Why the most useful statistical models are technically incorrect

Why the most useful statistical models are technically incorrect
Image: Intellect (generated)

In short: George E. P. Box, a titan of 20th-century statistics, famously argued that all models are inherently wrong. Yet, he transformed fields from quality control to time-series analysis by proving that imperfection does not preclude utility. Discover how a self-taught wartime experimenter reshaped the way we interpret data and reality.

Born in 1919, George E. P. Box began his academic journey in chemistry before World War II interrupted his studies. During the conflict, he was tasked with conducting experiments on small animals to test the effects of poison gas. Lacking formal training in data analysis, he taught himself statistics to interpret his findings. This pragmatic, problem-solving approach defined his career, leading him to earn a doctorate from the University of London in 1953 under the guidance of Egon Pearson and HO Hartley.

Box spent his career bridging the gap between theoretical mathematics and industrial application. While working for Imperial Chemical Industries between 1948 and 1956, he refined techniques that would later become industry standards. In 1960, he moved to the University of Wisconsin–Madison, where he founded the Department of Statistics and later established the Center for Quality and Productivity Improvement. His influence is cemented in the statistical lexicon through eponymous methods like Box–Jenkins models and Box–Cox transformations, which remain essential tools for researchers today.

Beyond his technical contributions, Box is remembered for his philosophical stance on scientific modeling. He frequently noted that while every model is a simplification and therefore technically wrong, some provide immense value by highlighting key patterns. This perspective encouraged scientists to focus on utility rather than chasing impossible perfection. His prolific career earned him numerous accolades, including the Guy Medal in Gold and a Fellowship of the Royal Society, before his death in 2013 at the age of 93.

His personal life was as intertwined with his field as his professional one. He was married three times, including a marriage to Joan Fisher, the daughter of the renowned statistician Ronald Fisher. Together, they authored a biography of her father in 1978. Box’s legacy persists not only through his foundational textbooks, such as Statistics for Experimenters, but through a generation of statisticians who continue to apply his pragmatic philosophy to complex, real-world problems.

Opens the Intellect+ feed with this card first. Swipe up for the next one.

Sources

To cite this card, link to https://intellectplus.co.uk/app/science-george-e-p-box.

How we source, check and correct cards: Editorial standards.

General information, not professional advice. Spotted an error? Flag the card or see corrections and complaints.

Intellect Plus is a visual educational discovery platform for exploring fascinating facts and ideas. Brain Food, Not Brain Rot. What is Intellect Plus?

Related

Finding something worth knowing…