WRU CoBE Statistics for Management II Handout Chapter 3 WRU CoBE Statistics for Management II Handout Chapter 3

CHAPTER THREE Or symbolically,

HYPOTHESIS TESTING H o :   330ml
H A :   330ml
3.1 Introduction: Basic Concepts If the inspector takes a random sample of bottles of this product and finds that the mean content per
bottle is much less than 330 ml, then he may conclude that the complaint of the customers is
In the preceding chapter, estimation, we used information obtained in a simple random sample to construct a
correct.
confidence interval estimate of the unknown value of a population parameter. In this chapter we shall start with an
assumed value of a parameter then we shall use sample evidence to decide whether the assumed value is Hypothesis testing is a procedure for checking the validity of statistical hypothesis. It is the process by
unreasonable and should be rejected, or whether it should be accepted. which we decide whether the null hypothesis should be rejected or not. The value computed from a sample
. that is used to determine whether the null hypothesis has to be rejected or not is called a test statistic.
Very often in practice we are called up n to make decisions about populations on the basis of sample Sometimes some known population quantities are used in the calculation of a test statistic alongside sample
values.
information. Such decisions are called statistical decisions.
Hypothesis is the assumptions we make about the value of population parameter. So hypothesis testing is a 3.2. Steps in Hypothesis Testing
procedure based on sample evidence and probability distribution used to determine whether the hypothesis
is a reasonable statement and should not be rejected, or is unreasonable. Statisticians follow a formal process to determine whether to reject a null hypothesis, based on sample data.
This process, called hypothesis testing, consists of four steps.
 A statistical hypothesis: A claim (or statement) about a population parameter is called a hypothesis. It is an
assumption about a population parameter. This assumption may or may not be true. 1. State the hypotheses. This involves stating the null and alternative hypotheses. The hypotheses are
Examples;
stated in such a way that they are mutually exclusive. That is, if one is true, the other must be false.
a) The mean daily profit of a supermarket is 1000 Birr or   1000 Birr.
2. Formulate an analysis plan. The analysis plan describes how to use sample data to evaluate the
b) The mean time to complete a certain assembly job is less than 2 hours or   2 hours. null hypothesis. The evaluation often focuses around a single test statistic.
c) The proportion of customers in this area who prefer this product is more than 75 percent or 3. Analyze sample data. Find the value of the test statistic (mean score, proportion, t-score, z-score,
P  75% . etc.) described in the analysis plan.
 Hypothesis testing: refers to the formal procedures used by statisticians to accept or reject statistical hypotheses. 4. Interpret results. Apply the decision rule described in the analysis plan. If the value of the test
 Statistical Hypotheses: The best way to determine whether a statistical hypothesis is true would be to statistic is unlikely, based on the null hypothesis, reject the null hypothesis.
examine the entire population.
3.3 Type I and Type II errors (concepts)
There are two types of statistical hypotheses.
Applying a hypothesis test may lead to a wrong conclusion. There are two kinds of possible errors, called
 Null hypothesis. A null hypothesis is a claim (or statement) about a population parameter that is type I error and type II error.
assumed to be true until it is declared false. It is denoted by Ho. The null hypothesis is usually the Type I error: Type I error occurs when a true null hypothesis is rejected. In the criminal trial, a Type I
hypothesis that sample observations result purely from chance. error is made when an innocent person is wrongly convicted. The value of  represents the probability of
 Alternative hypothesis. An alternative hypothesis is a claim about a population parameter that will committing this type of error; that is,
be true if the null hypothesis is false. The alternative hypothesis is denoted by HA or Ha or H1. The   P  H o is rejected \ H o is true  .
alternative hypothesis is the hypothesis that sample observations are influenced by some non-
random cause. The value of  represents the significance level of the test.

