Dataset for chapter "Quality and Bias"

Simpson's paradox is a major problem in data analysis. These graphs were created (via a MATLAB program with a fixed 'random' seed) to illustrate the paradox. The program and files are provided to allow the reader to experiment and see how sensitive Simpson's paradox is to the random seed or the various parameters used.

Dataset is a MATLAB program, and its output, for the graphs illustrating in Davenport's chapter "Quality and Bias" in the BCS-published book, edited by Adam Leon Smith, on "Artificial Intelligence and Software Testing". They are provided here, and linked from the book, to allow readers to experiment.

Artificial Intelligence, Bias, Simpson's Paradox
Information and communication technologies
Mathematical sciences

Cite this dataset as:
Davenport, J., 2022. Dataset for chapter "Quality and Bias". Bath: University of Bath Research Data Archive. Available from:


[QR code for this page]

application/zip (197kB)
Creative Commons: Attribution 4.0

Zip of data to support chapter "Quality and Bias".



University of Bath
Rights Holder


Data collection method:

This dataset goes with the graphs on Simpson's paradox in Davenport's chapter in the BCS-published book, edited by Adam Leon Smith, on "Artificial Intelligence and Software Testing". The .m file generates the .fig, from which the .jpg was saved. No additional manipulations were performed on the .fig between generation and saving the .jpg. 1729 (Hardy's taxi number) is explicitly set as the seed to ensure reproducibility. Data generated 14 November 2021, and images signed off 8 December 2021.


Publication details

Publication date: 10 March 2022
by: University of Bath

Version: 1


URL for this record:

Related papers and books

Black, R., Davenport, J., Olszewska, J., Rößler, J., Smith, A. L., and Wright, J., 2022. Artificial Intelligence and Software Testing: Building Systems You Can Trust. Swindon: BCS. Available from:

Contact information

Please contact the Research Data Service in the first instance for all matters concerning this item.


Faculty of Science
Computer Science

Research Centres & Institutes
UKRI CDT in Accountable, Responsible and Transparent AI