This variation calculator handles the six standard variation models taught in algebra: direct, direct with a power, inverse, inverse with a power, joint and combined. Give it one known set of values and it solves for the constant of variation k, writes out the resulting equation, and then predicts y at any new input you supply.
Arb Digital publishes it because variation problems are almost always set up the same way and almost always go wrong at the same step. Students find k correctly and then substitute into the wrong form, or they treat an inverse relationship as though halving the input halves the output. Showing the constant, the equation and the invariant quantity side by side makes both errors visible immediately.
What This Variation Calculator Does
It works in two stages, exactly as the textbook method does. First it uses your known values to solve for k. Then it substitutes k and your new inputs into the same equation to produce the missing value. It also reports the quantity that stays constant across every point on the relationship — y ÷ x for direct variation, x × y for inverse — which is the quickest hand check on any answer.
One thing to clear up before you go further: this page is about variation in the algebraic sense, a proportionality model. It has nothing to do with the coefficient of variation in statistics, which measures relative dispersion in a dataset and is handled by our coefficient of variation calculator. The two share a word and nothing else. Direct variation is also not the same as a general linear equation: y = kx passes through the origin, whereas the line y = mx + b handled by our slope calculator does not unless b is zero.
How to Use It
- Choose the model from the wording of the problem. "Varies directly as" means y = kx; "varies inversely as" means y = k/x; "varies jointly as" means a product; "varies directly as x and inversely as z" is the combined form.
- Enter the known point. This is the pair, or triple, the question gives you. It exists solely to pin down k.
- Set the power n if the model uses one. "Varies directly as the square of x" is n = 2; "inversely as the cube" is n = 3.
- Enter the new inputs and read y from the first grid item.
- Check the invariant. If the constant quantity in the third grid item does not hold at both points, you have chosen the wrong model.
The Formulas and How They're Calculated
Every form is the same two moves: rearrange to isolate k, then substitute forward. Direct: k = y ÷ x, then y = kx. Direct with a power: k = y ÷ xⁿ, then y = kxⁿ. Inverse: k = xy, then y = k ÷ x. Inverse with a power: k = y × xⁿ, then y = k ÷ xⁿ. Joint: k = y ÷ (xz), then y = kxz. Combined: k = yz ÷ x, then y = kx ÷ z.
Work the defaults through. Direct variation with the known point x = 4, y = 14 gives k = 14 ÷ 4 = 3.5, so the model is y = 3.5x. At the new value x = 10 the prediction is y = 3.5 × 10 = 35. The invariant y ÷ x is 3.5 at both points, as it must be. Notice that x rose by 150% and y rose by 150% too — equal proportional change is the defining property of direct variation.
Now switch the same numbers to inverse variation and watch what changes. k = 4 × 14 = 56, the model is y = 56 ÷ x, and at x = 10 the prediction is 5.6. The same 150% rise in x now produces a 60% fall in y. This is the step most people get wrong: inverse variation does not reverse the percentage, it reverses the ratio. Multiplying x by 2.5 divides y by 2.5, and 14 ÷ 2.5 = 5.6. The OpenStax College Algebra section on modeling using variation works through the same three forms with the constant identified explicitly at each step.
Reading the Problem: Which Model the Words Are Asking For
Nearly all the difficulty in variation questions is translation, not algebra. A small vocabulary does almost all the work. "Varies directly as", "is proportional to" and "varies as" all mean multiplication by k. "Varies inversely as" and "is inversely proportional to" mean division by k's partner. "Varies jointly as" means the product of two or more quantities, all in the numerator.
The combined form is the one worth reading twice, because it mixes both directions in a single sentence: "y varies directly as x and inversely as z" is y = kx ÷ z. The order of the clauses tells you which variable goes where, and swapping them produces a model that behaves in exactly the opposite way. The word "square", "cube" or "square root" attaches to whichever variable it follows, so "varies inversely as the square of the distance" is k ÷ d² and not (k ÷ d)².
Monroe Community College's college algebra notes on constructing and solving variation equations set out the same four-step method this tool follows: translate the sentence, find k from the given values, write the full formula, then solve for the unknown.
Why the Constant Is the Only Thing Worth Keeping
Students often treat k as an intermediate result to be discarded once y is found. It is the opposite: k is the whole model, and the individual data points are disposable. Once you have k you can answer any question about the relationship, in either direction, without ever returning to the original pair.
The constant also carries units, and those units are usually the most informative thing about it. If cost varies directly as weight and k comes out at 3.5, then k is 3.5 dollars per kilogram — a price. If distance varies directly as time and k is 3.5, then k is a speed. In an inverse relationship the units multiply instead: for pressure varying inversely as volume, k has units of pressure times volume, which is exactly the quantity Boyle's law holds constant.
