Partial-fraction decomposition is the single algebraic technique on the AP Calculus BC syllabus that most cleanly separates a 4 from a 5, and the reason is not that the integration is hard. The integrals themselves are linear, logarithmic, or arctangent — the kind a strong AB student could finish in two minutes. What costs the row, on a six-point FRQ written under timed conditions, is the algebra that has to come first. The set-up row, the coefficients row, and the constant-of-integration row each carry their own rubric logic, and BC candidates who treat partial fractions as a single technique rather than a sequence of graded micro-steps tend to drop a full point for an error that, in isolation, would have been a free mark.
This article walks through how the College Board rubric actually scores a partial-fraction problem on the AP Calculus BC FRQ. It covers the three denominator shapes the exam can hand you, the cover-up versus equating-coefficients decision a candidate has to make under time pressure, the four integration families that the decomposition unlocks, and the rows of the rubric where points are most often lost. The goal is concrete: by the end, a BC candidate should know exactly which algebraic move earns which point, and which short-cuts are unsafe to take on a graded response.
Why partial fractions are a BC-only topic and where they appear on the FRQ
Partial-fraction decomposition sits inside Unit 8 of the AP Calculus BC course framework, the unit that covers advanced integration techniques. It is the only integration topic on the BC exam that AB candidates are not expected to know, and the Course and Exam Description is unusually explicit about its scope. The unit description names three specific techniques: integration by parts, partial fractions, and the use of a CAS when appropriate. Of those three, partial fractions is the one that most reliably appears in a scored FRQ slot, because it is the only one whose assessment is fully algebraic — a grader can mark it in seconds, and a 1, 2, or 3 can be awarded without ambiguity.
On a typical AP Calculus BC exam, a partial-fraction problem shows up in one of two places. The first is a stand-alone question worth three to four points, usually early in the FRQ section, asking the student to evaluate a definite integral whose integrand is a rational function with a non-reducible denominator. The second is a sub-part of a larger accumulation or area problem, where the decomposition is the bridge between a setup the student has already completed and an evaluation row that the rubric will then score separately. The second format is the one most BC candidates underestimate: a student who would clear a stand-alone partial-fraction problem in four minutes can still lose a point inside a five-part question, because they read the setup as a single move rather than two graded moves.
The exam format is worth understanding here, because it shapes the scoring. The AP Calculus BC exam has two sections. Section I is multiple choice, 45 questions in 1 hour 45 minutes, with 30 of those questions being non-calculator. Section II is free response, six questions in 1 hour 30 minutes, with two of the six being calculator-active. A partial-fraction problem can technically appear in either section, but in practice the rubric only allows it to earn a partial-fraction row on the FRQ, because the multi-step algebra is what makes it a scoring problem. The College Board has historically placed it on the FRQ rather than the multiple choice, and the FRQ is where a student practicing for a 5 should expect to demonstrate competence with the technique.
For a student mapping a preparation strategy, the takeaway is that partial fractions is a low-volume, high-density topic. It will not appear in every BC administration, and on a given exam it is usually confined to a single FRQ or a single multi-part question. But the points it controls are disproportionately weighted relative to the study time it requires, because the rubric awards points for each algebraic micro-step rather than for the final answer. A candidate who has spent ten hours on polar area problems may never recover the time, but a candidate who has spent three hours on partial-fraction set-ups will see those points compound across multiple FRQs.
The three denominator shapes the AP Calculus BC rubric expects you to recognise
Before any integration happens, the grader is looking for one thing: did the student read the denominator and write the correct form of the decomposition? This is the row on the rubric that most candidates lose, because it is purely algebraic and it precedes the calculus. A student who writes a decomposition that does not match the denominator's factorisation is not in a position to earn any subsequent row, no matter how clean the integrals that follow.
The three shapes are these. The first is a product of distinct linear factors, the form (x − a)(x − b) or (ax + b)(cx + d) with no squared terms and no repeated roots. The decomposition is a sum of single constants over single factors, one term per factor. The second is a product that includes a repeated linear factor, the form (x − a)² or (x − a)²(x − b). The decomposition here needs a term with the highest power in the denominator plus a separate term for each lower power, because the algebraic identity requires it. A student who writes A/(x − a)² alone for a repeated factor has forgotten the second numerator and the row is gone. The third shape is a factor that does not factor further over the reals, an irreducible quadratic such as x² + 4 or x² + x + 1. The numerator over such a factor must be linear, Bx + C, never a single constant, because the polynomial long-division step that the rubric assumes in the background cannot reduce the degree of a linear numerator over a quadratic denominator.
