Why Statistics Matter In Everyday Life

The discipline of Statistics is a set of mathematical tools for analyzing the events that occur in everyday life. In today’s world, driven by data and rapidly changing technology, most of the information we see has been filtered through the lens of statistics. Statistics has near-endless applications in everyday life, including the following.

Image source: Pixabay.com

Clinical trials: Before a new drug can be approved for use, it must be shown to be safe and effective. This is done through a statistically rigorous program of carefully planned clinical trials. In phase I a drug must be shown to be safe. In phase II it must be shown to be effective, and in phase III it must be shown to work across a broad population of patients. In each phase the experimental results are noisy. The discipline of statistics is used to ensure that the conclusions in each phase are based on actual effects, and not experimental noise.

Education: Standardized tests are one of the main tools in educational policy, and as anyone who has applied to college knows they’re an important screening tool used by university admissions. Have you ever wondered how these tests are created? Test authors come up with carefully designed banks of test questions. The test questions are then tested on groups of students to determine their difficulty (how hard is it to get the right answer), their precision (how well do they separate students of different ability) and in some cases the ease with which the correct answer can be guessed. All this is part of a branch of statistics known as “item response theory" or IRT.

Once each question’s characteristics are known, groups of questions can be assembled into a test (like the SAT, for example) so as to gain as much information as possible about a student’s ability with a small number of questions. IRT is also used to adjust the scores of students who had different sets of questions, so that if Ally’s version of the SAT happened to have harder questions than Bob’s, Ally won’t be penalized because of the test design.

Online reviews IRT isn’t just useful for education. Online reviews (e.g. for restaurants) are similar to exams in many ways. Some reviewers like to use the whole scale, while others tend to give extreme reviews (very positive or very negative). When you have different sets of reviewers reviewing different restaurants, it can be hard to know how much to trust the reviews. Methods similar to IRT can be used to adjust reviewers’ scores based on their past reviews of other restaurants to make results comparable.

Manufacturing Statistics has been used in manufacturing for a long time. When precise control over a manufacturing process is needed a strategy known as “Six Sigma” can be used to ensure quality. “Sigma” is the Greek letter most often used for a standard deviation. If manufacturing errors follow a normal distribution (and in a well controlled manufacturing process they should), then slightly more then 99% of manufactured items will be within 3 sigma of the target specification. The goal of six sigma is to make the rest of the manufacturing process tolerant to errors twice this size, so that quality related defects become vanishingly rare.

Emergency preparedness: While manufacturing errors tend to be small, hurricanes and other natural disasters are not. When preparing for a natural disaster, authorities shouldn’t be planning for a typical case. They need to plan for a bad one! But how bad? By looking at past events (hurricanes, tornados, floods, etc) one can get a measure of the devastation accompanying past disasters. Every so often a new disaster sets a new record. Extreme value theory can help you predict how bad the next “worst disaster ever” will be, and to plan accordingly.

Government and politics: Some of the earliest uses of statistics were motivated by government policy, taxation, and public health. In England, the Domesday Book was a census commissioned by William the Conquerer so that he could levy appropriate taxes on his newly conquered lands. John Graunt made early contributions to demography while helping combat the great plague of London in 1665. Echos of these early contributions can be seen in today’s governments, where in the US the CDC, the US Census bureau, the department of education, department of labor, the federal reserve, and quite frankly most other branches of government all collect and analyze official statistics in an effort to better inform government officials and better serve the public interest.

Image source: Pixabay.com

Of course this list is incomplete. I haven’t even mentioned technology and big data, which is currently one of the hottest areas in statistics and data science at the moment.

Steven Scott is a Bayesian statistician, researcher, assistant professor, and sought-after public speaker who worked for companies such as Capital One Financial Corporation and Google. For similar reads, click here.