INTERNATIONAL UNIVERSITY (IU) Engineering Probability & Statistic

ISE Department Lecturer: Phan Nguyễn Kỳ Phúc

--------------------o0o------------------

HYPOTHESIS TESTING
Hypothesis testing is conducted when:

 The statement or the claim contradicts with the observation. For example, if a company claims that, the average weight of their
products is at least 50. However, based on the sample, the average weight of the sample is only 49. Consequently, the
hypothesis testing must be conducted to justify whether we are able to reject the company’s claim.
 It is often that, the claim is put in H0. Based on H0 , the alternative H1 is identified
 However, in case it is difficult to identify H0, H1 can be identified first through comparing the real observations with the value
in the claim. For the above example, the mean of the sample is 49<50. So H1 :   50  H 0 :   50

Meaning of reject of null hypothesis

 Under the null hypothesis, if the probability that we obtain the sample is too small, i.e., smaller than significant level  , we
think that it cannot occur, so we reject the null hypothesis.
 For example, if someone tell you that a coin is fair, i.e., Ho: P(Head) =0.5. You tossed a coins 10 times and you got 9 Heads,
and 1 Tails. You use the null hypothesis that H0: P(Head)= 0.5 to calculate the probability that in 10 tosses , we obtains less
than 1 tail is only 0.0107. Since it is extremely rare to occur, but it already occurred => the null hypothesis maybe wrong, so
we reject the null hypothesis

Test for Population Mean

In Test for Population Mean, there are 2 type of test

 H :   0
 2-tailed test:  0
 H1 :    0
 H :   0 (  0 )
 1-tailed test:  0
 H1 :    0 (    0 )

07 One Population Hypothesis Testing 1

 INTERNATIONAL UNIVERSITY (IU) Engineering Probability & Statistic
ISE Department Lecturer: Phan Nguyễn Kỳ Phúc
--------------------o0o------------------

When sampling a sample, 3 following parameters can be obtained from a sample

 Sample size: n
 Mean of sample: x
 Standard deviation of sample: s

2-TAILED TEST

std (σ) of population is given std (σ) of population is not given

Sample size: n Small sample size (n<30) Large sample size (n>30)

 H 0 :   0  H 0 :   0  H 0 :   0
Hypothesis   
 H1 :    0  H1 :    0  H1 :    0
 2   s2   s2 
Distribution of x z-distribution:  x,  t-distribution:  x,  , dof  n  1 z-distribution:  x, 
 n   n  n
x  0 x  0 x  0
Test statistic zt  tt  zt 
 n s n s n

Critical Values  z 2 t 2,n 1  z 2

Reject Ho if : Reject Ho if : Reject Ho if :
Decision
zt  z 2 or zt   z 2 tt  t 2, n 1 or tt  t 2, n 1 zt  z 2 or zt   z 2
Confident interval  s s
x  z x  t x  z
of population mean 2 n 2
,n 1 n 2 n

07 One Population Hypothesis Testing 2

 INTERNATIONAL UNIVERSITY (IU) Engineering Probability & Statistic
ISE Department Lecturer: Phan Nguyễn Kỳ Phúc
--------------------o0o------------------

1-TAILED TEST

std (σ) of population is given std (σ) of population is not given

Sample size: n Small sample size (n<30) Large sample size (n>30)

 H 0 :   0 (  0 )  H 0 :   0 (  0 )  H 0 :   0 (  0 )
Hypothesis   
 H1 :    0 (    0 )  H1 :    0 (    0 )  H1 :    0 (    0 )
 2   s2   s2 
Distribution of x z-distribution:  x,  t-distribution:  x,  , dof  n  1 z-distribution:  x, 
 n   n  n
x  0 x  0 x  0
Test statistic zt  tt  zt 
 n s n s n

Critical Values  z ( z ) t , n 1 (t , n 1 )  z ( z )

Reject Ho if : Reject Ho if : Reject Ho if :
Decision
zt   z ( zt  z ) tt  tn 1,  (tt  tn 1,  ) zt   z ( zt  z )
Confident interval  s s
x  z x  t x  z
of population mean 2 n 2
,n 1 n 2 n

07 One Population Hypothesis Testing 3

 INTERNATIONAL UNIVERSITY (IU) Engineering Probability & Statistic
ISE Department Lecturer: Phan Nguyễn Kỳ Phúc
--------------------o0o------------------

PROPROTION TEST

2-tailed Test 1-tailed Test

Sample size: n n

 H 0 : p  p0  H 0 : p  p0 ( p  p0 )
Hypothesis  
 H 1 : p  p0  H1 : p  p0 ( p  p0 )
 pq   pq 
Distribution of p z-distribution:  p,  z-distribution:  p, 
 n   n 
 
p  p p  p
0 0
zt  zt 
Test statistic p0 q0 p0 q0
n n

Critical Values  z 2  z ( z )
Reject Ho if : Reject Ho if :
Decision
zt  z 2 or zt   z 2 zt   z ( zt  z )

