Statistics is a relatively young scientific field. Some say that modern statistics started in the 1920s when ideas from various parts of mathematics came together to provide a basis for understanding randomness, for designing experiments, and for making “optimal” decisions. Even today, Statistics is often viewed as one of the mathematical sciences.
But Statistics has always differed from mathematics because of its very close connection with applications. Especially in recent years, many of the most important advances in Statistics have come directly from trying to solve difficult and sometimes messy problems in the real world.
Here are some examples:
- After the Second World War, Japanese industry was in ruins and Japanese goods had a reputation for having poor quality. The Japanese turned to statistician W. Edwards Deming, who developed new methods of quality management, based in large part on statistical principles. Now quality management is an important branch of Statistics, which is belatedly being widely-used in American industry.
- Ever since Statistics became its own field, the social and biological sciences have relied more and more heavily on statistical methods. Variability among individuals is a way of life for social and biological scientists, and statistics is the tool for drawing conclusions in the face of such apparently random variation. Lately, physical scientists have also discovered more and more situations in which randomness seems to be fundamental.
The trend continues today . New statistical methods are developed each year because of connections with other fields:
- Working with medical researchers, statisticians developed new methods of analyzing “categorical” data to assess the safety of an anesthetic.
- High energy physicists generate massive data sets that can be visualized only using advanced statistical methods of data display, largely developed by statisticians at the Stanford Linear Accelerator.
- Geographers find new insights by displaying and analyzing data “spatially” using GIS.
Advances in computer science have made possible “computer-intensive” methods that could not even have been conceived of 20 years ago. Recent specialized seminars in statistics have been held with the participation of experts in geology, astronomy, genetics, biophysics, criminal justice, human development, software reliability, etc., to seek statistical methods for solving new kinds of problems.
In the Statistics Department, we know from our own experience that it is easier to teach statistical ideas to people who already have some idea why they are useful. The decision to encourage Statistics majors to become familiar with an area of application was not made so that doing double majors would be easy. It was made because it helps us to produce better statisticians.
Some of the statistical methods you will study in basic courses were developed to answer practical problems. One of the most elementary and commonly used statistical procedures was invented by a mathematically-oriented brewmaster (William Gosset, usually known as “Student”, of Guinness Brewery) while trying to improve quality control in the making of ale. Modern experimental design was founded by Sir Ronald A. Fisher, an expert in agricultural genetics. Nurse/Statistician Florence Nightingale used data analysis to demonstrate a need for hospital reform in England.