The way to read a denominator quickly is to factor it, count the distinct linear factors, count the multiplicities, and then write a decomposition template before plugging in any numbers. The cover-up method, which I will return to in the next section, only works cleanly when the template is right, and a template error is the single most common source of a zero on a partial-fraction problem.
Worked example: distinct linear factors
Consider 1/((x − 1)(x + 2)). The correct decomposition is A/(x − 1) + B/(x + 2). The rubric awards a point for this form. To find A, multiply both sides by (x − 1) and evaluate at x = 1, giving A = 1/3. To find B, multiply both sides by (x + 2) and evaluate at x = −2, giving B = −1/3. The integral becomes (1/3) ∫ dx/(x − 1) − (1/3) ∫ dx/(x + 2), both logarithmic, both straightforward, both earning their own rubric row. This is the clean case, and it is the only case where the cover-up method works without modification. A student who tries to apply cover-up to a repeated factor or to an irreducible quadratic will produce an algebraically wrong answer that the grader will mark in a single pass.
Worked example: repeated linear factor
Consider 1/(x²(x − 3)). The decomposition is A/x + B/x² + C/(x − 3). The student must write all three terms. The cover-up method applied to x² alone is not enough: setting x = 0 yields 0, which is useless, so B must be found by an alternative route — either by multiplying through and equating coefficients, or by differentiating and then substituting. For C, cover-up works: set x = 3 to get C = 1/9. The integration step is three separate logarithmic terms, one of which is ∫ dx/x², a power rule integral that the rubric expects the student to handle without comment. The repeated factor row is where a BC candidate either demonstrates that they understand why the decomposition has its specific shape, or reveals that they are pattern-matching without the underlying algebra.
Worked example: irreducible quadratic
Consider 1/(x(x² + 4)). The decomposition is A/x + (Bx + C)/(x² + 4). The numerator over the quadratic must be linear. To find A by cover-up, set x = 0 and read off A = 1/4. For B and C, multiply through and equate coefficients of x², x, and the constant, then solve. The integral splits into a logarithmic piece from A/x, and the harder piece ∫ (Bx + C)/(x² + 4) dx, which becomes a logarithm plus an arctangent. The arctangent is the part of the problem that most clearly signals BC-level work, and the rubric row that awards the point for recognising the form (1/a) arctan(x/a) is one of the rows most often missed.
How the cover-up method and the equating-coefficients method are scored
Once the template is written, the student has to find the coefficients. The AP Calculus BC rubric accepts two methods. The first, cover-up (also called Heaviside), is faster and works perfectly for distinct linear factors. The second, equating coefficients by multiplying through by the common denominator, is slower but works for repeated factors, irreducible quadratics, and any case where cover-up gives a zero or an undefined value. The rubric does not award a point for the method itself — it awards a point for the correct coefficient values. But in practice, the method choice determines whether the student can find those values inside the time budget.
Cover-up is a one-step-per-coefficient method. To find the numerator of a particular linear factor, multiply both sides of the decomposition by that factor and substitute the root. The result is the coefficient, and the work takes about fifteen seconds per coefficient. For a three-factor distinct linear decomposition, the entire coefficient-finder step takes under a minute. The catch is that cover-up only works on distinct linear factors. A repeated factor like (x − 1)² produces a zero when you substitute x = 1, so the method returns nothing useful. An irreducible quadratic cannot be cover-up'd at all, because the factor never evaluates to zero over the reals.
Equating coefficients is the universal method. Multiply the decomposition through by the common denominator, expand the right side, and match coefficients of like powers on both sides. This produces a system of linear equations, which the student solves by substitution or elimination. The method is algebraically heavy and slow by hand, but it always works, and on a calculator-active FRQ section a student with a CAS can solve the system in seconds. The rubric row that follows is independent of the method: the grader is looking for the correct numerical values, written in a position where the rest of the work can read them.
The scoring trap in this section is a coefficient that depends on a previous coefficient. When a student makes an error in the first A and carries it into the expression for B, the rubric does not award partial credit for the second coefficient in the usual sense — the second value is treated as correct only if it follows from correct previous work. This is a subtle rubric point, and it is one of the reasons the College Board has historically insisted that the decomposition be shown before the integration. The grader needs to be able to see, line by line, which step produced which value, so that errors can be isolated and remaining rows can still be earned.
The four integration families the decomposition unlocks
Once the coefficients are known, the integral is a sum of four families, each with its own rubric row. The first family is ∫ dx/(x − a), a logarithm. The second is ∫ dx/(x − a)ⁿ for n ≥ 2, a power rule integral. The third is ∫ (Bx + C)/(x² + a²) dx, which splits into a logarithm and an arctangent. The fourth is ∫ (Bx + C)/((x − a)² + b²) dx, which also splits but with a completed-square denominator. A BC candidate should be able to recognise each family on sight and write down the antiderivative template without re-deriving it.
