59)	-3.98 (1.83)	0.993	ANOVA	0.04	Cohen's f	1	1	1	1	1	1	1	1	1	1	1	1	1	1	1
¹ n (%); Mean (SD)
² p-values use Welch t-test (two groups) or ANOVA (more than two groups) for continuous outcomes and chi-square or Fisher’s exact test, selected from expected counts for categorical outcomes. Pairwise p-values are bonferroni-adjusted within outcome.

You can also choose a referent and compare only to that

tbl_pairwise_referent <- MakeComparisonTable(
  df_Revalued,
  CompVariable = "Genotype",
  Variables = c(
    "Diagnosis",
    "age",
    "sex",
    "AXL",
    "Calbindin",
    "Ferritin",
    "MMP7"
  ),
  AddPairwise = TRUE,
  AddEffectSize = TRUE,
  Referent = "E3E3"
)

tbl_pairwise_referent

Comparison by Genotype (values: mean (SD)). Pairwise p-values are bonferroni-adjusted within outcome.
Characteristic	E2E2 N = 2¹	E2E3 N = 37¹	E2E4 N = 8¹	E3E3 N = 167¹	E3E4 N = 106¹	E4E4 N = 13¹	p-value²	Test²	Effect size	ES method	E3E3 - E2E2	E3E3 - E2E3	E3E3 - E2E4	E3E3 - E3E4	E3E3 - E4E4
Diagnosis							<0.001	Fisher (sim.)	0.25	Cramer's V	1	1	1	0.00475	0.0123
Control	2 (100%)	30 (81%)	7 (88%)	133 (80%)	65 (61%)	5 (38%)
Impaired	0 (0%)	7 (19%)	1 (13%)	34 (20%)	41 (39%)	8 (62%)
Age	54.00 (15.56)	72.30 (12.78)	78.13 (19.47)	72.04 (13.87)	72.85 (11.37)	74.83 (15.19)	0.305	ANOVA	0.14	Cohen's f	1	1	1	1	1
Sex							0.475	Fisher (sim.)	0.12	Cramer's V	0.826	1	1	1	1
Female	0 (0%)	22 (59%)	4 (50%)	100 (60%)	70 (66%)	8 (62%)
Male	2 (100%)	15 (41%)	4 (50%)	67 (40%)	36 (34%)	5 (38%)
AXL receptor tyrosine kinase	0.08 (0.64)	0.32 (0.42)	0.38 (0.66)	0.31 (0.47)	0.29 (0.41)	0.13 (0.33)	0.727	ANOVA	0.09	Cohen's f	1	1	1	1	0.404
Calbindin	20.24 (5.22)	21.98 (4.04)	23.36 (3.89)	22.04 (4.31)	22.70 (3.88)	21.55 (5.81)	0.676	ANOVA	0.10	Cohen's f	1	1	1	0.963	1
Ferritin	2.54 (1.06)	2.76 (0.80)	2.77 (0.91)	2.74 (0.81)	2.81 (0.75)	2.48 (0.62)	0.807	ANOVA	0.08	Cohen's f	1	1	1	1	0.926
Matrix metalloproteinase 7	-3.49 (3.68)	-3.80 (1.46)	-4.10 (1.54)	-3.82 (1.53)	-3.83 (1.59)	-3.98 (1.83)	0.993	ANOVA	0.04	Cohen's f	1	1	1	1	1
¹ n (%); Mean (SD)
² p-values use Welch t-test (two groups) or ANOVA (more than two groups) for continuous outcomes and chi-square or Fisher’s exact test, selected from expected counts for categorical outcomes. Pairwise p-values are bonferroni-adjusted within outcome.

