Download as pdf or txt
Download as pdf or txt
You are on page 1of 5

Test of Hypothesis

A hypothesis is a statement about a population. Statistical hypothesis, a collection of data, then


uses the data to test the assertion. A statement about a population parameter developed for the
purpose of testing. The method of statistical inference will be used to test the validity of a claim
about the value of a population parameter. The claim is generally called a hypothesis and the
methods is called test of hypothesis.

Hypothesis testing: A procedure is based on sample evidence and probability theory to determine
whether the hypothesis is a reasonable statement.

The null hypothesis is developed for the purpose of testing, either reject or fail to reject

The null hypothesis is a statement that is not rejected unless the sample data provide convincing
evidence that is false. If the given claim contains equality, or a statement of no change from the
given or accepted condition, then it is null hypothesis.

Alternative hypothesis describes what will conclude if rejected the null hypothesis. A statement
that is accepted if the sample data provide sufficient evidence that the null hypothesis is false.

The Null Hypothesis, H0


 States the assumption (numerical) to be tested
 Begin with the assumption that the null hypothesis is true
 Always contains “=” , “≤” or “” sign
 May or may not be rejected

The Alternative Hypothesis, H1

 Is the opposite of the null hypothesis


e.g., The average number of TV sets in U.S. homes is not equal to 3 (H1: μ ≠ 3 )
 Never contains the “=” , “≤” or “” sign
 May or may not be supported

Type I Error: Rejecting the null hypothesis where it is true.

The probability of Type I error is called an  risk (is called significance level).

  P(Rejecting H0 | H0is true)  P(TypeI error).

Type II Error: Accepting the null hypothesis where it is false.

The probability of Type II error is called a β risk (1- β is called power).


  P( Not Rejecting H 0 | H 0is true)  P(TypeII error).
1
Type I & II Error Relationship
 Type I and Type II errors can not happen at the same time
 Type I error can only occur if H0 is true
 Type II error can only occur if H0 is false
The null hypothesis is   0 ( 0 is some standard value)
The alternative hypothesis   0
One sided alternative hypothesis are   0 and   0 .
Using the population mean, there are three types of hypothesis tests:
(i) Two tailed (sided) test
H 0 :   0
H1 :   0 (two sided)
(ii) One tailed (sided) lower tail/ left tail test
H 0 :   0
H1 :   0 (left tailed)
(iii) One tailed (sided) upper tail/right tail test
H 0 :   0
H1 :   0 (right tailed)

For a two tailed test, there are two rejection regions, one on each side

Do not reject H 0 if L  X  R

Reject H 0 if X  L or X  R,

 
where, L  0  z and R  0  z
2 n 2 n

Equivalently,

Do not reject H 0 if  z / 2  z  z / 2

X  0
Reject H 0 if z   z / 2 or z  z / 2 , where z  .
/ n

Conclusion

Do not reject H 0 if | z | z / 2 ,

Reject H 0 if | z | z / 2 .

Problem The monthly household income of all the households in a specific community
was 2400 with a population standard deviation 480. An economist is interested in
2
knowing if this mean monthly household income has changed. He selected a random
sample of 144 households and the sample mean was found to be 2480. Conduct the
hypothesis test at   0.05.

Ans. The null and alternative hypothesis is

H 0 :   2400
H1 :   2400

X  0
z is approximately a standard normal distribution
/ n

We have   0.05, z / 2  1.96

Do not reject H 0 if | z |  1.96,

Reject H 0 if | z |  1.96.

X  0 2480  2400
But z    2.00
/ n 480 / 144

The conclusion is to reject H 0 because | z |  1.96.

 480
where, L  0  z  2400  1.96  2321.6 and
2 n 144

 480
R  0  z   2400  1.96  2478.4.
2 n 144

The decision rule is

Do not reject H 0 if 2321.6  X  2478.6

Reject H 0 if X  2321.6 or X  2478.4.

The conclusion is to reject H 0 because X  2478.4.

For a one tailed or lower tail test, the rejection region is in the lower tail side

Do not reject H 0 if X  L,

Reject H 0 if X  L ,

3

where, L  0  z
n

Equivalently,

Do not reject H 0 if z   z ,

X  0
Reject H 0 if z   z , where z  .
/ n

Problem A machine filling soft drink cans, when appropriately adjusted, fills cans with a
mean of at least 12 ounces and a standard deviation of 0.5 ounces. The contents of the
cans are normally distributed. A random sample of 25 cans is selected and the sample
mean is found to be 11.85 ounces. At a significance level of 0.05, test if there is
sufficiently evidence to believe that the population mean is less than 12 ounces.

Ans. The null and alternative hypothesis is

H 0 :   12
H1 :   12

X  0
z is approximately a standard normal distribution
/ n

We have   0.05, z  1.645.

Do not reject H 0 if z  1.645,

Reject H 0 if z  1.645.

X  0 11.85  12
But z    1.5.
/ n 0.5 / 25

The conclusion is not to reject H 0 because z  1.645.

 0.5
where, L  0  z  12  1.645  11.8355,
n 25

The decision rule is

Do not reject H 0 if X  11.8355,

Reject H 0 if X  11.8355.

4
The conclusion is not to reject H 0 because X  11.8355.

For a one tailed or upper tail test, the rejection region is in the upper tail side

Do not reject H 0 if X  R,

Reject H 0 if X  R,


where, R  0  z
n

Equivalently,

Do not reject H 0 if z  z ,

X  0
Reject H 0 if z  z , where z  .
/ n

You might also like