The AP Calculus integral test for convergence is one of the most tightly rubric-bound convergence tests a candidate can apply to an infinite series. The exam asks for it in both the multiple-choice section and the free-response section, often disguised inside a larger series question that also invokes the comparison test, the alternating series test, or the ratio test. The reason it scores so predictably is that the four preconditions are mechanical, the improper integral itself is a standard row in a calculator section or a non-calculator section, and the final conclusion is binary. A student who understands which rows the reader is hunting for can lose only one or two points even on a difficult item.
This article walks through the AP Calculus integral test for convergence exactly as the rubric reads it. We start with the four preconditions the test demands before the improper integral is even computed, then the integral itself, then the conclusion, and finally the way the scoring separates "converges" from "diverges" when the integral is improper. Worked examples are threaded through each H2 so the reader can see the row-by-row scoring on a real series. The article ends with a tactical MCQ triage and a checklist that turns the integral test into a sequence of checkboxes rather than a moment of judgement.
The four preconditions AP Calculus requires before the integral test is legal
For most candidates reading this, the first mistake happens before any integral is taken. The integral test is not a universal tool. The College Board scoring guidelines treat it as a conditional test, and the four preconditions are listed explicitly on the rubric, often as a single bullet worth one point. If a student applies the integral test to a series that fails one of these conditions, the answer is not "wrong" in a soft sense — it is unscorable, because the test simply does not apply.
The four preconditions, in the order the rubric reads them, are: the terms aₙ must be positive (or at least eventually positive), the function f(x) that interpolates aₙ must be continuous on the relevant interval, f(x) must be decreasing on that interval, and the improper integral must actually exist as a finite or infinite number. AP Calculus typically lists these as a single setup row. If the student is asked "Does the series converge by the integral test?", the first row of credit is for naming the test and checking these four conditions.
Positivity: the aₙ ≥ 0 row
The terms of the series must be non-negative. A series such as ∑ n / (n² + 1) satisfies this because every term is positive. A series like ∑ (-1)ⁿ / n does not, and applying the integral test to it scores zero on the legal-test row. The official rubric on FRQs that mix alternating and non-alternating series is unusually strict on this point. Many students lose the setup point by silently assuming the integral test works on alternating series.
Continuity: the f(x) on [1, ∞) row
The continuous function f(x) must be defined for every real x ≥ 1 (or ≥ N, the index at which the test starts to apply). A denominator that vanishes inside the interval breaks continuity, and the test fails before it begins. The series ∑ 1 / (x² − 1) interpolated as f(x) = 1 / (x² − 1) is discontinuous at x = 1, so the integral test is not applicable. Students who compute the integral anyway, even correctly, do not earn the convergence conclusion on the rubric — the row that awards the conclusion is the row that says "test is legal", and that row is closed by the discontinuity.
Decreasing: the f'(x) < 0 row
The function f(x) must be decreasing on the interval. The rubric accepts either a verbal justification ("f'(x) is negative for x ≥ 1") or an explicit derivative computation. A common AP Calculus MCQ trap is a series whose terms are positive and continuous but not decreasing: ∑ (2 + sin n) / 3ⁿ has terms that oscillate in numerator, but the geometric decay still drives convergence. The integral test is technically legal here, but it gives a complicated integral. A subtler trap is ∑ n / (n² + 1), which is positive, continuous, and decreasing, but only after a finite number of initial terms. The rubric accepts "eventually decreasing" as sufficient, but the test starts at the first index where decreasingness holds.
This section ends with a checkpoint: before writing any integral, the student should be able to fill in four lines — aₙ > 0, f continuous, f decreasing, improper integral definable. The first three are setup rows; the fourth is the substantive calculation that follows.
Setting up the improper integral the rubric actually scores
Once the four preconditions are met, the rubric moves to the integral itself. This is the row that carries the most weight on FRQs and the row that is most often fumbled on MCQs. The integral is improper at infinity, and the standard form the AP Calculus rubric expects is a limit of a definite integral with a finite upper bound. Concretely, for a series ∑ aₙ starting at n = 1, the integral test row is written as
limb → ∞ ∫1b f(x) dx
with the upper bound replaced by b and the limit taken after integration. If the series starts at n = N rather than n = 1, the lower bound of the integral is N, not 1, and the limit of integration is taken from N to b as b → ∞.
