The domain of a function is every input value that produces a real output, and the range is every output value it actually reaches. So for f(x) = √(x − 3), the domain is x ≥ 3 and the range is y ≥ 0 — one boundary, two different answers, because an input restriction doesn't automatically restrict the output the same way. Get that asymmetry straight and most textbook problems collapse into mechanical work.
Why the Domain Answers a Different Question Than the Range
Domain asks "what am I allowed to put in?" Range asks "what can come out?" Those are separate questions, and mistaking one for the other is the most common error in early algebra. A function can accept every real number and still produce only positive outputs — f(x) = x² does exactly that.
The domain is usually found by elimination. Start with all real numbers, then remove inputs that break arithmetic: division by zero, even roots of negatives, logarithms of zero or negatives. The range is found by construction — you solve for what the output can be, or you read the graph's vertical span. Elimination is mechanical; construction takes thought.
A quick numeric check keeps you honest. Pick three inputs inside your claimed domain and confirm each yields a real number on a calculator. Then pick one input just outside the boundary and confirm it errors out or goes imaginary. If the boundary value doesn't fail, your domain is too narrow.
One more habit pays off: state domain and range in the same notation style. Mixing inequality notation for one and interval notation for the other costs marks on exams even when both answers are right. Pick one, usually interval notation beyond Algebra II, and stay with it.
Interval Notation: The Brackets That Decide Half Your Marks
Interval notation compresses inequalities into bracket pairs, and the bracket type carries real meaning. A square bracket means the endpoint is included; a parenthesis means it is excluded. So x ≥ 3 becomes [3, ∞), while x > 3 becomes (3, ∞). Infinity always takes a parenthesis because it isn't a number you can reach.
| Inequality | Interval notation | Endpoint status |
|---|---|---|
| x ≥ 3 | [3, ∞) | 3 included |
| x > 3 | (3, ∞) | 3 excluded |
| −2 ≤ x < 5 | [−2, 5) | −2 included, 5 excluded |
| x ≠ 4, all else allowed | (−∞, 4) ∪ (4, ∞) | 4 excluded |
| All real numbers | (−∞, ∞) | No endpoints |
The union symbol ∪ joins separate pieces. A function with two vertical asymptotes at x = 1 and x = −1 has a domain of (−∞, −1) ∪ (−1, 1) ∪ (1, ∞) — three pieces, two symbols. This is the notation most graphing software and standardized tests expect, including the College Board's SAT math section, which routinely tests domain restrictions through rational and radical expressions.
Watch the square bracket on a restricted range. For f(x) = x² with domain [−2, 2], the range is [0, 4] — closed at both ends, because the function does reach 0 at x = 0 and 4 at both endpoints. Students often write (0, 4) by reflex. The function touches both values, so both brackets are square.
Set-builder notation, {x | x ≥ 3}, says the same thing in words and appears in precalculus textbooks before interval notation takes over. Know both; some instructors demand one specifically.
Linear Functions: All Reals Unless the Context Says Otherwise
For f(x) = mx + b with m ≠ 0, the domain is (−∞, ∞) and the range is (−∞, ∞). A non-horizontal line covers every input and every output. There is no restriction to hunt for, which makes linear functions the easy case and the one students often overthink.
The exception is a horizontal line, f(x) = b. Domain stays (−∞, ∞), but the range collapses to the single value {b}, written as [b, b]. A function that outputs 7 for every input has a range of exactly one number. That trips people up because the domain is still infinite; the two ends of the function behave completely differently.
Word problems change the answer, and this is where domain gets interesting. A taxi fare modeled as C(m) = 2.50 + 1.75m has a mathematical domain of all reals but a real-world domain of m ≥ 0, because negative miles don't exist. If the ride caps at 30 miles, the domain becomes [0, 30] and the range is [2.50, 55.00] — computed by substituting both endpoints into the formula.
Contextual domains are a tested concept on the SAT and in most state algebra standards. The rule: the mathematical domain is the default, and context narrows it. If a problem describes a real quantity, ask what values make physical sense before writing the interval.
