The AP Calculus second derivative test is a local-extrema classification method that uses the sign of the second derivative at a critical point to decide whether the point is a local maximum, a local minimum, or neither. On both the AB and BC exams it appears in the BC Local Extrema and Concavity unit, in multiple-choice stems asking students to interpret f″(c), and in free-response problems that require a written justify argument. Mastery depends on three practical habits: knowing exactly when the test is conclusive, knowing the rubric language that earns the second row of credit, and avoiding the common sign-error trap when the second derivative changes sign without passing through zero.
What the second derivative test actually says, in rubric-ready language
The second derivative test is a conditional decision procedure. It applies at an interior critical point c where f′(c) = 0 and f″ is continuous near c. Under those conditions the test is: if f″(c) is positive, f has a local minimum at c; if f″(c) is negative, f has a local maximum at c; if f″(c) = 0, the test gives no information. Most candidates lose points because they try to apply the test at points where f′(c) does not exist, or at points where f is not twice differentiable. The College Board rubric treats those as separate errors: the first row of credit goes to correctly identifying the critical numbers in the open interval, and the second row goes to a conclusion line that names a max or min with a sign on f″.
Students preparing for AP Calculus should memorise the test in a one-line rubric form: c is a local minimum because f′(c) = 0 and f″(c) > 0; c is a local maximum because f′(c) = 0 and f″(c) < 0; if f″(c) = 0, the test is inconclusive. That sentence is the entire answer an AP reader needs to see to award full credit on a typical 2-row local-extrema prompt. The hardest part is not the sentence itself but the algebra that produces f″(c); a sign error in differentiation will leave a candidate writing a contradictory sentence that loses both rows.
The BC-only Local Extrema and Concavity unit deepens the test by linking it to the second derivative's role as a concavity indicator. A point where f″ changes sign from positive to negative is, by definition, an inflection point, and a local max typically sits where f is concave down. Candidates who can move freely between the test and the geometric picture of concavity earn the third row of credit that BC prompts sometimes add, which asks for the point of inflection or for the open interval on which f is concave up. The transition between rows matters: the rubric is written so that a student who skips from f′(c) = 0 directly to "local minimum" without a sign argument loses a row even if the conclusion is right.
Where the test shows up on the AB and BC exams
On the AP Calculus AB exam, the second derivative test appears almost entirely as a multiple-choice question stem. A typical stem gives a polynomial or rational function, lists the value of f″ at a critical point, and asks whether f has a local max, a local min, or neither. The trap answer is the one that says "neither" whenever f″(c) = 0, which is wrong in general but tempting for students who over-apply the test. AB free-response problems occasionally include the test as one bullet inside a longer analysis question, but the heavy concentration of the test is in MCQ Section I.
On the AP Calculus BC exam, the test is more central. The Local Extrema and Concavity unit in the Course and Exam Description places the test in the first derivative test article as well, but the second derivative version is preferred when f is a transcendental or composite function whose first derivative test would require a sign chart with hard-to-factor roots. A typical BC FRQ scenario: a function f is given implicitly, and the prompt asks for all local extrema of f on a closed interval, requiring the candidate to solve f′(c) = 0, then classify each c using f″(c) where possible and the first derivative test where f″(c) = 0. The expected solution is often a two-column table with c, f′(c), f″(c), and the conclusion — that table format is what the rubric describes as a decision line per critical point.
For both exams the question types break into three families. Family one is the plug-and-chug question: a function is given, the candidate is told the critical number, and the only task is to evaluate f″ there and pick a max or min. Family two is the build-the-function question: the candidate must differentiate twice, solve a system, and then apply the test. Family three is the interpret-the-graph question: a graph of f is shown, and the candidate must read the second derivative's sign from the graph's concavity. Most preparation programmes triage these families in that order of difficulty, and AP Courses' AP Calculus BC programme designs its free-response practice around family two, which is the family where rubric credit is most often lost.
