WORKSHOP 5 SOLUTIONS

Q5.1

The lifetimes of the heating element in a Heatfast electric oven are normally distributed,
with a mean of 7.8 years and a standard deviation of 2.0 years.

(a) If the element is guaranteed for two years,

(i) What percentage of the ovens sold will need replacement in the guarantee period
because of element failure? Include a statement describing your answer.
Let X = lifetime of heating element (years) 𝑿~𝑵(𝟕. 𝟖, 𝟐. 𝟎𝟐 )
𝑿 − 𝝁 𝟐 − 𝟕. 𝟖
𝑷(𝑿 < 𝟐) = 𝑷 ( < ) = 𝑷(𝒁 < −𝟐. 𝟗𝟎) = 𝟎. 𝟎𝟎𝟏𝟗 = 𝟎. 𝟏𝟗%
𝝈 𝟐. 𝟎
0.19% of the ovens sold will need replacement in the guarantee period.

(ii) In a year in which 10,000 ovens are sold, how many ovens would you expect to have to
replace in the guarantee period because of element failure?
𝑷(𝑿 < 𝟐) × 𝟏𝟎𝟎𝟎𝟎 = 𝟎. 𝟎𝟎𝟏𝟗 × 𝟏𝟎𝟎𝟎𝟎 = 𝟏𝟗 𝐨𝐯𝐞𝐧𝐬

(b) What proportion of elements are expected to last for between 2 and 10 years? Write a
statement explaining your answer.
𝟐 − 𝟕. 𝟖 𝑿 − 𝝁 𝟏𝟎 − 𝟕. 𝟖
𝑷(𝟐 < 𝑿 < 𝟏𝟎) = 𝑷 ( < < ) = 𝑷(−𝟐. 𝟗𝟎 < 𝒁 < 𝟏. 𝟏𝟎)
𝟐. 𝟎 𝝈 𝟐. 𝟎
= 𝑷(𝒁 < 𝟏. 𝟏𝟎) − 𝑷(𝒁 < −𝟐. 𝟗𝟎) = 𝟎. 𝟖𝟔𝟒𝟑 − 𝟎. 𝟎𝟎𝟏𝟗 = 𝟎. 𝟖𝟔𝟐𝟒

0.8624 is the proportion of elements that are expected to last between 2 to 10 years.
(c) Heatfast is reconsidering the length of the guarantee period on heating elements.
Calculate the length of guarantee period such that Heatfast would expect to replace a
maximum of 1% of ovens due to element failure. Include a statement describing your
answer.
P(X < x*) = 0.01
Finding this unknown X value will give the lifetime value for which 1% of ovens are less
than (i.e. a maximum of 1%). As the statistical tables are for the Z distribution:
P(Z < z*) = 0.01

The critical value of Z which cuts off a 1% tail (from the bottom line of Table 2) is 2.327.
So to cut-off 1% in the left hand tail, the critical value will be -2.327. You can also use Table 1a.
To find the unknown value of X, the following formula is used:
𝑿 = 𝝁 + 𝝈𝒁
Here, x*= 𝟕. 𝟖 + (𝟐. 𝟎 × −𝟐. 𝟑𝟐𝟕) = 3.146
Since a guarantee period is always a round number, the appropriate choice is 3 years.
The length of guarantee period such that Heatfast would expect to replace a maximum of 1%
of ovens due to element failure is 3 years.

(d) Find the length of time such that it includes 95% of all ovens. Include a statement
describing your answer.
P(X < x*) = 0.95 OR P(X > x*) = 0.05
As the statistical tables are for the Z distribution:
P(Z < z*) = 0.95 OR P(Z > z*) = 0.05
Critical value of Z cutting off a 5% right tail (from the bottom line of Table 2) is 1.645.
You can also use table 1b. z*= 1.645.
To find the unknown value of X, the following formula is used:
𝑿 = 𝝁 + 𝝈𝒁
Here, x*= 𝟕. 𝟖 + (𝟐. 𝟎 × 𝟏. 𝟔𝟒𝟓) = 11.09
The length of time which includes 95% of all ovens is 11.1 years.
Video solution for Q5.1

Q5.2

Based on laboratory testing, the lifetime of a Tyrannosaurus brand tyre is taken to be

normally distributed, with mean 65,107 kilometres (km) and standard deviation of
2,582 km. The tyres carry a customer warranty for 60,000 km.

