A box-graph method for illustrating relative-size relationships in a 2 x 2 table

Area of Science:

• Biostatistics
• Data Visualization
• Epidemiological Methods

Background:

• Contingency tables are fundamental in statistical analysis, particularly in observational studies.
• Visualizing proportional relationships within these tables can enhance data interpretation.
• Existing graphical methods may not optimally display the nuances of 2x2 table data.

Purpose of the Study:

• To introduce two novel box graph methodologies for displaying the proportional relationships of four numbers in a 2x2 table.
• To provide a pictorial format for representing contingency counts analogous to established data visualization techniques.

Main Methods:

• Method 1: A 'unitary square' is sequentially divided based on denominator and then numerator proportions from study groups.
• Method 2: Four proportionately sized squares, based on the square root of each number, are arranged around a central reference point.

Main Results:

• The proposed box graphs offer a clear visual representation of the proportional distributions within 2x2 contingency tables.
• Both methods provide an intuitive way to grasp the relationships between different data categories.
• The graphical format facilitates comparison and understanding of epidemiological data.

Conclusions:

• The developed box graph techniques offer effective visual tools for analyzing 2x2 table data.
• These methods enhance the interpretability of contingency counts in epidemiological and statistical research.
• The pictorial representation aids in communicating complex data relationships.