Family-by-family signal patterns
- Plug-and-chug: f″(c) is positive or negative by a clean margin; the only real risk is misreading the sign from a calculator display that returns a small negative number rounded to zero.
- Build-the-function: a wrong chain rule or a missed negative sign in the second derivative is the dominant error; the candidate's sign on f″(c) flips and the conclusion contradicts itself.
- Interpret-the-graph: candidates misread concave up as a local minimum region; the test is about a single point c, not about an interval, and the rubric gives credit only when the candidate names the point.
The six-step rubric-ready argument for a local-extrema FRQ
The free-response argument that earns full credit on a typical AP Calculus local-extrema question has six steps, and a candidate who skips any one of them risks losing a row. The steps are not arbitrary; they map directly to the rubric bullets the AP readers use. Step one: find f′ and set it equal to zero in the open interval specified by the prompt. Step two: list the critical numbers, including any where f′ does not exist. Step three: compute f″ at each critical number. Step four: apply the second derivative test, writing the conclusion as a sentence with the sign of f″. Step five: state whether the test is conclusive, and where it is not, fall back to the first derivative test or a sign chart. Step six: list the local maxima and minima as ordered pairs (c, f(c)) with the correct value of f at the critical number.
A worked example makes the shape concrete. Let f(x) = x⁴ − 4x³ and suppose the prompt asks for all local extrema of f on (−∞, ∞). f′(x) = 4x³ − 12x² = 4x²(x − 3), so the critical numbers are x = 0 and x = 3. f″(x) = 12x² − 24x. At x = 0, f″(0) = 0, so the test is inconclusive; the candidate must fall back to the first derivative test or observe that f′ does not change sign at 0 (it is non-positive on both sides) and conclude that f has neither a local max nor a local min at 0. At x = 3, f″(3) = 12·9 − 72 = 36, which is positive, so x = 3 is a local minimum. f(3) = 81 − 108 = −27, so the local minimum is (3, −27). The rubric wants to see the inconclusive call at 0 written explicitly — saying "f″(0) = 0, test fails, so use the first derivative test" is the language that earns the second-row credit.
For most candidates the step-five moment is where preparation pays off. The instinct is to skip from f″(c) = 0 to a guess; the rubric treats that as an unstated assumption and docks a row. The correct move is to write f″(c) = 0, so the second derivative test is inconclusive; by the first derivative test, f has a local minimum at c (or whichever conclusion the sign chart supports). The exam rewards the candidate who is willing to spend fifteen extra seconds naming the failure of the test instead of guessing. The same logic applies to candidates who try to apply the second derivative test at a point where f′ does not exist: the rubric does not allow that move, and the candidate should fall back to the first derivative test without comment.
When the test fails: the four inconclusive cases and what to do instead
The second derivative test fails in four situations, and each one requires a specific replacement argument. Case one: f″(c) = 0 at a critical point. The fix is the first derivative test, using a sign chart on f′ around c. Case two: f″(c) does not exist because f is not twice differentiable at c. The fix is again the first derivative test, because the second derivative test requires f″ to be defined and continuous near c. Case three: c is an endpoint of the closed interval, in which case the second derivative test does not apply and the candidate must compare f(c) with f at interior critical points and at the other endpoint. Case four: the function is given as a piecewise expression, and the candidate cannot differentiate through the join. The fix is to evaluate the left and right derivatives separately and apply the first derivative test on each side.
For case one — the dominant case in AB and BC — a common error is to declare the test inconclusive and then stop. The rubric treats the conclusion as part of the same row as the inconclusive call. The candidate must write: f″(c) = 0, so the second derivative test is inconclusive. By the first derivative test, f′ changes from negative to positive at c, so f has a local minimum at c. That sentence is the entire second row of credit. A candidate who writes only the first half loses the row even though the analysis is correct, because the rubric rewards the conclusion line, not the analysis leading to it.
