Find an unknown from the coefficients
Find an unknown (a constant or n) from given coefficients
- In 9 of 20 recent exams
- 2% of the marks
- Paper 1 or 2
- 4 to 7 marks
How it is set in the exam
Paper 1 (no calculator) and Paper 2 (calculator), 4–7 marks, usually a whole Section A question. Common (9 of 20 exam versions, 6 of them on Paper 2 (calculator)): the commonest topic 1 type outside sequences. The algebra is the same on both papers; on Paper 2 (calculator) the GDC does the arithmetic and, for an unknown n, a table.
Exam phrases to spot (5)
“the coefficient of x3 in the expansion of … is …; find the value(s) of k”
“find the possible values of a”
“find the value of n”
“the coefficient of x5 is twice the coefficient of x4”
“find the value of n and the value of k”
In disguise (5)
- a bracket containing x1 or x: r comes first, from the power of x written as a function of r
- the whole expansion divided by a power of x: the term you need from the bracket has a different power from the one named
- the same bracket raised to two different powers, with one coefficient given for each
- the power of the bracket is the unknown, such as (x2+c)n+1
- the coefficient is given as an expression in the unknown ("… is equal to 3a4"), not as a number
Variants (7)
- a constant in the bracket, one coefficient given, an even power of the constant: ± ( and )
- a constant in the bracket and an odd power of it, after writing a root of x as a power ()
- a bracket over a power of x, the coefficient equal to an expression in the constant, 0 rejected ()
- one coefficient a multiple of another, with a binomial coefficient worked out first ()
- n and k from the coefficients of x and x2 in (1+kx)n: a quadratic in n ()
- n from one coefficient, when n is in the power of a constant too (, a table)
- two constants from the same bracket raised to two powers, with a sign condition ()
Looks like it, but isn't (3)
- no equation to solve: the coefficients are only written down, even when a letter stays in them ("in terms of h"): Expand, or find one term or coefficient
- the coefficients are consecutive terms of a sequence: Expand, or find one term or coefficient, then Three consecutive terms with an unknown
- a probability with an unknown p, or the least number of trials for "at least k" successes: Binomial probabilities: recognise B(n, p), then translate the words; the least number of trials for "at least one" is Find n: which term, how many terms, least n
A constant in the bracket, an even power: keep both signs
The coefficient of x3 in the expansion of (2x−k)5 is 360. Find the possible values of k. [5]
Try it yourself first, then compare.
Now try one like it
Write your working on paper or on screen. Paper Zero marks it step by step, like the worked example, and shows where each mark went. Free, with no account.
| The unknown | Method |
|---|---|
| A constant, with an even power of it in the equation (k2, k4) | ±: keep both, unless a condition (such as k>0) removes one |
| A constant, with an odd power of it (k3, k5) | one real root, and its sign comes out of the equation |
| n, when the equation is linear or quadratic in n ((1n), (2n), or an equation with (3n) that becomes linear or quadratic once a common factor n or n(n−1) cancels: Factorials and nCr) | by hand: write (rn) as a polynomial in n and solve |
| n, in any other equation (a cubic, or n in the power of a constant too) | : a table of the coefficient against n first (table of values); the solver is the alternative (solve an equation) |
| Two unknowns | two equations; eliminate one unknown, by cancelling a common factor such as k2 (Example 2.3) or by dividing one equation by the other (Example 2.4); then substitute back to find the other |
| A constant, when one coefficient is a multiple of another | equate, then factorise out the lower power; the root 0 makes both coefficients 0, so reject it with the question's reason (q=0, q>0). A final answer that keeps 0 loses the last mark. |
n must be a positive integer, and at least every r in the question (in (1+x)n there is no x3 term unless n≥3), so reject negative, non-integer and too-small values. You never need the factor theorem (AHL 2.12) for a cubic in n: at SL a common factor cancels, or uses a table.
