Hypothesis Testing

Hypothesis: A hypothesis is some statement about a population which we want to verify

on the basis of information available from a sample.
Example: The mean monthly income for systems analysts is $6,325.

Hypothesis testing: A procedure based on sample evidence and probability theory to

determine the hypothesis is a reasonable statement and should not be rejected, or is
unreasonable and should be rejected.

Steps in solving testing of hypothesis: The major steps involved in a hypothesis testing
are outlined as follows:

Step 1: State null and alternate hypotheses

Step 2: Select a level of significance

Step 3: Identify the test statistic

Step 4: Formulate a decision rule

Step 5: Arrive at a decision

Step 5: Arrive at a decision Reject null and accept alternate

DStep 4: Formulate a decision rule Do not reject null

oStep 3: Identify the test statistic

Example: Let us suppose that the bulbs manufactured under some standard
Step 2: Select a level of significance
manufacturing process have an average life of  hours and it is proposed to test a new
procedure for manufacturing light bulbs. If  0 is the average life of the bulbs
nStep 1: State null and alternate
hypotheses
manufactured by the new process. Thus we have two populations of bulbs, those
ot reject null
manufactured by standard process and those manufactured by the new process. In this
problem the following three hypothesis may be set up:

1. New process is better than standard process or New process is inferior to

standard process
2. There is difference between two process

-1- TM
 3. There is no difference between the process
Types of Hypotheses:
There are two types of hypothesis.
1. Null Hypothesis ( H 0 )
2. Alternative Hypothesis ( H 1 )

Null Hypothesis ( H 0 ): A statement about the value of a population parameter. That is a

null hypothesis is that statement which we want to reject. Null hypothesis denoted by H 0
For the above example the number 3 hypothesis is called hull hypothesis. That is there is
no difference between the two processes. Null hypothesis can be written as follows
H 0 :  = 0

Alternative Hypothesis ( H 1 ): A statement that is accepted if the sample data provide

evidence that the null hypothesis is false. Alternative hypothesis is usually denoted
by H 1 .
For the above example the 1 or 2 number hypothesis could be alternative hypothesis. So
alternative hypothesis can be written as
H1 :   0
or
H1 :   0
or H1 :    0
One tailed test: The alternate hypothesis H1 , states a direction.

In the above example if the alternative hypothesis is like

H1 :   0
or
H1 :   0
Then this type of test is called one tailed test.

-2- TM
 10- 18

at .05% level of Significance Region of

Do not
rejection
reject
[Probability =.95] [Probability=.05]

0 1.65
Critical value

One-Tailed Test

Two tailed test: No direction is specified in the alternate hypothesis H1 .

In the above example if the alternative hypothesis is like
H1 :    0

Then this type of test is called two tailed test.

10- 20

Region of Region of
Do not
rejection rejection
reject
[Probability=.025] [Probability=.025]
[Probability =.95]

-1.96 0 1.96
Critical value Critical value

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

Example: Suppose a person getting a positive result in a HIV test whether he is not HIV
positive.

-3- TM
Type II Error: Accepting the null hypothesis when it is actually false.
Example: Suppose a person getting negative result in HIV test whether he is HIV
positive.

Level of Significance: The probability of rejecting the null hypothesis when it is

actually true; i.e.the probability of type Ι error is called level of significance. It is denoted
as  .
 = Pr (Type  error)

Power of the test: The probability of Type II Error is defined as  .

That is  = Pr (Type II Error).Now (1-  ) is defined as the power of the test.

The difference between Type I Error and Type II Error given as below:

• Rejecting the null hypothesis when it is actually true is called type I error

On the other hand, accepting the null hypothesis when it is actually false is called
type II error.

• Type I error is known as false positive

On the other hand, Type II error is known as false negative.

• Type I error is more serious then type II error

Type II is less serious then type I error.

• The probability of type I error is called level of significance

On the other hand, if the probability of type II error is subtracted from 1 then that
term is called power of the test.

Test statistic: A value, determined from sample information, used to determine

whether or not to reject the null hypothesis is called test statistic.
Example: There are many types of test statistic. As like, z test, t test, F test,  2 test etc.

Decision Rule

-4- TM
If Computed value < critical value then Null Hypothesis is accepted.i.e. Alternative
hypothesis is rejected

If Computed value > critical value then Null Hypothesis is rejected. i.e. Alternative
hypothesis is rejected

t-test: Testing for a Population Mean where sample size is small i.e less than 30,
and Population Standard Deviation Unknown. The test statistic is the t distribution is

X −
t=
s/ n
The critical value of t is determined by its degrees of freedom equal to n-1.

