Statistics

Statistics is the scientific discipline used to collect, organize, analyze, and interpret data. It acts as a bridge between raw facts and meaningful information. While many fields use data, statistics focuses specifically on making decisions under conditions of uncertainty. It is often described as both the science of uncertainty and a technology for extracting information. This field can be viewed as a branch of mathematics or a distinct mathematical science. In the Mathematics Subject Classification, it is indexed at number 62. It is closely tied to probability theory, which studies random phenomena.

To begin any statistical study, researchers must define a statistical population. A population is the entire group being studied. This could be a diverse group of people, such as every citizen in a specific country. It could also be a group of objects, like every atom within a crystal. Because it is often impossible to measure every single member of a population, statisticians use a sample. A sample is a smaller subset of the population. For the results to be valid, the sample must use representative sampling. This ensures that conclusions drawn from the sample can be reasonably extended to the whole population.

There are two primary ways to conduct research: experimental and observational studies. In an experimental study, researchers actively manipulate the system under study. They take initial measurements, change a specific variable, and then take more measurements to see the effect. For example, the Hawthorne study at the Western Electric Company investigated how light levels affected worker productivity. Researchers changed the illumination to see if it caused a change in output. In contrast, an observational study does not involve any manipulation. Instead, researchers simply gather data and look for correlations, such as the link between smoking and lung cancer.

Data analysis is generally divided into two main methods: descriptive and inferential statistics. Descriptive statistics aim to summarize the characteristics of a data set. They often focus on two properties: central tendency and dispersion. Central tendency, or location, seeks to find a typical or central value, such as the mean. Dispersion, or variability, describes how much the data points depart from that center. Inferential statistics take a different approach. This method uses data from a sample to make predictions or induce statements about a larger population. This process relies heavily on the framework of probability theory.

When testing relationships between data sets, statisticians use a formal procedure involving hypotheses. A researcher proposes a hypothesis about a relationship between two sets of data. This is compared against a null hypothesis, which assumes there is no relationship between the sets. Statistical tests quantify how much the data contradicts this null hypothesis. This process can lead to two specific types of errors. A Type I error, or a "false positive," occurs when the null hypothesis is rejected even though it is actually true. A Type II error, or a "false negative," occurs when the null hypothesis is not rejected even though it is actually false.

The history of the term is rooted in political and social descriptions. The word originates from the Latin "status," meaning a condition in society or a state. The German word "statistik" was coined by Gottfried Achenwall to describe a summary of how things stand. In the 1770s, the term entered English through German, originally referring to political arrangements. By the 1790s, the term gained its modern meaning through the works of John Sinclair. Today, the term "statistik" in German is synonymous with mathematical statistics.

Statistical processes are naturally prone to various types of errors. Some errors are classified as random, often referred to as "noise." Others are systematic, known as bias, which can result in skewed or unfair estimates. Errors can also occur through human blunders, such as an analyst reporting incorrect units. Missing data or censoring can also lead to biased results. Statisticians have developed specific techniques to address these problems and improve the accuracy of their models.