Hypothesis testing
Exercise 1
Module : Estimation Techniques for Engineers
Exercise N°1
Statement
A company uses an insulating material to manufacture industrial control
devices. It buys insulating components from a certain supplier who
certifies that the mean thickness of his components is 7.3 millimeters. To
see if the supplier is living up to its commitment, the company measures
the thickness of 10 components taken at random from the delivery. The
results, in millimeters, are :
It is assumed that the thickness in millimeters of one of these
components can be modeled by a normal random variable with unknown
mean µ and standard deviation σ . We want to test if the supplier
respects his commitments.
1
Exercise N°1
1. Give a point estimate of the mean µ.
2. Give a point estimate of the variance σ 2 .
3. Can we consider, at the risk of 5%, that the supplier respects his
commitments? (Detail the different steps of the test that must be
constructed).
4. Assuming now that the standard deviation σ = 0.1 millimeters, at
the risk of 1%, would the decision be different?
2
Exercise N°1
Solution
1. Give a point estimate of the mean µ. A point estimator of the
mean µ is:
n
1X
Xn = Xi
n
i=1
Thus, a point estimate is given by :
n 10
1X 1 X
xn = xi = xi ≃ 7.16
n 10
i=1 i=1
2. Give a point estimate of the variance σ 2 . A point estimator of
the variance is :
n
2 1 X
S = (Xi − Xn )2
n−1
i=1
So a point estimate is given by :
n 10
1 X 1X
s2 = (xi − xn )2 = (xi − 7.16)2 ≃ 0.12
n−1 9 3
i=1 i=1
Exercise N°1
3. Can we consider, at the risk of 5%, that the supplier respects
his commitments? (Detail the different steps of the test that
must be constructed).
We need to construct a two-tailed test on the mean.
Step 1 : Choice of hypothesis
(
H0 : µ = 7.3 mm
H1 : µ ̸= 7.3 mm
Step 2: Decision variable and reduced deviation.
We have n = 10 < 30 , and σ unknown but estimated by ≃ 0.35.
Xn − µ
T = ∼ T(n−1) Student of (n-1) dof.
√s
n
4
Exercise N°1
Step 3: Critical regions.
We determine t α2 ,n−1 from the Student’s law table.
α
α = 0.05 −→ = 0.025 −→ t α2 ,9 = 2.262
2
We calculate the value t0 taken by T :
xn − µ 7.16 − 7.3
t0 = s = 0.35
≃ −1, 26
√ √
n 10
Step 4: Decision
5
Exercise N°1
−t α2 ,n−1 = −2.262 < t0 = −1, 26 < t α2 ,n−1 = 2.262
So, we are in the zone of acceptance of H0 and rejection of H1 , so the
supplier respects its commitments.
4. Assuming now that the standard deviation σ = 0.1 millimeters,
at the risk of 1%, would the decision be different?
• n = 10 < 30 and σ known, then :
σ 0.1
Xn ∼ N (µ = 7.3; √ = √ )
n 10
The reduced deviation is
Xn − 7.3
Z= 0.1
∼ N (0, 1)
√
10
6
Exercise N°1
• We calculate the value z0 taken by Z :
7.16 − 7.3
z0 = 0.1
= −4.43
√
10
α
• α = 0.01 −→ = 0.005 −→ z α2 = 2.58
2
then we have:
z0 = −4.43 < −z α2 = −2.58
Conclusion : We are in the zone of rejection of H0 and acceptance of
H1 , then the decision would be different.