(a) What proportion of the tyres is expected to fail before the warranty expires?
Let the random variable X = lifetime of a tyre (km) 𝑿~𝑵(𝟔𝟓𝟏𝟎𝟕, 𝟐𝟓𝟖𝟐𝟐 )
𝑿 − 𝝁 𝟔𝟎𝟎𝟎𝟎 − 𝟔𝟓𝟏𝟎𝟕
𝑷(𝑿 < 𝟔𝟎𝟎𝟎𝟎) = 𝑷 ( < ) = 𝑷(𝒁 < −𝟏. 𝟗𝟖) = 𝟎. 𝟎𝟐𝟑𝟗
𝝈 𝟐𝟓𝟖𝟐

Thus 0.0239 is the proportion of tyres that fail before the warranty expires.

(b) Obtain the 50th percentile of tyre life. Explain how you obtained this answer.
th
50 percentile = median = mean since the normal distribution is symmetric.
th
Therefore 50 percentile = 65,107 km.
(c) Explain to a non-statistician what your answer from (b) means
This means that 50% of the tyres last less than 65,107km and 50% last beyond this
number of kilometres

(d) Calculate the probability that a tyre would last for 145,000 km or more
145,000km is (145000 – 65107)/2582 = 30.94 standard deviations above the mean. The
tables tell us that to 4 decimal places the probability of being more than 4 standard
deviations above the mean is 0, and so the probability of being 31 standard deviations
above the mean is vanishingly small
Video solution for Q5.2

Q5.3

A statistical analyst who works for a large insurance company is in the process of examining
several pension plans. Company records show that the age at which its male clients retire is
approximately normally distributed with a mean of 63.7 years and a standard deviation of 3.1
years.

(a) Calculate the probability that a randomly selected male client will retire before the age
of 65 years.
Let X = the retirement age of male client (years) 𝑿~𝑵(𝟔𝟑. 𝟕, 𝟑. 𝟏𝟐 )
𝑿 − 𝝁 𝟔𝟓 − 𝟔𝟑. 𝟕
𝑷(𝑿 < 𝟔𝟓) = 𝑷 ( < ) = 𝑷(𝒁 < 𝟎. 𝟒𝟐) = 𝟎. 𝟔𝟔𝟐𝟖
𝝈 𝟑. 𝟏
(b) If a random sample of 50 male clients is to be selected from the company database, what
is the probability that the sample mean will be less than 65 years?
Let the random variable 𝑿 = mean retirement age of samples of 50 male clients (years)
𝟑. 𝟏𝟐
𝑿~𝑵 (𝟔𝟑. 𝟕, )
𝟓𝟎

𝑿−𝝁 𝟔𝟓 − 𝟔𝟑. 𝟕
𝑷(𝑿 < 𝟔𝟓) = 𝑷 ( < ) = 𝑷(𝒁 < 𝟐. 𝟗7) = 𝟎. 𝟗𝟗𝟖𝟓
𝝈⁄ √ 𝒏 𝟑. 𝟏⁄√𝟓𝟎

(c) Close examination of the ages of recent retirees shows that the assumption of a normal
distribution may be false.
Which, if either, of your answers above would be changed by this information, and why?
The assumption of a normal distribution was important in part (a) but not in part (b),
because the central limit theorem tells us that for a large enough sample, (at least 30)
𝑿 will be approximately normally distributed.

Video solution for Q5.3

 Further practice (independent practice for students):

Q5.4

When using tables to obtain standard normal probabilities, values of Z can only be specified
to two decimal places. Draw sketches for all answers and shade the relevant areas. Use
Table 1a or 1b in the statistical tables provided on the website to obtain the following
probabilities.