The antiderivative row
The antiderivative is a separate row on the rubric. The reader awards the setup point for writing the limit, and a separate point for evaluating the antiderivative correctly. For a series like ∑ 1 / (n² + 1), the antiderivative is arctan(x) plus a constant, and the rubric gives credit for the arctan even if the limit evaluation is missed. For a series like ∑ ln(n) / n², the antiderivative requires integration by parts; the rubric awards one point for the by-parts setup and a second point for the resulting expression.
In my experience this is the row that separates a 4 from a 5 on FRQs that combine the integral test with a comparison or ratio test. The student who knows how to do the integration by parts cleanly scores both rows; the student who struggles with the technique loses the antiderivative row and then loses the conclusion row because the limit cannot be evaluated without it.
The limit row
Once the antiderivative is written, the limit is taken as b → ∞. If the limit is finite, the integral converges, and the series converges. If the limit is infinite or does not exist, the integral diverges, and the series diverges. The rubric scores this as a single row: "limit exists and is finite" versus "limit is infinite or does not exist". There is no partial credit between these two states.
A common student error is to write the limit expression, then substitute infinity into the antiderivative, then stop. The rubric requires the limit notation, and on FRQs a bare substitution without the lim symbol is read as a missing row. Always write limb → ∞ [F(b) − F(1)] explicitly, even when the answer is obviously F(∞) − F(1).
Reading the conclusion row: converges versus diverges
The conclusion row of the AP Calculus integral test for convergence is binary. The series either converges or diverges. The rubric does not award half credit for "the series probably converges". It does not award credit for an unjustified conclusion. It does not award credit for a conclusion that contradicts the integral.
The four legitimate states of the conclusion row are: the integral converges to a finite number, so the series converges; the integral diverges to infinity, so the series diverges; the integral does not exist as a real number, so the test is inconclusive; and one of the four preconditions fails, so the test is inapplicable. Of these four, only the first two are scorable conclusions. The third and fourth are the most common reasons a student who has done the integration correctly still loses the conclusion point.
When the integral test is inconclusive
If the antiderivative does not have a finite limit, the integral test does not prove divergence — it only fails to prove convergence. The student must then move to a different test: the comparison test, the limit comparison test, or the ratio test. A common AP Calculus BC item combines the integral test with a follow-up comparison; the rubric splits the credit so that the integral test is worth one or two rows, and the comparison is worth one or two additional rows.
For example, a series like ∑ 1 / (n ln n) is positive, continuous, and decreasing, but its integral ∫ 1 / (x ln x) dx diverges. The integral test therefore does not prove convergence. The student who writes "diverges by the integral test" loses the conclusion row, because the integral test does not establish divergence — divergence of the integral proves divergence of the series, but here the integral diverges, which is exactly the case where the test does yield a conclusion. The right reading is: the integral diverges, so the series diverges. The rubric reads these two as the same row.
Distinguishing "integral diverges" from "test is inconclusive"
The two states are often confused. The integral test is inconclusive when one of the four preconditions fails; the integral diverges when the preconditions hold and the improper integral equals infinity. The rubric makes this distinction because the student who recognises that the test is inapplicable has demonstrated deeper understanding than the student who pushes the calculation forward regardless. A good FRQ answer always names the test, checks the preconditions, computes the integral, and states the conclusion in that order.
Worked example: the integral test on a non-standard series
Consider the series ∑n=2∞ 1 / (n (ln n)²). This is a classic AP Calculus BC item because it requires a u-substitution inside the integral and a careful limit. The four preconditions are: aₙ = 1 / (n (ln n)²) is positive for n ≥ 2; f(x) = 1 / (x (ln x)²) is continuous on [2, ∞); f(x) is decreasing on [2, ∞) (since both x and ln x are positive and increasing, the denominator grows and the function shrinks); the improper integral is definable. All four preconditions hold, so the test is legal.
