Skip to main content Skip to page footer

Always New

A brewer in Dublin was trying to improve the quality of the barley for beer production.  A welfare worker attached to the British army in the Crimean War needed to communicate critical facts on how disease was such a big killer.  A mathematician was planning where to best apply armour to bombers in World War II.

These were all real problems.  The last two life or death problems, and some would say the first is as well.  They remain relevant today as examples of real problems motivating new statistical ideas, an approach that motivates our work at Data Analysis Australia. 

We examine the above three cases and their links to statistical problem solving below and reflect on how their approaches still remain applicable in our times.

The brewer was William Gosset.  To understand how his experimental data should be properly interpreted, he developed the mathematics of distributions of data, publishing under the name “Student” as his employer (Guinness Breweries) did not typically allow employees to publish their work.  Today, Student’s t-test is still one of the fundamental statistical techniques learned by every statistician.

The welfare worker was Florence Nightingale.  Best known as a founder of modern nursing, she also had a strong interest in mathematics.  Her “coxcomb diagrams” showing monthly fatalities in the British army highlighted a major problem, leading to critical reforms.  She became the first female Fellow of the Royal Statistical Society and an advocate of “visualisation” to convey the meaning of public health data.

The mathematician was Abraham Wald.  It was suggested that armour should be applied where planes returning from missions were most frequently damaged.  Wald realised that the data represented only the planes that had survived.  Planes damaged elsewhere had not survived, suggesting that those parts were in more critical need for protection.  This suggested placing armour on those parts not damaged in surviving planes.  Today statisticians remain aware that biases in available data must be understood and allowed for.

Today the challenges are different in detail – the advent of “big data” and the Internet of Things creates challenges of computation and data management as well as statistics – but the way of doing statistics is the same.  Real problems continue leading to ever more advanced statistical approaches.  It guides the approach of Data Analysis Australia in all we do, constantly balancing new and older methods to develop unique solutions for our clients. 

For more information on how Data Analysis Australia can help you solve your data problems, please contact

December 2017