(i) P(Z < -0.443)

= P(Z < -0.44) = 0.3300

(ii) P(Z < 1.522)

= P(Z < 1.52) = 0.9357

(iii) P(Z > -1.944)

= P(Z > -1.94) = 1 - P(Z < -1.94) = 1 - 0.0268 = 0.9738
Alternative solution: P(Z > -1.94) = P(Z < 1.94) = 0.9738
(iv) P(0 < Z < 1.282)
= P(0 < Z < 1.28) = P(Z < 1.28) – P(Z < 0) = 0.8997 – 0.5 = 0.3997

Q5.5
Use Table 1a or 1b in the statistical tables provided on the website to find the value for
which
(i) P(Z < z*) = 0.895
Look in the body of the table for the value closest to 0.895. 0.8944 is closest and
corresponds to z* = 1.25

(ii) P(Z < z*) = 0.054

The closest value is 0.0537 and corresponds to z* = -1.61
(iii) P(Z > z*) = 0.025
So P(Z < -z*) = 0.025. The closest value is 0.0250, corresponding to –z*= -1.96
So z*= 1.96

(iv) P(0 < Z < z*) = 0.400

So P(Z < z*) = 0.400 + 0.5 = 0.9
The closest value is 0.8997, corresponding to z* = 1.28

Q5.6

An automatic machine in a manufacturing process is operating properly if the lengths of an

important subcomponent are normally distributed, with mean μ = 117 cm and standard
deviation σ = 2.1 cm. If the machine is operating correctly:

(a) Find the probability that one randomly selected unit has a length greater than 120 cm
Let the random variable X = length of subcomponent (cm)
If the machine is operating correctly then 𝑿~𝑵(𝟏𝟏𝟕, 𝟐. 𝟏𝟐 )

𝑿 − 𝝁 𝟏𝟐𝟎 − 𝟏𝟏𝟕
𝑷(𝑿 > 𝟏𝟐𝟎) = 𝑷 ( > ) = 𝑷(𝒁 > 𝟏. 𝟒𝟐𝟖𝟔) ≈ 𝑷(𝒁 > 𝟏. 𝟒𝟑)
𝝈 𝟐. 𝟏
= 𝟏 − 𝑷(𝒁 < 𝟏. 𝟒𝟑) = 𝟏 − 𝟎. 𝟗𝟐𝟑𝟔 = 𝟎. 𝟎𝟕𝟔𝟒

Alternative solution: 𝑷(𝒁 > 𝟏. 𝟒𝟑) = 𝑷(𝒁 < −𝟏. 𝟒𝟑) = 𝟎. 𝟎𝟕𝟔𝟒
(b) Find the probability that if three units are randomly selected, their mean length
exceeds 120 cm
Let the random variable 𝑿 = mean length of samples of 3 subcomponents (cm)
𝟐. 𝟏𝟐
𝑿~𝑵 (𝟏𝟏𝟕, )
𝟑

𝑿−𝝁 𝟏𝟐𝟎 − 𝟏𝟏𝟕

𝑷(𝑿 > 𝟏𝟐𝟎) = 𝑷 ( > ) = 𝑷(𝒁 > 𝟐. 𝟒𝟕𝟒𝟒) ≈ 𝑷(𝒁 > 𝟐. 𝟒𝟕)
𝝈⁄ √ 𝒏 𝟐. 𝟏⁄√𝟑
= 𝟏 − 𝑷(𝒁 < 𝟐. 𝟒𝟕) = 𝟏 − 𝟎. 𝟗𝟗𝟑𝟐 = 𝟎. 𝟎𝟎𝟔𝟖

Alternative solution: 𝑷(𝒁 > 𝟐. 𝟒𝟕) = 𝑷(𝒁 < −𝟐. 𝟒𝟕) = 𝟎. 𝟎𝟎𝟔𝟖

(c) Explain the differences between parts (a) and (b)

In (i) we are considering the probability that a single item is longer than 120cm. The answer
is about 0.076.
In (ii) we are considering three separate items, but the question is whether their mean
length is greater than 120cm. This will certainly happen if all three are longer than 120cm,
but it can also happen when one or two are shorter than 120, so long as the average is
above. However, the distribution of the mean is less variable than the distribution of
individual values, so the probability of the mean being above 120cm will be lower than the
probability for an individual item.