dichotomized

Dichotomized is a term used to describe something divided into two parts or categories that are mutually exclusive and collectively exhaustive. The concept comes from the Greek roots dichotomos, meaning divided in two, and is widely used across disciplines to describe a binary partition of a set or variable.

In statistics and data analysis, dichotomizing a variable means converting it from a continuous or multi-category form into two categories, often by applying a threshold or a rule such as median split or clinically meaningful cutpoints. For example, age might be dichotomized as under 65 and 65 or older; a biomarker value as positive/negative.

Advantages include simpler interpretation and compatibility with models that require a binary input. However, dichotomization can lead to loss of information, reduced statistical power, and arbitrary boundary choices that may distort relationships and introduce bias. Critics advocate using the original scale or exploring multiple thresholds or nonbinary modeling approaches to preserve information.

In logic and philosophy, dichotomies describe pairs of related concepts with a clear binary separation, such as true/false or mind/body. The term dichotomized is also used informally to describe the act of imposing such a binary division on data or concepts. See also: dichotomy, binary variable, binarization.