For the logarithm family, the rule is ∫ dx/(x − a) = ln|x − a| + C. The absolute value is required for the rubric row, and so is the constant. For the power-rule family, ∫ dx/(x − a)ⁿ = (x − a)⁻ⁿ⁺¹/(−n + 1) + C for n ≠ 1. A student who attempts to integrate (x − 1)² by logarithms will lose the row, because the rubric expects the recognition that the exponent on the denominator determines the family. For the arctangent family, the rule is ∫ dx/(x² + a²) = (1/a) arctan(x/a) + C, and the coefficient 1/a is the row that most students forget. For a linear numerator Bx + C over x² + a², the integration splits into (B/2) ln(x² + a²) and (C/a) arctan(x/a), both with their own constants, both with their own rows.
The reason the rubric awards separate rows for these families is that they are the discrete skills a BC student is supposed to demonstrate. A student who integrates the logarithmic part correctly and the arctangent part incorrectly can still earn the logarithmic row, and that is the kind of partial credit that turns a 2 into a 3 on a three-point problem. The integration rows are independent, and a good preparation strategy is to drill each family in isolation before combining them in a single integrand.
How definite integrals with partial fractions are graded differently from indefinite ones
On a definite integral FRQ, the decomposition is the same and the integration is the same, but the evaluation row is new. The student has to substitute the bounds into the antiderivative, compute the difference, and write the answer as an exact value. The exact-value-versus-decimal distinction is something the rubric is firm about, and a student who writes a decimal when the question asks for an exact value loses a point even if the decimal is correct to the displayed precision.
For a partial-fraction definite integral, the bounds can be substituted into each of the four family templates in turn. The logarithm template evaluates to ln|b − a| − ln|c − a|, the power-rule template to (b − a)⁻ⁿ⁺¹/(−n + 1) − (c − a)⁻ⁿ⁺¹/(−n + 1), and the arctangent template to (1/a) arctan(b/a) − (1/a) arctan(c/a). The student has the option to simplify using log laws, but the rubric does not require it; what it requires is that the evaluation row shows the substitution explicitly, so that a grader can follow it. A student who writes a single line of arithmetic without showing the substitution is in a position where a grader cannot award the row, because the work is invisible.
Another definite-integral subtlety is the constant of integration. On a definite integral, the +C is conventionally not written, because it cancels in the subtraction. The College Board rubric reflects this convention. A student who writes +C inside a definite integral does not lose a point, but a student who treats the constant as a graded object and tries to evaluate it at a bound is wasting time and creating a place for an arithmetic error. The clean way to handle the constant on a definite partial-fraction problem is to omit it.
Common pitfalls and how to avoid them
The first pitfall is a wrong template. A student who writes A/(x − 1) + B/(x + 2) for a denominator of (x − 1)²(x + 2) has written a decomposition for a different problem, and the rubric cannot award any row from that point forward. The fix is mechanical: factor first, count the powers, write the template, then find the coefficients. Skipping the factoring step is the most expensive short-cut on the exam.
The second pitfall is a wrong numerator over an irreducible quadratic. A student who writes A/(x² + 4) instead of (Bx + C)/(x² + 4) is treating a degree-two denominator as if it were a degree-one factor. The integration that follows will not produce the required arctangent, and the row for the arctangent family will be lost. The fix is to remember that any quadratic factor that does not split must have a linear numerator, and the linear numerator is the only way to produce the arctangent term.
The third pitfall is forgetting the 1/a in front of the arctangent. The formula ∫ dx/(x² + a²) = arctan(x) + C is the AB version and is wrong when a ≠ 1. The BC rubric expects (1/a) arctan(x/a). A student who writes arctan(x/2) for a = 2 has lost the coefficient, and the row is gone.
The fourth pitfall is dropping the absolute value in the logarithm. ∫ dx/(x − 1) = ln(x − 1) + C is a domain error that the rubric penalises. The form ln|x − 1| + C is the only one that earns the row on a graded FRQ. A student practising for a 5 should treat the absolute value as part of the template, not as a stylistic choice.
The fifth pitfall is failing to match coefficients when the cover-up method is not available. A student who tries to cover-up an irreducible quadratic gets a non-zero number that is not the coefficient, and a student who tries to cover-up a repeated factor gets zero. The fix is to switch methods: multiply through, expand, match coefficients, solve. The cover-up method is a tool, not a universal rule.