Assumptions for t-test:

1. When sample size is less than 30, we can use t-test.

2. The sample observation are independent that is sample is random
3. The population standard deviation is unknown.

The Steps are included in t-test given as below:

Step 1
State the null and the alternative hypotheses that is we have to state.
H 0 :  =  0 and H 1 :    0 or
   0 or
  0
Step 2
Select the significance level.
Step 3
Calculate test statistic. Here we have to calculate z test.
Step 4
State the decision rule. That is

Re ject H0 if t  t0 or -t  −t0
Or
Accept H 0 if t  t0 or -t  −t0
Step 5
Make a decision and interpret the results.
Example-1: A drug manufacturer has installed a machine which fills automatically 5
grms in each phile.Randomly 10 sample was taken and found to contain 5.02 grams
on an average in a phial. The Standard deviation of the sample was 0.002 grms.Test at
5% levels of significance, if the adjustment in the machine is in order.

-5- TM
Sol:
Here the null and the alternative hypotheses are
H 0 :  = 5 i.e. the adjustment in the machine is in order
H1 :   5 i.e. the adjustment in the machine is not in order
X −
Now the test statistic is: t =
s/ n
Here; x = 5.02,  = 5, s = 0.002, n = 10
5.02 − 5
= = 31.62
0.002 10
 5%
Here (n-1) =9, = = 0.025
2 2
From the t table, the tabulated value of t is, ttab =  2.262.
As, the calculated value is greater than the tabulated value, that is,
tcal =31.62> ttab =2.262.
So null hypothesis is rejected.
Hence, the adjustment in the machine is not in order.

Example-2: An automobile tyre manufacturer company said that the average life of
certain grade of tyre is 25,000 km.A random sample of 15 tyres were tested and mean
and standard deviation of 27,000 and 5000 kms respectively, were computed. At
5%level of significance is it reasonable to conclude that the average life of certain
grade of tyre is greater than 25,000 km.

Sol: Here the null and the alternative hypotheses are

H 0 :  = 25000 i.e. the average life of certain grade of tyre is 25,000 km.
H1 :   25000 i.e. the average life of certain grade of tyre is greater than 25,000 km.
X −
Now the test statistic is: t =
s/ n
Here; x = 27000,  = 25000, s = 5000, n = 15
27000 − 25000
= = 1.549 = 1.55
5000 15

Here (n-1) =14,  = 5% = .05

From the t table, the tabulated value of t is, ttab =1.761.
As, the calculated value is less than the tabulated value, that is,
tcal =1.55< ttab =1.761.
So null hypothesis is accepted.
Hence, the average life of certain grade of tyre is 25,000 km

Z-test: Test for the population mean from a large sample ( (n  30) . The Z statistics
is defined as

-6- TM
 ( X − )
z=

n

Here,
X is sample mean.
 is population mean.
 is population standard deviation.
s is sample standrad deviation.
n sample size.

Assumptions for z-test:

1. When sample size is greater than 30,we can use z-test

2. The population standard deviation is known

Problem: The mean life time of a sample of 100 light tubes produced by a company
is found to be 1580 hours and the population standard deviation is 90 hours. Test the
hypothesis at the 5% level of significance that the mean of the tubes produced by the
company is 1550 hours.

Solution:

Here the null and the alternative hypotheses are

H 0 :  = 1550
H1 :   1550

Here given that 5 % level of significance. The tabulated value of z 0 is  1.96.

Here the test statistic is z.

So,
X −  1580 − 1550
z= = = 3.33
 n 90 100

We know,
Re ject H 0 if z  z 0

Otherwise accept H 0

-7- TM
Here z = 3.33 and tabulated value z 0 is  1.96.
As 3.33>1.96 that is z  z 0 ,

So null hypothesis is rejected. Hence the mean life time of the tubes produced by the
company may not be 1550 hours.

 2 -test: For testing the population variance  2 ,based on sample variance s 2 ,  2

test is used. The text statistic is given as below:
 n2−1 =
(n − 1)s 2
2
Here s 2 is sample variance,  2 is population variance and degrees of freedom is
(n-1).

Assumptions:
1. The sample should contain at least contain 50 observations.
2. The sample observation should be independent of each other
3. It is use for variance test.

-8- TM