Confident interval pq pq

p  z p  z
of population mean 
n

n
2 2

07 One Population Hypothesis Testing 4

 INTERNATIONAL UNIVERSITY (IU) Engineering Probability & Statistic
ISE Department Lecturer: Phan Nguyễn Kỳ Phúc
--------------------o0o------------------

VARIANCE TEST

2-tailed Test 1-tailed Test

Sample size: n n
2 2 2 2 2 2
 H 0 :    0  H 0 :    0 (   0 )
Hypothesis  2 2  2 2 2 2
 H1 :    0  H1 :    0 (   0 )
(n  1)s 2
Distribution of χ2 -distribution: dof  n  1 χ2 -distribution: dof  n  1
2
2 (n  1) s 2 2(n  1) s 2
Test statistic  
t  
t
 02  02

Critical Values 12 /2,n 1 or 2 / 2,n 1 12 ,n 1 (2 ,n 1 )

Reject Ho if : Reject Ho if :
Decision
t2  12 /2,n 1 or  t2  2 / 2,n 1 t2  12 ,n 1 ( t2  2 ,n 1 )
 2 2
  2 2

Confident interval of  (n  1) s (n  1)s   (n  1) s (n  1)s 
 2 , 2   2 , 2 
population mean    
 
 2 ,n 1 1 , n 1 
2   2 ,n 1 1 , n 1 
2 

07 One Population Hypothesis Testing 5

 INTERNATIONAL UNIVERSITY (IU) Engineering Probability & Statistic
ISE Department Lecturer: Phan Nguyễn Kỳ Phúc
--------------------o0o------------------

Assignments

Question 1: A producer specifies that the mean lifetime of a certain type of battery is at least 240
hours. A sample of 18 such batteries yielded the following data.

237 242 232 242 248 230

244 243 254 262 234 220

225 236 232 218 228 240

Assuming that the life of the batteries is approximately normally distributed, do the data indicate
that the specification is not being met?

Question 2: The mean breaking strength of a certain type of fiber is required to be at least 200
psi. Past experience indicates that the standard deviation of breaking strength is 5 psi. If a sample
of 8 pieces of fiber yielded breakage at the following pressures,

210 198 195 202

197.4 196 199 195.5

Would you conclude, at the 5 percent level of significance, that the fiber is unacceptable? What
about at the 10 percent level of significance?

Question 3: In a certain chemical process, it is very important that a particular solution that is to
be used as a reactant have a pH of exactly 8.20. A method for determining pH that is available
for solutions of this type is known to give measurements that are normally distributed with a
mean equal to the actual pH and with a standard deviation of .02. Suppose 10 independent
measurements yielded the following pH values:

8.18, 8.17, 8.16, 8.15, 8.17,

8.21, 8.22, 8.16, 8.19, 8.18

1. What conclusion can be drawn at the α = .10 level of significance?

2. What about at the α= .05 level of significance?

Question 4: An advertisement for a new toothpaste claims that it reduces cavities of children in
their cavity-prone years. Cavities per year for this age group are normal with mean 3 and
standard deviation 1. A study of 2,500 children who used this toothpaste found an average of
2.95 cavities per child. Assume that the standard deviation of the number of cavities of a child
using this new toothpaste remains equal to 1.

07 Regression Analysis 6
 INTERNATIONAL UNIVERSITY (IU) Engineering Probability & Statistic
ISE Department Lecturer: Phan Nguyễn Kỳ Phúc
--------------------o0o------------------

1. Are these data strong enough, at the 5 percent level of significance, to establish the claim of
the toothpaste advertisement?

2. Do the data convince you to switch to this new toothpaste?

Question 5: The weights of salmon grown at a commercial hatchery are normally distributed
with a standard deviation of 1.2 pounds. The hatchery claims that the mean weight of this year’s
crop is at least 7.6 pounds. Suppose a random sample of 16 fish yielded an average weight of 7.2
pounds. Is this strong enough evidence to reject the hatchery’s claims at the

1. 5 percent level of significance;

2. 1 percent level of significance?

3. What is the p-value?

Question 6: A medical scientist believes that the average basal temperature of (outwardly)
healthy individuals has increased over time and is now greater than 98.6 degrees Fahrenheit (37
degrees Celsius). To prove this, she has randomly selected 100 healthy individuals. If their mean
temperature is 98.74 with a sample standard deviation of 1.1 degrees, does this prove her claim
at the 5 percent level? What about at the 1 percent level?

07 Regression Analysis 7