How To Tell If A Linear Approximation Is Over Or Under?

A linear approximation is over or under the true value depending on the shape of the curve at the point you are approximating. If the curve bends upward (concave up) near that point, the tangent line sits below the curve, so the approximation is an underestimate. If the curve bends downward (concave down), the tangent line sits above the curve, so the approximation is an overestimate. The mathematical test is the second derivative: positive means under, negative means over.

That rule holds for the standard tangent-line approximation used in calculus. It is not a guess or a shortcut. It comes directly from how curvature works, and it gives you a fast way to check your answer without graphing anything.

How To Tell If A Linear Approximation Is Over Or Under Using the Second Derivative

The second derivative is the deciding factor. When you approximate a function f(x) near a point a using its tangent line, the error depends on how the function curves away from that line.

Here is the logic in plain terms. The tangent line matches the function’s value and its slope at the point a. What it cannot capture is curvature. If the function curves upward as you move away from a, the real values climb above the line. If it curves downward, the real values fall below the line.

• If f”(a) > 0 (concave up), the approximation is an underestimate.
• If f”(a) < 0 (concave down), the approximation is an overestimate.
• If f”(a) = 0, the test is inconclusive and you need to check further.

This is the same concavity test used across calculus. A positive second derivative means the slope is increasing, which bends the curve upward. A negative second derivative means the slope is decreasing, which bends the curve downward.

Why Concavity Determines the Direction of the Error

Concavity describes which way a curve bends, and that bend is exactly what the tangent line misses.

Picture a curve that opens upward, like the bottom of a bowl. At any point on that bowl, the tangent line touches the curve and then runs underneath it on both sides. Since the line stays below the curve, any value you read off the line is smaller than the true value. That is an underestimate.

Now picture a curve that opens downward, like the top of a hill. The tangent line touches the curve and runs above it on both sides. Values on the line are larger than the true values. That is an overestimate.

This relationship is not a coincidence. It follows from the geometry of the tangent line and the definition of concavity. When the second derivative is positive, the function is bending toward higher values faster than the straight line can follow. When it is negative, the function is bending toward lower values.

One clarification worth keeping in mind: this rule applies to the tangent-line approximation at a single point. It tells you the direction of the error locally, near that point. Over a large interval, a function can switch concavity, and the direction of the error can change with it.

What If the Second Derivative Is Zero?

When the second derivative equals zero at the point, the simple rule breaks down. You cannot conclude over or under from concavity alone at that spot.

A second derivative of zero means the curve is momentarily flat in its bending. It could be an inflection point, where concavity switches from up to down or the reverse. Right at that point, the tangent line may sit on one side, but the direction of the error nearby depends on higher-order behavior.

In these cases, the practical approach is to check a specific value. Pick a point close to a, compute the true function value, and compare it to the approximation. If the true value is larger, you have an underestimate. If it is smaller, you have an overestimate.

You can also look at the third derivative or examine the sign of the second derivative just to the left and right of the point. If concavity is the same on both sides, the direction of the error is consistent. If it flips, the error direction flips too.

How To Check Over or Under Without Calculus

You do not always need derivatives. A direct numerical comparison answers the question just as well.

Follow these steps:

• Write down the tangent-line approximation at your chosen point.
• Pick a nearby input value where you can compute the true function output.
• Compare the two numbers directly.
• If the true value is greater, the approximation is under. If it is less, the approximation is over.

This method works for any function you can evaluate. It is slower than checking the second derivative, but it requires no calculus and it never gives a wrong answer for the point you tested.

The limitation is that it only tells you about the point you tested. The second derivative tells you about the whole neighborhood around a, as long as concavity does not change there. That is why the derivative test is preferred when you want a general answer rather than a single data point.

Common Mistakes When Deciding Over or Under

The most frequent error is confusing the sign of the first derivative with the sign of the second. The first derivative tells you whether the function is increasing or decreasing. It says nothing about whether the approximation is over or under.

A function can be increasing and concave up, which gives an underestimate. It can be increasing and concave down, which gives an overestimate. The direction of the function and the direction of the error are separate questions.

Another mistake is assuming the approximation is always under or always over for a given function type. Exponential growth curves are concave up everywhere, so tangent-line approximations are always underestimates. But many functions change concavity, and the error direction changes with it.

A third error is ignoring the size of the interval. Even when the direction of the error is correct, the size of the error grows as you move farther from the point of tangency. The over-or-under answer stays the same, but the gap between approximation and truth widens.

Worked Example: Checking a Square Root Approximation

Suppose you approximate the square root of 9.2 using the tangent line at x = 9. The function is f(x) = √x, and its second derivative is negative for all positive x, since the curve is concave down.

Because the second derivative is negative, the tangent line sits above the curve. The approximation should be an overestimate.

Checking directly: the tangent-line value comes out slightly above the true square root of 9.2. The prediction from concavity matches the numerical result. This is the kind of cross-check that builds confidence in the rule.

Try the same logic with f(x) = x² at x = 3, approximating near 3.1. The second derivative is 2, which is positive, so the curve is concave up and the approximation is an underestimate. Comparing the tangent-line value to the true value of 3.1² confirms it.

Why This Matters Beyond the Classroom

Knowing the direction of the error tells you whether your estimate is too high or too low. That matters in engineering, physics, and any field where a linear model stands in for a curved reality.

If you know your approximation is an underestimate, you know the true value is at least as large as your number. That gives you a bound. Bounds are useful when you need to guarantee a result rather than just estimate it.

Linear approximation is also the foundation for more advanced methods. Newton’s method for finding roots, Euler’s method for differential equations, and many numerical techniques all rest on the same tangent-line idea. Understanding when the approximation runs high or low is the first step toward understanding the error in those methods.