Pacing, preparation, and the role of the CAS on a partial-fraction FRQ
On a calculator-active FRQ section, the student has access to a CAS, and the temptation is to let the calculator do the decomposition. The CAS will produce the correct coefficients in a fraction of a second, and on a problem where the only graded skill is the integration, this is the right move. But on a problem where the decomposition itself is part of the scoring, the student has to show the algebraic work by hand, because the rubric needs to see the template and the coefficients. A calculator answer with no supporting work will not earn the rows that depend on the set-up.
The best preparation strategy is to drill the template step by hand on a sequence of denominators, getting fast enough that the algebra takes under two minutes for a standard problem. The cover-up method, applied only where it works, is the fastest path through the coefficient-finding step, and the equating-coefficients method is the fallback when cover-up is not available. Time yourself on three or four problems per family, and the speed will come.
Scoring on a partial-fraction problem is usually distributed across three or four rows: one for the correct decomposition, one for the correct coefficients, one or two for the integration, and one for the final answer. A student who can earn three of those rows is already at the boundary of a 4, and a student who can earn all four is at the boundary of a 5. The points are not given for elegance, they are given for the rubric rows, and a student who knows the rows is a student who can recover from a single mistake.
How partial fractions connect to the rest of the BC syllabus
Partial fractions is rarely an isolated question on a BC exam. The technique typically appears as a step inside a larger problem: an area calculation, a volume of revolution, an accumulation function, or a logistic differential equation. The decomposition is the bridge between the set-up the student has already completed and the evaluation row that the rest of the rubric will score. A student who cannot do the decomposition cannot finish the larger problem, and a student who can do the decomposition but takes six minutes to do it has used up the time budget for the rest of the question.
The connection to differential equations is the most important one. The logistic model dy/dt = ky(1 − y/M) separates, but the resulting integral 1/(y(1 − y/M)) dy requires partial fractions. The arctan model in a second-order differential equation produces a quadratic denominator that may need decomposition. The BC syllabus includes both, and the partial-fraction row is the common prerequisite. A student who has drilled partial fractions as a stand-alone topic will find these larger problems tractable; a student who has not will find them impossible regardless of the differential equation content.
The connection to improper integrals is also worth noting. A definite integral with an infinite bound or a vertical asymptote can be evaluated by partial fractions, and the rubric treats the limit step as a separate row. A student practising for a 5 should be able to combine a partial-fraction decomposition with an improper-integral limit evaluation in a single FRQ, and the rows for both should be earned in a single pass through the rubric.
Final preparation check: a worked BC-style FRQ with full rubric annotation
Consider the problem: evaluate the definite integral from 0 to 1 of 1/((x + 1)(x² + 1)) dx. The denominator is a product of a linear factor and an irreducible quadratic. The correct decomposition is A/(x + 1) + (Bx + C)/(x² + 1). The first row of the rubric is the form, and the student earns it by writing the template. The second row is the coefficients, found by cover-up for A and by equating for B and C: A = 1/2, B = −1/2, C = 1/2. The third row is the integration of A/(x + 1) to (1/2) ln|x + 1|, with the absolute value. The fourth row is the integration of (Bx + C)/(x² + 1) to a sum of (B/2) ln(x² + 1) and C arctan(x). The fifth row is the evaluation at 1 minus the evaluation at 0, with the simplification to an exact value. Each row is independent, and a student who loses one row can still earn the others.
The final answer is (1/2) ln 2 + (π/8). The student who writes this as a decimal loses the row, because the question asks for an exact value. The student who writes it as a sum of two terms with ln and arctan symbols earns the row, and the grader can check the work in seconds.
A BC candidate practising for a 5 should be able to do the above problem in under seven minutes on a calculator-active section, with the decomposition and coefficients taking about half that time. If the decomposition is taking longer than two minutes, the student needs more drill on the template step. If the integration is taking longer than two minutes, the student needs more drill on the four families. The time budget is a diagnostic, and a timed practice session will tell the student which row to focus on next.
Conclusion and next steps
Partial-fraction integration on the AP Calculus BC FRQ is a sequence of small algebraic moves, each with its own rubric row, and the scoring rewards students who treat it as a sequence rather than a single technique. The denominator shape determines the template, the method chosen determines how the coefficients are found, and the four integration families determine the antiderivative form. A student who can read a denominator, write the correct decomposition, find the coefficients, integrate each family, and evaluate a definite integral in under seven minutes is in a strong position to earn the full row allocation on the FRQ.
For students targeting a 5, the next concrete step is a focused drill on the template step across all three denominator shapes, with a stopwatch. Ten problems, four minutes each, and the algebraic set-up will become automatic. AP Courses' one-to-one AP Calculus BC programme builds a per-student decomposition error log from timed FRQ practice, identifying which denominator shape is costing the most rubric rows and turning the partial-fraction set-up into a reliable scored step.