This is why the units check is such a good error detector. A k whose units are nonsense means the model is wrong, no matter how clean the arithmetic was. It is the same instinct that makes the ratio calculator useful for sanity-checking a proportion before you build anything on top of it.
Inverse Square Laws, and Why the Exponent Changes Everything
The power forms are not a classroom embellishment. A great deal of physics is inverse variation with n = 2: gravitational attraction, electrostatic force, and the intensity of light or sound radiating from a point source all fall off as the square of distance. The reason is geometric — a fixed quantity spreading over the surface of a sphere is diluted by an area that grows as r².
The practical consequence is that intuition built on n = 1 fails badly at n = 2. Under plain inverse variation, doubling the distance halves the quantity. Under an inverse square law, doubling the distance leaves a quarter, and tripling it leaves a ninth. Set the model to inverse with a power, put n = 2, and try it: with x = 4 and y = 14, k = 224, and at x = 8 the value is 3.5 rather than 7.
The same sharpening applies to direct variation with a power. Area varies as the square of a linear dimension and volume as the cube, which is why doubling the side of a cube multiplies its surface by four and its volume by eight. If you are working with the exponents themselves, the exponent calculator handles the arithmetic, and the percentage change calculator converts the results back into the proportional language these problems are usually posed in.
When a Variation Model Is the Wrong Tool
A variation model assumes the relationship passes exactly through the origin, or through the corresponding point for an inverse form, and that a single constant describes it everywhere. Real data rarely obliges. If the true relationship is y = 2x + 30, fitting y = kx to a point at x = 10 gives k = 5, which then badly mispredicts at x = 100. The tell is that the invariant is not invariant: compute y ÷ x at two different points and the numbers disagree.
That disagreement is the diagnostic. If it holds within rounding, direct variation is a fair model. If it drifts steadily, there is a constant term the model has no way to represent, and a fitted line through several points is the honest approach — which is what our linear regression calculator is for. A variation model is a one-parameter fit through a single point; regression is a two-parameter fit through many. Neither is more correct in the abstract; they answer different questions and assume different things.
Arb Digital checks whether the relationship it is assuming actually holds before scaling spend against it.
Browse All Free Tools Talk To Our TeamCommon Mistakes to Avoid
- Reversing the percentage for inverse variation — multiplying x by 2.5 divides y by 2.5, which is a 60% fall, not a 150% one.
- Attaching a power to the wrong thing — "inversely as the square of d" is k ÷ d², never (k ÷ d)².
- Confusing direct variation with a general straight line — y = kx must pass through the origin, so any relationship with a non-zero intercept is not direct variation.
- Discarding k — the constant is the model. Keeping it lets you answer in either direction without re-deriving anything.
- Mixing up the two meanings of "variation" — the coefficient of variation is a statistical dispersion measure and has nothing to do with these models.
Related Free Tools From Arb Digital
Measure relative dispersion with the coefficient of variation calculator, work with lines that have an intercept using the slope calculator, simplify a proportion with the ratio calculator, handle powers with the exponent calculator, express results proportionally with the percentage change calculator, or fit a relationship through many points with the linear regression calculator. The full free online tools hub lists every maths tool we publish.
Frequently Asked Questions
It is the fixed multiplier k that turns one quantity into another in a proportional relationship. For direct variation it equals y divided by x, and for inverse variation it equals x multiplied by y.
Read what happens as the input grows. If the output grows with it, the relationship is direct; if the output shrinks, it is inverse. The phrases "varies directly as" and "varies inversely as" say the same thing explicitly.
A relationship where one quantity varies directly as the product of two or more others, written y equals k times x times z. The volume of a cylinder varying jointly as its height and the square of its radius is a standard example.
A model that is direct in one variable and inverse in another at the same time, written y equals k times x divided by z. Reading the sentence carefully matters, because swapping the two variables reverses the behaviour entirely.
Not quite. Direct variation is the special case of a straight line that passes through the origin. A line with a non-zero y-intercept is linear but is not direct variation.
It is inverse variation with the power set to two, so the output falls as the square of the input. Doubling the distance leaves one quarter of the original quantity rather than one half.
Compute the invariant at both points. For direct variation, y divided by x should be identical at the known point and the predicted one; for inverse variation, the product x times y should match.
This page explains a standard algebraic modelling technique for educational purposes only. Whether a variation model correctly describes a real system is a question for the subject it comes from, not for the arithmetic.
