2019 H2 Mathematics Paper 2 Question 9

Hypothesis Testing

Answers

The manager should carry out a 2-tail test as, in testing whether the mean resistance is in fact 750, the actual population mean resistance could be greater or less than 750.

H0:μ=750{\textrm{H}_0: \mu = 750}
H1:μ750{\textrm{H}_1: \mu \neq 750}

p-value=0.0897{p\textrm{-value} = 0.0897}
There is insufficient evidence at the 5%{5\%} level of significance to conclude whether the mean resistance of the resistors is 750{750} ohms
We will need to take a larger sample of size at least 30.{30.}
Since the population distribution of the 1250{1250} ohm resistors is unknown, in order for the sample mean test statistic X{\overline{X}} to be normally distributed (approximately) to carry out at Z-{Z\textrm{-}}test, we will need a large sample size of at least 30{30} so that the Central Limit Theorem can be applied.

Full solutions

(i)

The manager should carry out a 2-tail test as, in testing whether the mean resistance is in fact 750, the actual population mean resistance could be greater or less than 750. {\blacksquare}
Let μ{\mu} denote the population mean resistance of the resistors, X{X} denote the random variable of the resistance of the resistors, x{\overline{x}} denote the mean resistance of the sample of 8,{8, } and H0{\mathrm{H}_0} and H1{\mathrm{H}_1} be the null and alternative hypothesis respectively.
H0:μ=750{\textrm{H}_0: \mu = 750}
H1:μ750  {\textrm{H}_1: \mu \neq 750 \; \blacksquare}

(ii)

x=xn=756\begin{align*} \overline{x} &= \frac{\sum x}{n} \\ &= 756 \end{align*}
Under H0,{\textrm{H}_0, } test statistic
Z=XμσnN(0,1)Z= \frac{\overline{X} - \mu}{\frac{\sigma}{\sqrt{n}}} \sim N(0,1)
p-value=0.089686{p\textrm{-value} = 0.089686}
Since p-value>0.05,{p\textrm{-value} > 0.05, } we do not reject H0{\textrm{H}_0}
There is insufficient evidence at the 5%{5\%} level of significance to conclude whether the mean resistance of the resistors is 750{750} ohms {\blacksquare}

(iii)

We will need to take a larger sample of size at least 30.{30.}
Since the population distribution of the 1250{1250} ohm resistors is unknown, in order for the sample mean test statistic X{\overline{X}} to be normally distributed (approximately) to carry out at Z-{Z\textrm{-}}test, we will need a large sample size of at least 30{30} so that the Central Limit Theorem can be applied. {\blacksquare}