Case two shows up in piecewise FRQs and in problems where f is defined implicitly. If f is given by a piecewise rule with a corner, f″ is undefined at the corner and the test cannot run. The candidate's move is to evaluate f′ on each side and run the first derivative test. Case three is the closed-interval extreme value problem, where the test is a comparison of function values rather than a derivative test at all. Case four is the rarest on the AP exam but appears in BC problems with absolute value or floor functions. In practice, the candidate who internalises the four inconclusive cases can plan the FRQ response in under 90 seconds: list critical numbers, classify each by which test applies, and reserve a sign-chart argument for the inconclusive ones.
A four-row decision table for the inconclusive case
| Situation at c | Is the second derivative test conclusive? | Replacement argument | Rubric language that earns the row |
|---|---|---|---|
| f′(c) = 0 and f″(c) = 0 | No | First derivative test using a sign chart on f′ | "f″(c) = 0, so the test is inconclusive; by the first derivative test, f has a local [max/min] at c." |
| f′(c) = 0 but f″(c) does not exist | No | First derivative test | "f″(c) is undefined, so the second derivative test does not apply; by the first derivative test…" |
| c is an endpoint of a closed interval | No | Compare f(c) with f at interior critical points and the other endpoint | "The candidates for absolute extrema on [a, b] are f(a), f(b), and f at interior critical numbers." |
| f is piecewise and not twice differentiable at c | No | Evaluate one-sided derivatives and apply the first derivative test on each side | "By the first derivative test on (−∞, c) and (c, ∞)…" |
Common pitfalls and how to avoid them
The single most common pitfall is a sign error in the second derivative that flips the conclusion. A student differentiating f(x) = sin(x) cos(x) by the product rule produces f′(x) = cos²(x) − sin²(x) = cos(2x) and f″(x) = −2 sin(2x). At the critical number x = π/4, f″(π/4) = −2 sin(π/2) = −2, which is negative, so the candidate should conclude a local maximum. A student who mistakenly writes f″(x) = 2 sin(2x) gets a positive value and concludes a local minimum — a full-row deduction for a sign error that the rubric will catch because the conclusion contradicts the actual shape of the function. The mitigation is to test the sign of f″ at one easy point as a sanity check before writing the conclusion.
The second pitfall is treating f″(c) = 0 as automatically inconclusive. The test is inconclusive at that c, but a candidate who writes "the test is inconclusive for the whole problem" because of one f″(c) = 0 entry has misapplied the test. The rubric gives credit row by row, and a single inconclusive call at one critical number does not contaminate the test's use at another. A third pitfall is forgetting to compute f(c) and reporting the extremum as just a number on the x-axis. The rubric wants the ordered pair; on the BC exam this is a non-negotiable formatting requirement.
A fourth pitfall is to apply the second derivative test at a point where f is not differentiable. The test requires f′(c) = 0 and f″ defined near c. Candidates sometimes test at a corner where f′ does not exist, and the rubric gives no credit for that move. A fifth pitfall is to skip the first derivative test entirely and rely on the second derivative test alone. The first derivative test is the more general method and is the only one that works when f″(c) = 0; a candidate who refuses to use it will lose credit on every problem with a higher-order critical point. The way to avoid these pitfalls is to internalise the six-step argument and to plan the response before writing, so the candidate knows in advance which test applies at each critical number.
How scoring actually works: rows, decisions, and the conclusion line
On a typical AP Calculus BC FRQ that asks for all local extrema on an open interval, the rubric has three rows. Row one is the critical number row: the candidate must list every c in the open interval where f′(c) = 0 or f′ is undefined. Row two is the classification row: the candidate must decide for each c whether it is a local max, a local min, or neither, with a sign argument or a sign chart. Row three is the conclusion row: the candidate must list the local extrema as ordered pairs (c, f(c)). A candidate who finds all the critical numbers but writes no decision loses row two. A candidate who decides correctly but writes the extrema as x-values rather than ordered pairs loses row three. A candidate who confuses a critical point with an inflection point loses row two because the rubric rewards classification as a max or min specifically.