Quadratics and Polynomials: Same Domain, Range That Depends on the Vertex
Every polynomial — quadratic, cubic, quartic, any degree — has a domain of (−∞, ∞). Polynomials never divide by a variable, never take even roots of variables, and never take logs, so nothing ever breaks. That's the whole domain story for this entire family.
Range is where polynomials diverge. An even-degree polynomial has a range bounded on one side; an odd-degree polynomial has a range of (−∞, ∞). For a quadratic in vertex form, f(x) = a(x − h)² + k, the vertex sits at (h, k). If a > 0 the parabola opens up and the range is [k, ∞). If a < 0 it opens down and the range is (−∞, k]. The k value comes straight off the equation.
Take f(x) = 2(x − 3)² − 5. The vertex is (3, −5), the parabola opens up, so the range is [−5, ∞). The domain is (−∞, ∞). Nothing else needs computing — the vertex form hands you the answer. If the quadratic arrives in standard form, complete the square or use x = −b/(2a) to find the vertex input, then substitute to get k.
Cubics behave differently. f(x) = x³ − 2x has a range of (−∞, ∞) because the cube term drags the output to both infinities. A quartic like x⁴ − 3x² has a minimum somewhere and rises in both directions, so its range is [m, ∞) for some minimum m. The pattern: even degree has one open end, odd degree has two. That rule holds for every polynomial you'll meet before calculus.
Rational Functions: Where the Denominator Sets the Boundary
A rational function is a ratio of polynomials, and its domain is all reals except the inputs that make the denominator zero. For f(x) = 1/(x − 4), the denominator vanishes at x = 4, so the domain is (−∞, 4) ∪ (4, ∞). One excluded point, two intervals.
The range is trickier because of horizontal asymptotes. For f(x) = 1/(x − 4), the function never equals 0 — as x grows large, the output approaches 0 but never reaches it. So the range is (−∞, 0) ∪ (0, ∞). The horizontal asymptote at y = 0 becomes an excluded value in the range.
Not every excluded input excludes an output. For f(x) = (x² − 1)/(x − 1), the denominator is zero at x = 1, so 1 is excluded from the domain. But the numerator also factors as (x − 1)(x + 1), so the function simplifies to x + 1 everywhere except at x = 1. The graph is a line with a single hole. The range is all reals except 2, because the hole sits at (1, 2) — the output 2 is the one value the function never produces.
That's a genuine trap. Many students simplify the fraction, forget the hole, and report the range as (−∞, ∞). The simplified form hides the restriction that the original expression enforces. Always find excluded inputs on the original denominator before cancelling anything.
For a rational function with a horizontal asymptote at y = c, c is excluded from the range only if the function never crosses it. Some rational functions cross their asymptote, so don't assume — solve for x in terms of y and see which y values produce no solution.
Radicals and Logarithms: The Restrictions That Flip the Usual Order
Square roots require nonnegative inputs. For f(x) = √(x − 3), the radicand x − 3 must be ≥ 0, so the domain is [3, ∞). The range is [0, ∞) because a principal square root never returns a negative value. Input restricted from below, output restricted from below — a rare case where both boundaries sit on the same side.
Cube roots behave completely differently. f(x) = ∛(x − 3) accepts every real input, so the domain is (−∞, ∞), and it produces every real output, so the range is also (−∞, ∞). Odd roots carry no domain restriction at all. Students who blanket-apply "radicals restrict the domain" get cube-root problems wrong; the restriction belongs only to even roots.
Logarithms require positive arguments. For f(x) = log(x − 2), the argument x − 2 must be > 0, so the domain is (2, ∞) — open at 2, because log(0) is undefined, not zero. The range of any basic logarithm is (−∞, ∞), since the curve climbs without bound and dives toward negative infinity as x approaches the vertical asymptote from the right.
Compound these carefully. For f(x) = √(x − 1)/(x − 5), two restrictions stack: x ≥ 1 from the root and x ≠ 5 from the denominator. The domain is [1, 5) ∪ (5, ∞). Work each restriction separately, then intersect. Skipping the intersection step is how multi-restriction problems lose points even when both individual restrictions were found correctly.
