CAIE FP2 2016 June — Question 10 11 marks

Exam BoardCAIE
ModuleFP2 (Further Pure Mathematics 2)
Year2016
SessionJune
Marks11
PaperDownload PDF ↗
Mark schemeDownload PDF ↗
TopicLinear regression
TypeRelate two regression lines
DifficultyStandard +0.3 This is a straightforward application of standard regression formulas from A-level Further Maths statistics. Students need to recall and apply formulas for correlation coefficient and regression lines, then interpret results. All steps are routine calculations with no conceptual challenges or novel problem-solving required, making it slightly easier than average.
Spec5.08a Pearson correlation: calculate pmcc5.09c Calculate regression line5.09e Use regression: for estimation in context

10 For a random sample of 6 observations of pairs of values \(( x , y )\), where \(0 < x < 21\) and \(0 < y < 14\), the following results are obtained. $$\Sigma x ^ { 2 } = 844.20 \quad \Sigma y ^ { 2 } = 481.50 \quad \Sigma x y = 625.59$$ It is also found that the variance of the \(x\)-values is 36.66 and the variance of the \(y\)-values is 9.69 .
  1. Find the product moment correlation coefficient for the sample.
  2. Find the equations of the regression lines of \(y\) on \(x\) and \(x\) on \(y\).
  3. Use the appropriate regression line to estimate the value of \(x\) when \(y = 6.4\) and comment on the reliability of your estimate.

Question 10:
Part (i):
AnswerMarks Guidance
Answer/WorkingMarks Guidance
\((\sum x)^2 = 6(844.20 - 6 \times 36.66) = 3775.44\)M1 A1 Find \(\sum x\) and \(\sum y\) (M1 for either)
\(\sum x = 61.2\) or \(\bar{x} = 10.2\) SC: Allow M1 if 5 used instead of 6 (max 4/11), giving \(62.97\), \(50.97\) and \(r = 0.961[5]\)
\((\sum y)^2 = 6(481.5 - 6 \times 9.69) = 2540.16\)A1
\(\sum y = 50.4\) or \(\bar{y} = 8.4\)
\(S_{xy} = 625.59 - 61.2 \times 50.4/6 = 111.51\) Find correlation coefficient \(r\)
\(S_{xx} = 844.20 - 61.2^2/6\) or \(6 \times 36.66 = 219.96\)M1 A1
\(S_{yy} = 481.50 - 50.4^2/6\) or \(6 \times 9.69 = 58.14\)
\(r = S_{xy}/\sqrt{(S_{xx}\,S_{yy})} = 0.986\) Total: [5]
Question (ii):
AnswerMarks Guidance
Answer/WorkingMarks Guidance
\(b_1 = S_{xy} / S_{xx} = 0.507\)B1 Calculate gradients in both lines
\(b_2 = S_{xy} / S_{yy}\) or \(r^2/b_1 = 1.918\)
\(y = 50.4/6 + 0.507(x - 61.2/6) = 0.507x + 3.23\)M1 A1 Find one regression line, e.g. \(y\) on \(x\)
\(x = 61.2/6 + 1.918(y - 50.4/6) = 1.92y - 5.91\)A1 Find other regression line, e.g. \(x\) on \(y\)
[4]SC: Use PA \(-1\) if results only correct to 2 s.f.
Question (iii):
AnswerMarks Guidance
Answer/WorkingMarks Guidance
\(x = 6.36\) (allow \(6.37\) or \(6.38\))B1 Find \(x\) when \(y = 6.4\)
Reliable since \(r \approx 1\) [\(\sqrt{}\) on \(r\)] *or* \(6.4\) within range of \(y\)B1 Valid comment on reliability
[2]
Question 11A:
AnswerMarks Guidance
Answer/WorkingMarks Guidance
\(W \times a\sin\Phi + 2W \times (3a/2)\sin\Phi = T \times AE\)M1 A1 Take moments for rod about \(A\) (AEF)
\(W \times a\sin 2\theta + 2W \times (3a/2)\sin 2\theta = T \times 2a\sin\theta\)A1 Substituting \(\Phi = 2\theta\) and \(AE = 2a\sin\theta\)
\(T = 2W\sin 2\theta / \sin\theta\) \([= 4W\cos\theta]\)A1 Hence find tension \(T\) in terms of \(\theta\)
\(T = 3W(BC - 3a/2)/(3a/2) = W(8\cos\theta - 3)\)M1 A1 A1 Find \(T\) in terms of \(\theta\) using Hooke's Law
\(4\cos\theta = 8\cos\theta - 3\) \(\cos\theta = \frac{3}{4}\) A.G.M1 A1 Equate to find \(\cos\theta\)
\(T = 3W\)B1 Hence find \(T\) (allow assumption of \(\cos\theta = \frac{3}{4}\))
[10]
\(X = T\sin\theta \left[= 3W\sqrt{7}/4\right]\)B1 Find horizontal force \(X\) at \(A\) (ignore sign)
\(Y = 3W - T\cos\theta \left[= 3W/4\right]\) Find vertical force \(Y\) at \(A\) (ignore sign)
\(\sqrt{X^2 + Y^2} = \frac{3}{\sqrt{2}}W\) *or* \(3\sqrt{2}/2\ W\) *or* \(2.12\ W\)M1 A1 Find magnitude of force at \(A\)
[4]
Question 11B:
AnswerMarks Guidance
Answer/WorkingMarks Guidance
\(\Sigma x = 12\bar{x} = 12 \times \frac{1}{2}(25.17 + 26.83) = 12 \times 26 = 312\)M1 A1 M1 Find \(\Sigma x\) via sample mean \(\bar{x}\)
\(t\ s_A / \sqrt{12} = \frac{1}{2}(26.83 - 25.17)\), \(s_A = 0.83\sqrt{12}/2.201 = 1.306\) or \(s_A^2 = 1.706\)A1 Find estimated s.d. \(s_A\) or \(s_A^2\); with \(t = t_{11,\ 0.975} = 2.201\) (to 3 s.f.)
\(s_A^2 = \{\Sigma x^2 - (\Sigma x)^2/12\}/11\); \(\Sigma x^2 = 11s_A^2 + (\Sigma x)^2/12 = 8130[.8]\)M1 A1 Find \(\Sigma x^2\) from \(s_A\) or \(s_A^2\) (M0 for \(s_A^2 = \{\ldots\}/12\))
[6]
\(H_0: \mu_A = \mu_B\), \(H_1: \mu_A > \mu_B\)B1 State hypotheses
\(s_B^2 = (4507.62 - 177^2/7)/6 = [5.341]\)M1 Estimate \(B\)'s population variance (to 3 d.p.); allow biased: \(4.578\)
\(s^2 = (11s_A^2 + 6s_B^2)/17 = (18.77 + 32.05)/17 = 2.989\) or \(1.729^2\) (to 3 s.f.)M1 A1 Find pooled estimate of common variance
\(t = (26 - 177/7)/s\sqrt{1/12 + 1/7} = (26 - 25.29)/s\sqrt{1/12 + 1/7} = 0.7143/0.8222 = 0.869\)M1 A1 *B1 Calculate value of \(t\) (or \(-t\)) (to 3 s.f.)
\(t_{17,\ 0.9} = 1.333\) State or use correct tabular \(t\) value (to 3 s.f.)
\(t <\) tabular value, so Petra's belief not supported *or* wing span of \(A\) not greaterB1\(\checkmark\) Consistent conclusion (AEF, \(\sqrt{}\) on \(t\), dep *B1)
[8]
## Question 10:

### Part (i):
| Answer/Working | Marks | Guidance |
|---|---|---|
| $(\sum x)^2 = 6(844.20 - 6 \times 36.66) = 3775.44$ | M1 A1 | Find $\sum x$ and $\sum y$ (M1 for either) |
| $\sum x = 61.2$ or $\bar{x} = 10.2$ | | **SC**: Allow M1 if 5 used instead of 6 (max 4/11), giving $62.97$, $50.97$ and $r = 0.961[5]$ |
| $(\sum y)^2 = 6(481.5 - 6 \times 9.69) = 2540.16$ | A1 | |
| $\sum y = 50.4$ or $\bar{y} = 8.4$ | | |
| $S_{xy} = 625.59 - 61.2 \times 50.4/6 = 111.51$ | | Find correlation coefficient $r$ |
| $S_{xx} = 844.20 - 61.2^2/6$ or $6 \times 36.66 = 219.96$ | M1 A1 | |
| $S_{yy} = 481.50 - 50.4^2/6$ or $6 \times 9.69 = 58.14$ | | |
| $r = S_{xy}/\sqrt{(S_{xx}\,S_{yy})} = 0.986$ | | Total: [5] |

# Question (ii):

| Answer/Working | Marks | Guidance |
|---|---|---|
| $b_1 = S_{xy} / S_{xx} = 0.507$ | B1 | Calculate gradients in both lines |
| $b_2 = S_{xy} / S_{yy}$ or $r^2/b_1 = 1.918$ | | |
| $y = 50.4/6 + 0.507(x - 61.2/6) = 0.507x + 3.23$ | M1 A1 | Find one regression line, e.g. $y$ on $x$ |
| $x = 61.2/6 + 1.918(y - 50.4/6) = 1.92y - 5.91$ | A1 | Find other regression line, e.g. $x$ on $y$ |
| | [4] | SC: Use PA $-1$ if results only correct to 2 s.f. |

# Question (iii):

| Answer/Working | Marks | Guidance |
|---|---|---|
| $x = 6.36$ (allow $6.37$ or $6.38$) | B1 | Find $x$ when $y = 6.4$ |
| Reliable since $r \approx 1$ [$\sqrt{}$ on $r$] *or* $6.4$ within range of $y$ | B1 | Valid comment on reliability |
| | [2] | |