The scoring for multiple-choice questions is simpler but follows the same logic. A typical AB MCQ stem gives a function and the value of f″(c) and asks for the conclusion. The four options are usually: local max, local min, neither (because f″(c) = 0), and "cannot be determined." The trap is the cannot be determined option, which is correct only when f″(c) = 0 and the first derivative test would also fail; for the standard second derivative test application, cannot be determined is the wrong answer and is the most commonly selected wrong answer on AB MCQ. The scoring for MCQ is binary, so a single sign error costs a full point.
In the BC exam scoring, the decision line matters because a single c might be a critical number where f″(c) = 0, in which case the candidate's decision line should read inconclusive, use first derivative test, f has a local [max/min] at c. The rubric reader is trained to look for that exact three-part structure. The candidate who writes f″(c) = 0, so c is neither a max nor a min loses a row because the conclusion is wrong; the test being inconclusive does not mean no extremum exists, it means the test cannot decide. AP Courses' AP Calculus BC programme trains students to write the three-part decision line as a habit so the rubric language is automatic under time pressure.
A 6-minute preparation strategy for the second derivative test
For most candidates, the second derivative test is best prepared in a 6-minute practice loop on a single FRQ. The loop has four parts. Part one (60 seconds): read the problem, identify the open interval, and list the operations the candidate will need to perform on f. Part two (90 seconds): compute f′ and solve f′(c) = 0, listing every critical number. Part three (120 seconds): compute f″ at each critical number and apply the test, writing a one-sentence decision per critical number. Part four (90 seconds): state the local extrema as ordered pairs and check for sign errors by re-reading the conclusion line.
The first three parts are mechanical. The fourth part is where preparation pays off. A candidate who finishes the loop in 6 minutes has 90 seconds of slack to check work; a candidate who finishes in 9 minutes has no slack. In practice the difference between a 4 and a 5 on the AP Calculus BC exam often comes down to whether the candidate has the 90-second slack on local-extrema problems. The way to build that slack is to practice the six-step argument on five to seven FRQ-style problems in the weeks before the exam, until the candidate can produce the decision line without conscious thought.
For the AB exam, the preparation is shorter because the second derivative test is concentrated in MCQ. A candidate who can answer ten plug-and-chug MCQ items in 15 minutes is well prepared. For the BC exam, the candidate should add five FRQ-style problems to that base, with a focus on the build-the-function family. The full 6-minute loop should be rehearsed at least three times before the exam, with the candidate timing each part. AP Courses' AP Calculus AB and BC programmes both build this kind of timed loop into their unit reviews, with rubric-aligned feedback on the conclusion line and on the sign of f″(c).
Mistakes to expect on test day, and how to recover from each
The test-day mistake that costs the most points is a sign error in f″ that goes uncorrected. The way to recover is to test the sign of f″ at one easy point — for example, at x = 0 — before writing the conclusion line. If the sign at the easy point contradicts the candidate's general sign analysis, the candidate has made a sign error and should redo f″ before continuing. A second test-day mistake is to write the extrema as x-values rather than ordered pairs; the recovery is to go back and add the f(c) value to each pair. A third mistake is to apply the second derivative test at a point where f is not differentiable; the recovery is to fall back to the first derivative test without arguing the point further.
A fourth mistake is to spend too much time on the second derivative test when the first derivative test would be faster. If the candidate has a polynomial of degree four or higher with a critical number at c = 0, f″(0) is often 0 and the second derivative test is inconclusive. The candidate should switch to the first derivative test immediately. A fifth mistake is to confuse a critical point with an inflection point; the recovery is to check the sign of f″ on each side of c, and if the sign does not change, c is not an inflection point. If the sign does change and f′(c) = 0, c is a critical point and an inflection point, and the candidate should report both.