Reading Domain and Range Off a Graph Without Solving
The graph tells you both answers directly if you read it the right way. Sweep left to right along the x-axis: every x value where the curve exists is in the domain. Sweep bottom to top along the y-axis: every y value the curve touches is in the range. Holes, gaps, and asymptotes show up as breaks in the sweep.
A parabola opening upward with vertex at (2, −3) has a domain of (−∞, ∞) — the arms extend forever horizontally — and a range of [−3, ∞). You didn't solve anything. You read the lowest point and noted the arms rise. That's often faster than algebra, which is why graphing calculators and Desmos are standard tools in Algebra II classrooms.
Closed dots mean included endpoints; open circles mean excluded. A graph that starts at a closed dot at (−4, 1) and rises to the right includes x = −4 in its domain and y = 1 in its range. If that dot is open, neither value counts, and both intervals take parentheses. This visual grammar matters more than any formula for piecewise functions, where the domain is stitched together from each piece's interval.
Piecewise notation rewards careful reading. A function defined as x² for x < 0 and x + 1 for x ≥ 0 has a domain of (−∞, ∞) — the pieces cover everything. Its range requires evaluating each piece on its own interval: x² on x < 0 gives (0, ∞), and x + 1 on x ≥ 0 gives [1, ∞). The union is [1, ∞), because the first piece's outputs all fall above 1. Piecewise ranges are unions, and the union can shrink what either piece alone would suggest.
Four Function Families Side by Side
The table below is the comparison most textbooks spread across four chapters. Same columns, one view — that's the point of building it.
| Function family | Example | Domain | Range | What sets the boundary |
|---|---|---|---|---|
| Linear (m ≠ 0) | 2x + 1 | (−∞, ∞) | (−∞, ∞) | Nothing restricts either end |
| Quadratic, a > 0 | (x − 3)² − 5 | (−∞, ∞) | [−5, ∞) | Vertex y-value |
| Rational | 1/(x − 4) | (−∞, 4) ∪ (4, ∞) | (−∞, 0) ∪ (0, ∞) | Zero denominator, horizontal asymptote |
| Square root | √(x − 3) | [3, ∞) | [0, ∞) | Radicand ≥ 0, principal root ≥ 0 |
| Cube root | ∛(x − 3) | (−∞, ∞) | (−∞, ∞) | No restriction at all |
| Logarithm | log(x − 2) | (2, ∞) | (−∞, ∞) | Argument > 0 |
Scan the domain column: only three families produce restrictions. Scan the range column: four families produce restrictions. The asymmetry is the whole lesson — domain restrictions come from forbidden operations, range restrictions come from the shape of the output.
If you're choosing how to study this, work the restricted families first. Linear and polynomial domains are free points; rational, radical, and log functions carry the actual reasoning. A student who can do those three cold can reconstruct the rest.
One caveat on rational ranges: not every rational function excludes its horizontal asymptote value. f(x) = x²/(x² + 1) has a horizontal asymptote at y = 1 and its range is [0, 1), because the output climbs toward 1 from below but never reaches it, and the numerator allows 0. The asymptote and the excluded value coincide here, but only because the function approaches from one side. Verify by solving for x rather than assuming.
Three Mistakes That Cost the Most Marks
The first mistake is treating domain and range as the same answer. They aren't, and the square-root case proves it: [3, ∞) for the domain, [0, ∞) for the range, same function. Students who copy the domain into the range field lose the marks even though they found the hard part.
The second is forgetting holes after simplifying. The rational function (x² − 1)/(x − 1) becomes x + 1, and the simplified form's domain looks like all reals. The original excludes 1. Always check the original denominator before you cancel.
The third is writing ∞ with a square bracket. Brackets on infinity show up constantly and are wrong in every notation system — infinity isn't a value the function reaches, so it can never be included. If you catch yourself writing [3, ∞], change the second bracket to a parenthesis.