- 1
Write the general term
with the unknown in it, and choose r for each power of x you need from the bracket: from the power equation if a term contains a power of x other than x, or both terms contain x. If the whole expansion is multiplied or divided by a power of x, first work out which power of x the bracket must supply (divided by x3, the term in x2 comes from the bracket's term in x5), and divide the coefficient by any number in the divisor too.
- 2
Write each coefficient
as (rn) times the powers of the constants.
- 3
Equate
it to the given value, or to the given multiple of another coefficient. Equating the whole terms, with their powers of x, also earns the method mark; cancel the power of x straight away, because some markschemes give the accuracy mark only to the equation in the unknown alone.
- 4
Solve:
a power equation in the constant; a polynomial in n, by hand when it is linear or quadratic (perhaps once a common factor cancels); on , a table of values for any other equation in n (a cubic, or n also in the power of a constant); or two equations with one unknown eliminated (cancel a common factor, or divide one equation by the other), then substituted back for the other.
- 5
Keep only the values the question allows,
and say why the others go: both signs for an even power unless a condition rejects one; 0 rejected when the question excludes it; n a positive integer, and at least every r you used (a term with br exists only when n≥r: in (1+x)n there is no x3 term unless n≥3).
Only the positive root
when the power of k is even and nothing rules the negative out (Example 2.1): the final mark is lost. Write ±. (Compare Solve a logarithmic equation: ln(x2)=ln9 keeps both signs too, but 2lnx=ln9 keeps only x=3.)
± when it doesn't belong
an odd power has one real root (Example 2.2), and a condition such as a>0 fixes a sign (Example 2.4). The extra value costs the final mark.
A zero left in the answer
when one coefficient is a multiple of another. From 10q5=30q4, factorise, 10q4(q−3)=0, and reject q=0 with the question's reason (these questions state one, such as q=0 or q>0). Leaving q=0 in the answer loses the final mark. (Markschemes also accept dividing by q4 once q=0 is stated, but factorising is the method here, as in Three consecutive terms with an unknown: it shows where the 0 went.)
More mistakes (10)
The wrong power of x
when a term contains a power of x other than x (x2, x1), when both terms contain x, or when the expansion is divided by a power of x, the power named is not r or n−r. Write the power of x as a function of r first.
The number in the divisor forgotten
dividing the expansion by 3x2 lowers each power of x by 2 and divides each coefficient by 3.
The constant's power forgotten
in (2x−k)5 the coefficient of x3 is 10×23×k2, not 10k2.
Binomial coefficients left as symbols in the equation
work each (rn) out before you solve. On its value often has its own ("seen anywhere": (25)=10 in Example 2.1), and the for the equation has gone only to the equation with the numbers worked out, 80k2=360. On , (37)k3=−4.375 has been accepted, but you still need (37)=35 to solve it.
A "show that" for (rn) with no working
write the cancelling, as in (310)=3×2×110×9×8=120, or Pascal's triangle written out down to the row needed. The value alone, or a check that starts from it, loses the mark.
Adding the pieces
of a term instead of multiplying them: nothing scores.
Keeping a negative, non-integer or too-small n
(n=−9 in Example 2.3; n=0 and n=1 if (3n)=2(2n) is multiplied out to n(n−1)(n−8)=0, where both coefficients would be 0 because the terms don't exist): n is a positive integer, and at least the largest r used.
The wrong n when the power is n+1
in (4+x)n+1 the term in x2 is (2n+1)4n−1x2, not (2n)4n−2x2. The binomial coefficient and the power of the constant both use the bracket's power, n+1. If you solve for the power first, subtract 1 at the end.
Reading n off Pascal's triangle with no equations
the answer can still score, but the set-up marks are lost. Write the equations first.
Hunting for a factor of a cubic in n
you never need to (the factor theorem is AHL 2.12). On the equation in n will be linear or quadratic, perhaps once a common factor such as n(n−1) cancels; otherwise, on , use a table.