The recovery moves for each of these mistakes are mechanical, and a candidate who has rehearsed the six-step argument will perform them automatically. In my experience coaching AP Calculus students, the candidates who earn a 5 on the BC exam are the ones who internalise the conclusion line and the inconclusive-case table; the candidates who earn a 3 are the ones who try to apply the second derivative test in every situation and lose credit on the inconclusive cases. The test is a tool, not a religion; the rubric rewards the candidate who knows when to use the first derivative test instead.
Worked example: a BC-style FRQ in full
Let f be defined by f(x) = x − 2 sin(x) on the closed interval [0, 2π]. Find all local extrema of f and classify each as a local max or local min using the second derivative test, falling back to the first derivative test where the second derivative test is inconclusive. The expected AP-style answer proceeds in six steps. Step one: f′(x) = 1 − 2 cos(x). Step two: f′(x) = 0 when cos(x) = 1/2, so the critical numbers in [0, 2π] are x = π/3 and x = 5π/3. Step three: f″(x) = 2 sin(x). Step four: f″(π/3) = 2 sin(π/3) = √3, which is positive, so f has a local minimum at x = π/3. f″(5π/3) = 2 sin(5π/3) = −√3, which is negative, so f has a local maximum at x = 5π/3. Step five: f(π/3) = π/3 − √3 and f(5π/3) = 5π/3 + √3. Step six: the local minimum is (π/3, π/3 − √3) and the local maximum is (5π/3, 5π/3 + √3).
The worked example shows the rhythm of a high-scoring response. Each step is one or two lines. The decision line for each critical number is explicit: f″(π/3) = √3 > 0, so f has a local minimum at x = π/3. The conclusion line at the end names the ordered pair. The candidate who writes this response in 6 minutes has 90 seconds of slack. The candidate who writes it in 9 minutes does not. AP Courses' AP Calculus BC programme uses a similar worked example as the model response in its free-response review, with rubric annotations on each step, so candidates learn to write the response in the rubric's language rather than in their own.
Connecting the second derivative test to the rest of the BC syllabus
The second derivative test is one of three classification tools in the AP Calculus BC syllabus, sitting alongside the first derivative test and the closed-interval extreme value theorem. The course and exam description places it in the Local Extrema and Concavity unit, which is one of the BC-only units that distinguish the BC exam from AB. The same unit covers inflection points, concavity intervals, and the second derivative test for local extrema; a candidate who is strong on one of these is usually strong on all of them because they share a single underlying idea: the sign of f″ controls concavity, and concavity at a critical number controls the local shape of the graph.
The second derivative test also connects to the curve-sketching FRQ, which is a common BC free-response format. The candidate who can apply the test, identify inflection points, and read concavity intervals from f″ is in a strong position to draw the graph. The test is the gateway to a complete graph analysis, and AP preparation should treat it as a foundational skill rather than an isolated fact. AP Courses' AP Calculus BC programme integrates the test into its curve-sketching unit, so candidates see the connection between classification and graphing in a single block of study time, rather than treating them as separate topics.
For AB candidates, the second derivative test is one of several tools to know for the MCQ section, and the preparation is lighter. For BC candidates, the test is a load-bearing skill for free-response questions on local extrema, and the preparation should be heavier. In both cases, the second derivative test is not memorisable as a formula; it is a method that requires the candidate to know when to use it, what to write, and when to fall back. The candidate who treats it as a method, not a fact, will earn full credit on the test.
Conclusion and next steps
The AP Calculus second derivative test is a high-leverage method: it is fast to apply, it appears in both AB and BC exams, and it earns a clean row of credit when the candidate writes the rubric language. Mastery depends on three habits: knowing the four inconclusive cases, writing the three-part decision line for each critical number, and reporting the local extrema as ordered pairs. For most candidates, the path to a 5 on the BC exam runs through the six-step argument rehearsed on five to seven FRQ-style problems, with a 90-second slack on each. AP Courses' AP Calculus BC programme analyses each student's local-extrema FRQ responses against the rubric, identifies the row where credit is lost, and turns a 5 target into a concrete preparation plan with timed practice on the second derivative test decision line.