# Question 11A:

| Answer/Working | Marks | Guidance |
|---|---|---|
| $W \times a\sin\Phi + 2W \times (3a/2)\sin\Phi = T \times AE$ | M1 A1 | Take moments for rod about $A$ (AEF) |
| $W \times a\sin 2\theta + 2W \times (3a/2)\sin 2\theta = T \times 2a\sin\theta$ | A1 | Substituting $\Phi = 2\theta$ and $AE = 2a\sin\theta$ |
| $T = 2W\sin 2\theta / \sin\theta$ $[= 4W\cos\theta]$ | A1 | Hence find tension $T$ in terms of $\theta$ |
| $T = 3W(BC - 3a/2)/(3a/2) = W(8\cos\theta - 3)$ | M1 A1 A1 | Find $T$ in terms of $\theta$ using Hooke's Law |
| $4\cos\theta = 8\cos\theta - 3$ $\cos\theta = \frac{3}{4}$ **A.G.** | M1 A1 | Equate to find $\cos\theta$ |
| $T = 3W$ | B1 | Hence find $T$ (allow assumption of $\cos\theta = \frac{3}{4}$) |
| | [10] | |
| $X = T\sin\theta \left[= 3W\sqrt{7}/4\right]$ | B1 | Find horizontal force $X$ at $A$ (ignore sign) |
| $Y = 3W - T\cos\theta \left[= 3W/4\right]$ | | Find vertical force $Y$ at $A$ (ignore sign) |
| $\sqrt{X^2 + Y^2} = \frac{3}{\sqrt{2}}W$ *or* $3\sqrt{2}/2\ W$ *or* $2.12\ W$ | M1 A1 | Find magnitude of force at $A$ |
| | [4] | |

# Question 11B:

| Answer/Working | Marks | Guidance |
|---|---|---|
| $\Sigma x = 12\bar{x} = 12 \times \frac{1}{2}(25.17 + 26.83) = 12 \times 26 = 312$ | M1 A1 M1 | Find $\Sigma x$ via sample mean $\bar{x}$ |
| $t\ s_A / \sqrt{12} = \frac{1}{2}(26.83 - 25.17)$, $s_A = 0.83\sqrt{12}/2.201 = 1.306$ or $s_A^2 = 1.706$ | A1 | Find estimated s.d. $s_A$ or $s_A^2$; with $t = t_{11,\ 0.975} = 2.201$ (to 3 s.f.) |
| $s_A^2 = \{\Sigma x^2 - (\Sigma x)^2/12\}/11$; $\Sigma x^2 = 11s_A^2 + (\Sigma x)^2/12 = 8130[.8]$ | M1 A1 | Find $\Sigma x^2$ from $s_A$ or $s_A^2$ (M0 for $s_A^2 = \{\ldots\}/12$) |
| | [6] | |
| $H_0: \mu_A = \mu_B$, $H_1: \mu_A > \mu_B$ | B1 | State hypotheses |
| $s_B^2 = (4507.62 - 177^2/7)/6 = [5.341]$ | M1 | Estimate $B$'s population variance (to 3 d.p.); allow biased: $4.578$ |
| $s^2 = (11s_A^2 + 6s_B^2)/17 = (18.77 + 32.05)/17 = 2.989$ or $1.729^2$ (to 3 s.f.) | M1 A1 | Find pooled estimate of common variance |
| $t = (26 - 177/7)/s\sqrt{1/12 + 1/7} = (26 - 25.29)/s\sqrt{1/12 + 1/7} = 0.7143/0.8222 = 0.869$ | M1 A1 *B1 | Calculate value of $t$ (or $-t$) (to 3 s.f.) |
| $t_{17,\ 0.9} = 1.333$ | | State or use correct tabular $t$ value (to 3 s.f.) |
| $t <$ tabular value, so Petra's belief not supported *or* wing span of $A$ not greater | B1$\checkmark$ | Consistent conclusion (AEF, $\sqrt{}$ on $t$, dep *B1) |
| | [8] | |
10 For a random sample of 6 observations of pairs of values $( x , y )$, where $0 < x < 21$ and $0 < y < 14$, the following results are obtained.

$$\Sigma x ^ { 2 } = 844.20 \quad \Sigma y ^ { 2 } = 481.50 \quad \Sigma x y = 625.59$$

It is also found that the variance of the $x$-values is 36.66 and the variance of the $y$-values is 9.69 .\\
(i) Find the product moment correlation coefficient for the sample.\\
(ii) Find the equations of the regression lines of $y$ on $x$ and $x$ on $y$.\\
(iii) Use the appropriate regression line to estimate the value of $x$ when $y = 6.4$ and comment on the reliability of your estimate.

\hfill \mbox{\textit{CAIE FP2 2016 Q10 [11]}}