A fourth error is subtler: applying even-root rules to odd roots. ∛(x − 3) has no domain restriction, full stop. The radicand can be negative because odd roots of negatives are real numbers. Students who restrict cube-root domains to x ≥ 3 report a domain far smaller than the truth, and the range shrinks with it.
A Worked Domain Problem You Can Copy
Here's the routine I'd hand to a student, applied to the compound function f(x) = √(x − 1)/(x − 5).
1: Identify every operation that can fail. The square root fails when the radicand is negative. The denominator fails when it equals zero. Nothing else in this expression can break. 2: Set the radicand's condition. x − 1 ≥ 0, so x ≥ 1. That's the first restriction, [1, ∞). 3: Set the denominator's condition. x − 5 ≠ 0, so x ≠ 5. That's the second restriction, all reals except 5. 4: Intersect the two restrictions. The first says x ≥ 1; the second removes 5 from any set. The intersection is [1, 5) ∪ (5, ∞). 5: Test the boundaries. At x = 1 the radicand is 0, so the function outputs 0 divided by −4, which equals 0 — a real number, so 1 belongs in the domain with a square bracket. At x = 5 the denominator is zero, so 5 is excluded with a parenthesis. 6: Confirm with one interior and one exterior value. x = 2 gives √1/(−3) = −1/3, a real output. x = 0 gives √(−1)/(−5), an imaginary radicand, confirming 0 is outside the domain.
That six-step sequence handles any function with stacked restrictions. Apply it once or twice and the intersection step becomes automatic.
FAQ
Does every function have a domain of all real numbers?
No — only polynomials, odd roots, exponentials, and lines do. The 2026 SAT math reference materials test domain restrictions through rational, radical, and logarithmic expressions specifically because those are the families that break. If a function contains division, an even root, or a log, assume a restriction exists until you prove otherwise.
Can the range be larger than the domain?
No. The range is the set of outputs the function actually produces, so it can never exceed the set of inputs in size for a well-defined function, and for most functions it's strictly smaller or equal. The domain is the invitation list; the range is who shows up.
Why is log(0) undefined instead of zero?
Because no exponent makes any positive base equal zero — 10 raised to any real power stays positive. Graphs of y = log(x) dive toward negative infinity as x approaches 0 from the right, never touching the y-axis. That's why the domain of log(x) is (0, ∞) with a parenthesis, not a bracket.
How do I find the range of a rational function without graphing?
Solve y = f(x) for x in terms of y, then ask which y values make that solution impossible. If the algebra forces a division by zero for some y, that y is excluded. For 1/(x − 4), solving gives x = 4 + 1/y, which fails at y = 0 — so 0 is excluded and the range is (−∞, 0) ∪ (0, ∞).
Is the domain of a cube root really all real numbers?
Yes, all reals, and the range is also all reals. ∛(−8) = −2 is a valid real output, unlike √(−8). The 2026 SAT reference sheet lists odd-root functions among the no-restriction families for exactly this reason, and state standards in Texas and California treat cube roots this way from Algebra II onward.
What's the difference between a hole and an asymptote in the range?
A hole is a single missing output at one input; an asymptote is a value the function approaches without ever reaching across an entire direction. For (x² − 1)/(x − 1), the hole at (1, 2) excludes exactly y = 2. For 1/(x − 4), the asymptote at y = 0 excludes 0 across the whole curve. Both appear as excluded values in the range, but only holes come from cancelled factors.
Do real-world problems ever change the mathematical domain?
Frequently, and that's a tested skill. A revenue function R(p) = p(100 − 4p) has a mathematical domain of all reals, but a business context narrows it to prices that yield nonnegative demand, roughly 0 ≤ p ≤ 25. Always check whether a problem describes a physical quantity before reporting (−∞, ∞).
How much of the SAT math section covers domain and range?
The College Board's four content domains — algebra, advanced math, problem-solving and data analysis, and geometry and trigonometry — place domain and range primarily in the algebra and advanced math categories. Those two together account for roughly 70% of the math section's questions, so restriction problems appear in some form on nearly every test.