Statistical Data Extraction from ggstatsplot

Overview

The jextractggstats function provides a comprehensive interface for extracting statistical data from ggstatsplot analyses within the jamovi framework. This function allows researchers to access detailed statistical results computed by ggstatsplot, making them available for custom reporting, meta-analysis, or integration with other analytical workflows.

Note: This function is currently being updated for compatibility with the latest versions of ggstatsplot. The examples below demonstrate the intended functionality.

Key Features

Multiple Analysis Types: Supports between-groups comparisons, within-subjects analysis, histogram distributions, correlations, scatterplots, categorical data analysis, contingency table analysis, and one-sample tests
Comprehensive Data Extraction: Extracts subtitle data, caption information, pairwise comparisons, and descriptive statistics
Flexible Statistical Testing: Supports parametric, non-parametric, robust, and Bayesian methods
Customizable Output: Multiple extraction components and formatting options
Natural Language Interpretation: Provides automated interpretation of results

Loading Required Packages

library(ClinicoPath)
library(dplyr)
library(ggplot2)

# Load example data
data(histopathology, package = "ClinicoPath")

Basic Usage

Direct ggstatsplot Examples

For now, we demonstrate the underlying ggstatsplot functionality that jextractggstats is designed to work with:

library(ggstatsplot)

# Histogram with statistical details
gghistostats(
  data = histopathology,
  x = Age,
  type = "parametric",
  centrality.plotting = TRUE
)

# Between-groups comparison
ggbetweenstats(
  data = histopathology,
  x = Sex,
  y = Age,
  type = "parametric",
  pairwise.comparisons = TRUE
)

# Correlation analysis
ggscatterstats(
  data = histopathology,
  x = Age,
  y = OverallTime,
  type = "parametric"
)

Intended jextractggstats Usage

Once the compatibility issues are resolved, the function will work as follows:

# Basic histogram analysis with statistical tests
result_histogram <- jextractggstats(
  data = histopathology,
  dep_var = "Age",
  group_var = NULL,
  analysis_type = "histogram",
  extract_components = "all",
  detailed_results = TRUE,
  show_interpretation = TRUE
)

# Between-groups comparison
result_between <- jextractggstats(
  data = histopathology,
  dep_var = "Age",
  group_var = "Sex",
  analysis_type = "between_stats",
  statistical_test = "parametric",
  extract_components = "all",
  pairwise_comparisons = TRUE,
  pairwise_correction = "holm",
  detailed_results = TRUE
)

# Correlation analysis
result_correlation <- jextractggstats(
  data = histopathology,
  dep_var = "Age",
  group_var = "OverallTime",
  analysis_type = "correlation",
  statistical_test = "parametric",
  extract_components = "all",
  detailed_results = TRUE,
  show_interpretation = TRUE
)

Function Parameters

The jextractggstats function supports the following key parameters:

Core Parameters

data: The dataset to analyze
dep_var: Dependent variable (outcome variable)
group_var: Grouping variable (independent variable, optional)
analysis_type: Type of analysis (“histogram”, “between_stats”, “correlation”, etc.)

Statistical Options

statistical_test: “parametric”, “nonparametric”, “robust”, or “bayes”
effect_size_type: “eta”, “omega”, “cohens_d”, “hedges_g”, etc.
conf_level: Confidence level (default: 0.95)

Extraction Options

extract_components: “all”, “subtitle_data”, “caption_data”, “pairwise_data”, “descriptive_data”
pairwise_comparisons: Enable pairwise comparisons (TRUE/FALSE)
pairwise_correction: Multiple comparison correction method

Output Options

detailed_results: Include detailed statistical summary (TRUE/FALSE)
show_interpretation: Include natural language interpretation (TRUE/FALSE)
output_format: “table”, “dataframe”, or “json”

Statistical Methods Supported

Analysis Types

Histogram Analysis: Single variable distribution with statistical tests
Between-Groups: Compare means/distributions across groups
Within-Subjects: Paired/repeated measures comparisons
Correlation: Relationship between continuous variables
Scatterplot: Enhanced correlation with marginal distributions
Bar Chart: Categorical data analysis
Contingency: Two categorical variables association
One-Sample: Test against theoretical value

Statistical Tests

Parametric: t-tests, ANOVA, Pearson correlation
Non-parametric: Mann-Whitney, Kruskal-Wallis, Spearman correlation
Robust: Trimmed means, robust statistics
Bayesian: Bayesian equivalents with Bayes factors

Clinical Applications

The function is designed for various clinical and pathological research applications:

Biomarker Analysis

Compare biomarker levels between patient groups
Analyze marker correlations with clinical outcomes
Extract effect sizes for meta-analysis

Survival Analysis Components

Compare survival times between groups
Analyze prognostic factors
Extract hazard ratios and confidence intervals

Pathological Grade Analysis

Compare continuous variables by tumor grade
Analyze categorical associations
Multiple comparison corrections

Current Status and Future Development

The jextractggstats function is currently being updated to ensure compatibility with the latest versions of ggstatsplot and related dependencies. The core functionality has been implemented and will be fully functional once the compatibility issues are resolved.

Key Features in Development:

Enhanced error handling for edge cases
Support for additional ggstatsplot functions
Improved data extraction algorithms
Better integration with jamovi framework

Summary

The jextractggstats function will provide a comprehensive solution for extracting statistical data from ggstatsplot analyses within the jamovi framework. Key advantages include:

Comprehensive Extraction: Access to all statistical components computed by ggstatsplot
Multiple Analysis Types: Support for various statistical analyses and visualizations
Flexible Configuration: Extensive customization options for different research needs
Integration Ready: Extracted data suitable for meta-analysis and custom reporting
Clinical Applications: Designed for biomarker analysis, survival studies, and pathological research

Key Applications

Research Documentation: Extract detailed statistics for manuscripts and reports
Meta-Analysis: Gather effect sizes and confidence intervals for systematic reviews
Custom Visualization: Access underlying data for specialized plotting
Quality Control: Verify statistical computations and assumptions
Educational Use: Access step-by-step statistical calculations for teaching

This function bridges the gap between ggstatsplot’s powerful statistical capabilities and the need for detailed, accessible statistical output in clinical and pathological research contexts.

Session Information

sessionInfo()

ClinicoPath

2025-07-13