N2025 P2 Q10

N2025 P2 Q10

17 marks
Free

In a food processing plant, two different machines are used to produce packets of flour. Over a long period of time it has been established that the masses, in \(\text{kg}\), of packets produced by Machine \(P\) follow the distribution \(\mathrm{N}(2.2, 0.1^2)\) and the masses, in \(\text{kg}\), of packets produced by Machine \(Q\) follow the distribution \(\mathrm{N}(2.1, 0.05^2)\). It is assumed that these two distributions are independent.

  1. Explain why, in the context of the question, it is reasonable to assume that the two distributions are independent.[1]
  2. Find the probability that a randomly chosen packet produced by Machine \(P\) has a greater mass than a randomly chosen packet produced by Machine \(Q\).[3]
  3. Find the probability that \(3\) randomly chosen packets produced by Machine \(P\) and \(5\) randomly chosen packets produced by Machine \(Q\) have a total mass greater than \(17\text{ kg}\).[3]

Following an adjustment to Machine \(P\), the production manager wishes to test if the mean mass of packets produced by that machine now differs from \(2.2\text{ kg}\).

  1. State hypotheses for the production manager's test, defining any parameter that you use.[2]

The production manager finds the masses of a random sample of \(30\) packets produced by Machine \(P\). The masses, \(x\text{ kg}\), are summarised below.

\(\sum (x - 2) = 4.5 \qquad \sum (x - 2)^2 = 1.11\)

  1. Calculate unbiased estimates of the population mean and variance of the masses of packets after Machine \(P\) has been adjusted.[2]
  2. Carry out the production manager's test at the \(5\%\) level of significance. Show the values you use to carry out the test and give your conclusion in the context of the question.[4]
  3. Explain why the production manager would have less confidence in the conclusion of the hypothesis test if a sample of fewer than \(30\) packets had been used.[1]
  4. Explain why the production manager would have less confidence in the conclusion of the hypothesis test if the sample had not been chosen randomly.[1]

Default solution

  1. The machines operate separately, so one machine’s packet mass does not affect the other’s.
  2. Let \(P\) and \(Q\) be independently selected packet masses in kilograms.
    \(D=P-Q\)\(E(D)=2.2-2.1=0.1\)\(\operatorname{Var}(D)=0.1^2+0.05^2=0.0125\)\(D\sim\mathrm N(0.1,0.0125)\)
    \(P(P>Q)=P(D>0)=0.8144533152\ldots\approx0.814\)
  3. Let \(T\) be the total mass of three \(P\) packets and five \(Q\) packets. For independent packet masses,
    \(E(T)=3(2.2)+5(2.1)=17.1\)\(\operatorname{Var}(T)=3(0.1^2)+5(0.05^2)=0.0425\)\(T\sim\mathrm N(17.1,0.0425)\)
    \(P(T>17)=0.6861870975\ldots\approx0.686\)
  4. Let \(\mu\) be the population mean packet mass in kilograms from Machine \(P\) after adjustment.
    \(H_0:\mu=2.2,\qquad H_1:\mu\ne2.2\)
  5. \(\bar{x}\)\({}=2+\frac{4.5}{30}=2.15\)\(s^2\)\({}=\frac1{29}\left[1.11-\frac{4.5^2}{30}\right]=0.015\)
    The estimates are 2.15 kg for the mean and 0.015 kg² for the variance.
  6. Using (d) and (e), the independent random sample of 30 permits the approximate normal mean test by the central limit theorem.
    \(z\)\({}=\frac{2.15-2.2}{\sqrt{\frac{0.015}{30}}}\)\(z\)\({}=-\sqrt5=-2.236067977\ldots\)\(p\)\({}=2P(Z\le-\sqrt5)\)\(p\)\({}=0.0253473187\ldots\)
    Since \(p<0.05\), reject \(H_0\). There is sufficient evidence at the 5% level that Machine \(P\)’s population mean packet mass has changed after adjustment.
  7. The post-adjustment distribution is unknown. With fewer than 30 packets, the central limit approximation is less secure.
  8. A non-random sample may be systematically unrepresentative, making the inference about the population mean unreliable.
Answer:(a) Separate machines support independence. (b) 0.814. (c) 0.686. (d) \(H_0:\mu=2.2,H_1:\mu\ne2.2\), post-adjustment mean \(\mu\) in kg. (e) Mean 2.15 kg, variance 0.015 kg². (f) \(z=-\sqrt5,p\approx0.0253473\); reject at 5%, evidence the mean has changed. (g) Weaker large-sample normal approximation. (h) Possible selection bias.

Need help? Join our JC Math tuition classes.

Learn more