9 Summary

MakeComparisonTable() supports:

Descriptive statistics
Parametric analyses
Nonparametric analyses
Effect sizes
Covariate adjustment
Pairwise testing
Reference-group comparisons
Automatic variable labeling

RevalueData()
CreateCodebook()
PlotBoxPlot()
PlotVolcanoEffects()
CompareDatasets()

11 Session information

sessionInfo()

R version 4.6.1 (2026-06-24)
Platform: x86_64-pc-linux-gnu
Running under: Ubuntu 24.04.4 LTS

Matrix products: default
BLAS:   /usr/lib/x86_64-linux-gnu/openblas-pthread/libblas.so.3
LAPACK: /usr/lib/x86_64-linux-gnu/openblas-pthread/libopenblasp-r0.3.26.so;  LAPACK version 3.12.0

locale:
 [1] LC_CTYPE=C.UTF-8       LC_NUMERIC=C           LC_TIME=C.UTF-8
 [4] LC_COLLATE=C.UTF-8     LC_MONETARY=C.UTF-8    LC_MESSAGES=C.UTF-8
 [7] LC_PAPER=C.UTF-8       LC_NAME=C              LC_ADDRESS=C
[10] LC_TELEPHONE=C         LC_MEASUREMENT=C.UTF-8 LC_IDENTIFICATION=C

time zone: UTC
tzcode source: system (glibc)

attached base packages:
[1] stats     graphics  grDevices utils     datasets  methods   base

other attached packages:
[1] dplyr_1.2.1           SciDataReportR_21.0.0

loaded via a namespace (and not attached):
  [1] Exact_3.3              ggstatsplot_1.0.0      sjlabelled_1.2.0
  [4] tidyselect_1.2.1       rootSolve_1.8.2.4      farver_2.1.2
  [7] statsExpressions_2.0.0 S7_0.2.2               fastmap_1.2.0
 [10] bayestestR_0.18.1      labelled_2.16.0        digest_0.6.39
 [13] estimability_2.0.0     lifecycle_1.0.5        lmom_3.3
 [16] magrittr_2.0.5         compiler_4.6.1         rlang_1.3.0
 [19] sass_0.4.10            tools_4.6.1            yaml_2.3.12
 [22] gt_1.3.0               data.table_1.18.4      knitr_1.51
 [25] xml2_1.6.0             RColorBrewer_1.1-3     abind_1.4-8
 [28] expm_1.0-0             withr_3.0.3            purrr_1.2.2
 [31] nnet_7.3-20            grid_4.6.1             datawizard_1.3.1
 [34] xtable_1.8-8           e1071_1.7-17           gtsummary_2.5.1
 [37] paletteer_1.7.0        ggplot2_4.0.3          emmeans_2.0.4
 [40] scales_1.4.0           MASS_7.3-65            dichromat_2.0-1
 [43] insight_1.5.2          cli_3.6.6              mvtnorm_1.4-2
 [46] rmarkdown_2.31         generics_0.1.4         otel_0.2.0
 [49] RcppParallel_6.2.0     rstudioapi_0.19.0      httr_1.4.8
 [52] tzdb_0.5.0             parameters_0.29.2      commonmark_2.0.0
 [55] readxl_1.5.0           gld_2.6.8              proxy_0.4-29
 [58] effectsize_1.0.3       cellranger_1.1.0       base64enc_0.1-6
 [61] vctrs_0.7.3            Matrix_1.7-5           boot_1.3-32
 [64] sandwich_3.1-3         jsonlite_2.0.0         carData_3.0-6
 [67] car_3.1-5              litedown_0.10          hms_1.1.4
 [70] patchwork_1.3.2        rstatix_1.1.0          Formula_1.2-6
 [73] correlation_0.8.8      tidyr_1.3.2            glue_1.8.1
 [76] rematch2_2.1.2         gtable_0.3.6           tibble_3.3.1
 [79] pillar_1.11.1          htmltools_0.5.9        R6_2.6.1
 [82] evaluate_1.0.5         lattice_0.22-9         readr_2.2.0
 [85] markdown_2.0           haven_2.5.5            backports_1.5.1
 [88] cards_0.8.1            broom_1.0.13           snakecase_0.11.1
 [91] rstantools_2.7.0       DescTools_0.99.60      class_7.3-23
 [94] Rcpp_1.1.2             coda_0.19-4.1          xfun_0.60
 [97] fs_2.1.0               zoo_1.9-0              forcats_1.0.1
[100] pkgconfig_2.0.3