The integral is ∫2∞ 1 / (x (ln x)²) dx. Substitute u = ln x, so du = dx / x, and the integral becomes ∫ln 2∞ 1 / u² du. This is a standard antiderivative: −1 / u. The limit as b → ∞ of [−1 / u] from u = ln 2 to u = b is [0 − (−1 / ln 2)] = 1 / ln 2, a finite number. The integral converges, so the series converges.
The rubric would award: one point for naming the integral test and checking preconditions, one point for the u-substitution, one point for the antiderivative, and one point for the conclusion. This is a four-row question, and a clean answer picks up all four.
Worked example: the integral test on a series where the test is inconclusive
Consider ∑n=1∞ sin(n) / n². The four preconditions: aₙ = sin(n) / n². The first precondition — positivity — fails immediately, because sin(n) is negative for many values of n. The integral test is therefore inapplicable. A student who proceeds to compute the integral of sin(x) / x² from 1 to ∞ is doing an interesting calculation, but the rubric does not award the conclusion point, because the test is not legal.
The correct move on the AP Calculus exam is to switch tests. The comparison test, using 0 ≤ |sin(n) / n²| ≤ 1 / n² and the convergent p-series with p = 2, is the natural follow-up. The rubric on a question that combines these two tests awards one point for recognising that the integral test is inapplicable, one point for setting up the comparison, and one or two points for the absolute convergence conclusion.
The tactical lesson is that the integral test is conditional. When the conditions fail, the test is not "wrong" — it is silent. The student who recognises silence and switches to a different test scores the conclusion row; the student who forces the integral test scores nothing on that row.
How the integral test interacts with the comparison and ratio tests on FRQs
On the AP Calculus BC exam, the integral test rarely appears in isolation. It usually appears as one of three or four convergence tests in a multi-part FRQ, and the rubric is designed to separate the points. A typical setup is: part (a) asks for the integral test, part (b) asks for the comparison test, and part (c) asks for the radius or interval of convergence. The student who can move between tests without losing the structure of the answer is the student who scores a 5.
Comparison test as a follow-up to a failed integral test
When the integral test is inconclusive — either because a precondition fails or because the integral diverges — the comparison test is the natural follow-up. The rubric reads this transition as: "the integral test is not applicable, so we use the comparison test with the p-series ∑ 1 / n²". One point is awarded for naming the comparison, one point for the inequality, and one point for the conclusion.
Ratio test as an alternative when the integral test is awkward
For series whose terms involve factorials or exponentials, the integral test is usually possible but produces an unwieldy integral. The ratio test is almost always cleaner. The AP Calculus rubric allows the student to choose whichever test is most efficient, but the integral test is preferred only when the resulting integral is a standard form (arctan, ln, 1/x, 1/x², or arctan derivatives). When the integral becomes a special function (erf, Ei, Si), the ratio test is the better choice.
Id personally pick the ratio test on a series like ∑ n! / nⁿ, because the integral of (x!)/x^x does not have a closed form. The rubric rewards the correct test choice with full credit on the conclusion row, regardless of which test is used. There is no penalty for choosing the ratio test over the integral test, as long as the chosen test is applied correctly.
Common pitfalls and how to avoid them on the integral test
The most common AP Calculus integral test errors fall into a small number of patterns. Recognising them in advance saves the conclusion row on a difficult item.
- Forgetting the positivity check. Many students apply the integral test to alternating series without noticing. The test is inapplicable to series whose terms are not eventually positive. The fix is to write the positivity row explicitly on the FRQ: "aₙ = f(n) is positive for n ≥ 1". If you cannot write this row, the test is not legal.
- Misidentifying the lower bound. For a series that starts at n = 2, the integral is from 2, not 1. For a series with a singularity inside the interval, the integral must be split. The fix is to read the first index of summation before writing the integral.
- Dropping the limit notation. On FRQs, bare substitution of infinity into the antiderivative loses the limit row. The fix is to write limb → ∞ F(b) explicitly.
- Concluding without justification. "The series converges" without naming the test loses the conclusion row. The fix is to write "by the integral test, the series converges" or "the series diverges because the integral diverges".
