Continuous variable¶

A quantitative variable able to take every real value between any two attainable values within the interval under consideration.

Version: v1 · 2026-09-28 · History
Domain-specific #: 8695
Domain group: Formal Sciences
Origin domain: Experimental Design & Statistics
Subdomains: Variable Types, Measurement Scales → Experimental Design & Statistics

Core Idea¶

A continuous variable has an interval-like set of possible numerical values: between any two attainable values in the stated range, every real intermediate value is also admissible. Continuity is therefore a claim about the variable's value space, not about how many decimal digits are recorded.

The distinction is local and model-relative. A variable may be continuous on one range, discrete on another, or have a mixed distribution with an atom and a continuous component. Calculus, density functions, differential equations, and continuous optimization rely on this interval structure, while observed rounding and digital storage create separate discretization layers.

How would you explain it like I'm…

Any-Amount-in-Between

Count the marbles in a jar and you get 3 or 4, never three and a bit. But pour water into a glass and the water can be at any height at all, with a height in between any two heights. Things like the water height are continuous variables.

Values With No Gaps

A variable is something that can take different values, like height or number of siblings. A continuous variable is one where, between any two possible values, every value in between is also possible. Your height could be 140 cm, 141 cm, or anything between, like 140.37 cm. It is still continuous even if your ruler only shows whole centimeters; how precisely you measure is a separate question from what values are possible. Number of siblings is not continuous, because you can't have 2.4 siblings.

Interval-Valued Variable

A continuous variable is one whose possible values form an interval-like range: between any two values it can take, every real number in between is also allowed. This is a claim about the space of possible values, not about how many decimal places get written down; rounding to one decimal or storing a number in a computer creates a separate discretization layer on top. Whether a variable counts as continuous can depend on the range and the model. A variable might be continuous over one range and discrete over another, or have a mixed distribution, with a chunk of probability piled on one exact value plus a continuous spread elsewhere. Tools like calculus, probability density functions, and differential equations depend on this in-between structure.

A continuous variable is one whose admissible values form an interval-like subset of the real numbers: for any two attainable values in the stated range, every real intermediate value is also admissible. Continuity is thus a property of the value space, not of recording precision. Rounding in observation and finite-precision digital storage introduce distinct discretization layers but do not change the underlying variable's type. The classification is local and model-relative: a variable may be continuous on one range and discrete on another, or follow a mixed distribution that combines an atom (a point with positive probability) with a continuous component. Methods such as calculus, density functions, differential equations, and continuous optimization rely on this interval structure. Treating a variable as continuous is therefore a modelling commitment that should be stated along with its range.

Scope of Application¶

Statistics. Continuous supports motivate density-based models.
Optimization. Decision variables vary over real intervals.
Dynamics. Continuous time and state support differential equations.
Measurement. Latent continuity is separated from instrument resolution and rounding.

Clarity¶

State the variable, units, admissible set, interval, endpoint convention, physical or modeled meaning, observation resolution, rounding, censoring, and probability-law components. Do not infer continuity from notation or discreteness from stored data type alone. Inclusion test: A variable is continuous over a stated interval when every real intermediate value is admissible whenever two endpoints are admissible. Exclusion test: An integer count, binary indicator, or value set with isolated gaps is excluded on that range. Nearest boundary: A rounded temperature reading is the closest near miss because the observation is discrete while the modeled physical quantity can remain continuous. Exit condition: The identity exits where admissible values acquire positive gaps, the variable becomes categorical, or continuity is claimed solely from a decimal display without specifying support. Common misclassifications: It is not any variable written with decimals. It is not a categorical or integer-valued variable. It is not made discrete solely by finite measurement precision. It is not the same as an everywhere-continuous probability distribution when point masses are also present. Nearest named distinctions: Continuous function: Describes how outputs vary with inputs, not the admissible support of one variable. Continuous distribution: Is a probability-law property and can diverge from the variable's broader value space. Floating-point value: Comes from a finite representable set despite decimal appearance. Ordinal variable: Has ordered categories without requiring real between-values.

Manages Complexity¶

The label replaces an uncountable value set with one interval property, enabling calculus and density reasoning. That compression hides bounds, forbidden subranges, atoms, finite precision, and whether continuity is ontological, operational, or a modeling approximation.

Abstract Reasoning¶

Define the quantitative variable and units.
Specify the exact interval or admissible set.
Take arbitrary attainable values and test every intermediate value for admissibility.
Locate any isolated points or positive gaps.
Separate underlying support from measurement and storage resolution.
Classify the probability law as continuous, discrete, or mixed independently.
Choose calculus or discrete methods only after the support convention is explicit.

Knowledge Transfer¶

Interval-density reasoning transfers across physical quantities, statistical models, and control variables when units and admissible support are literal. It stops at categories, counts, and merely rounded records. The cargo is between-value closure on a stated range.

Neighborhood in Abstraction Space¶

Continuous variable sits in a crowded region of the domain-specific corpus (39^th percentile for distinctiveness): several abstractions share nearly its structure, so a description that fits it tends to fit its neighbors too.

Family — Matrices, Measures & Numeric Structures (30 abstractions)

Nearest neighbors

Computed from structural-signature embeddings · 2026-10-08