- Confusing divergence of the integral with inconclusiveness. If the integral diverges, the series diverges. The test is not inconclusive in this case; the conclusion is that the series diverges. The fix is to read the divergence case as a positive result of the test, not a failure.
MCQ triage for the integral test under time pressure
The AP Calculus multiple-choice section gives roughly 90 seconds per question, and the integral test items are designed to be tractable in that window. A useful triage is to identify the answer choice that violates a precondition first. If one of the answer choices is "the integral test is inapplicable because f(x) is not decreasing", that is often the correct choice on a tricky MCQ, even when the integral is computable. The exam writers know that students will compute the integral; the test-legal choice is the one that catches the precondition failure.
For a quick comparison of the integral test against the other major convergence tests, the table below summarises when each is preferred.
| Test | Best applied when | Common AP Calculus trap | Rubric rows typically awarded |
|---|---|---|---|
| Integral test | aₙ = f(n) with f integrable in closed form | Applying to alternating or non-positive series | 1 setup + 1 antiderivative + 1 limit + 1 conclusion |
| Comparison test | aₙ is bounded by a known convergent or divergent series | Using a series that does not actually bound aₙ | 1 comparison + 1 inequality + 1 conclusion |
| Limit comparison test | Direct comparison is awkward but a ratio with a known series is computable | Choosing the wrong comparison series (p-series vs geometric) | 1 ratio setup + 1 limit + 1 conclusion |
| Ratio test | Terms involve factorials or exponentials | Misidentifying the limit as greater than 1 when it is exactly 1 | 1 ratio setup + 1 limit + 1 conclusion |
| Alternating series test | Series is of the form ∑ (-1)ⁿ bₙ with bₙ decreasing to 0 | Forgetting the limit-to-zero condition | 1 alternation + 1 decreasing + 1 limit + 1 conclusion |
This table is the operational summary of the integral test for convergence on the AP Calculus exam. For most candidates, the integral test is the right choice when the antiderivative is standard, the comparison test is the right choice when a clean bound exists, the limit comparison test is the fallback when neither of those is obvious, and the ratio test is reserved for factorial-and-exponential series. The alternating series test is the only one that handles sign changes, and the integral test is the only one that requires a closed-form antiderivative.
Building a preparation plan around the integral test for convergence
A focused AP Calculus preparation strategy for the integral test should be built around three pillars: preconditions, antiderivative technique, and conclusion writing. The preconditions pillar is a series of short MCQ drills — given a series, name which preconditions hold and which fail. The antiderivative pillar is integration drill on the standard forms (1/x, 1/x², ln x / x, 1 / (x² + 1), 1 / √(1 − x²)). The conclusion pillar is a writing drill — given a worked integral, write the conclusion row in the exact form the rubric wants.
A reasonable time budget for a 12-week AP Calculus BC preparation plan is roughly 90 minutes per week on series convergence, with the integral test receiving about a third of that time. The other two-thirds goes to the comparison and ratio tests. For AP Calculus AB, the integral test appears less frequently; the same 30 minutes per week is usually sufficient, with the focus on MCQ triage rather than FRQ depth.
Scoring on the AP Calculus exam is on a 1–5 scale, and the integral test contributes to roughly one of the six FRQ points on the BC exam and two to three MCQ points across the two sections. A student who has the integral test mastered can usually pick up all of these points, which translates to a meaningful lift on the composite score. In practice, students who lose the conclusion row on the integral test also tend to lose it on the comparison and ratio tests, so fixing this single row tends to improve the entire series-convergence block of the exam.
Conclusion and next steps
The AP Calculus integral test for convergence is a four-row test: preconditions, antiderivative, limit, conclusion. The preconditions row is the most often missed, because students skip directly to the integral. The conclusion row is the most often mis-stated, because students write "converges" without naming the test. Mastering the integral test means writing all four rows explicitly on every FRQ and using the precondition check as a triage tool on every MCQ.
AP Courses' one-to-one AP Calculus BC programme walks each student through the four-row structure of the integral test on a set of curated FRQs, scores the conclusion row against the official rubric, and turns a target score of 5 into a concrete series of drills on the preconditions, the standard antiderivatives, and the conclusion-writing template.