Example: A soft drink bottling company’s advertisement states that a bottle of its products contains 330 Type II error: Type II error occurs when a false null hypothesis is not rejected. A Type II error occurs
milliliters (ml.). But customers are complaining that the company is under filling its products. To check when a guilty defendant is acquitted. The value of  represents the probability of committing a type II
whether the complaint is true or not, an inspector may test the following hypotheses: error; that is:
Ho: The average content of a bottle of this product is no less than 330 ml, against,   P  H o is not rejected \ H o is false  .
HA: The average content of a bottle of this product is less than 330ml

Department of Management Page 1 Department of Management Page 2

 WRU CoBE Statistics for Management II Handout Chapter 3 WRU CoBE Statistics for Management II Handout Chapter 3

We often refer to those two possible errors as alpha error ''  " and the beta error "  ." Decision Rules in Hypothesis
 Error- the probability of making type I error
 Error - the probability of making type II error The analysis plan includes decision rules for rejecting the null hypothesis. In practice, statisticians describe
these decision rules in two ways - with reference to a P-value or with reference to a region of acceptance.
The value of 1   is called the power of the test. It represents the probability of not making a type II error.
The two types of errors that occur in tests of hypotheses depend on each other. The error probabilities are  P-value: The strength of evidence in support of a null hypothesis is measured by the P-value.
inversely related, meaning that any attempt to reduce one will increase the other. We cannot lower the Suppose the test statistic is equal to S. The P-value is the probability of observing a test statistic as
values of  and  simultaneously for a test of hypothesis for a fixed sample size. Lowering the value of  extreme as S, assuming the null hypothesis is true. If the P-value is less than the significance level,
will raise the value of  and lowering the value of  will raise the value of  , However, we can decrease we reject the null hypothesis.
 Region of acceptance: The region of acceptance is a range of values. If the test statistic falls
both  and  simultaneously by increasing the sample size.
within the region of acceptance, the null hypothesis is not rejected. The region of acceptance is
The following table presents the possible conclusions and errors in performing a test.
defined so that the chance of making a Type I error is equal to the significance level.
Actual situation
Ho is true Ho is false The set of values outside the region of acceptance is called the region of rejection. If the test
statistic falls within the region of rejection, the null hypothesis is rejected. In such cases, we say that
Decision Do not reject Ho Correct Type II or
decision  error the hypothesis has been rejected at α level of significance.
Reject Ho Type I or Correct
 error decision 3.4. One-Tailed and Two-Tailed Hypothesis Tests

Type of Tests A test of a statistical hypothesis, where the region of rejection is on only one side of the sampling
Based on the form of the null and alternative hypotheses, we have two types of tests: one-sided (one tailed) distribution, is called a one-tailed test. One tail hypothesis test further can be classified as right one tail test
tests and two-sided (two tailed) tests. and left one tail test. The basis to decide the type of test is mainly the sign of comparison used in the
alternative hypothesis part.
i) Test of the form
For example, suppose the null hypothesis states that the mean is less than or equal to 10. The alternative
 H o :   o Ho :   o
  hypothesis would be that the mean is greater than 10. The region of rejection would consist of a range of
 H A :   o H A :   o numbers located on the right side of sampling distribution; that is, a set of numbers greater than 10.
OR OR
 H o :   o  H o :   o When Ho contains only the = sign, Ha will contain the  sign. The test then is called a two-tail test
  because the rejection consists of two tails of the sampling distribution.
 H A :   o  H A :   o
Where the alternative hypotheses are inequality (ies) type are called one-sided (one-tailed) tests. Table 3.1: Mathematical symbols and type of test
o is the hypothesized (assumed) mean.
ii) Test of the form Mathematics symbols in Ha Type of test
H o :   o ≤ < Left one tail test
≥ > Right one tail test
H A :   o ,
≠ Two tail test
in which the critical region (rejection region) includes both large and small values of the test statistic are
called two-sided or two-tailed tests. Example: Identify the types of tail tests for the following pairs of hypothesis:
In hypothesis testing we reach at a conclusion based on sample evidence. We cannot be sure of our A) : > 0.4 : ≥ 0.45
conclusion and an error is made if a true Ho is rejected or a false Ho is accepted. B) : ≥ 0.12 : < 0.12
C) : = 24 : ≠ 24

Department of Management Page 3 Department of Management Page 4

 WRU CoBE Statistics for Management II Handout Chapter 3 WRU CoBE Statistics for Management II Handout Chapter 3

Solution: Here the appropriate test statistics is Z because though the population standard deviation is unknown, the
A) Right one tail test because the alternative hypothesis has ≥ sample size is large at 100.
B) Left one tail test because the alternative hypothesis has <
C) Two tail test because the alternative hypothesis has≠ Decision rule: Accept the null hypothesis if the is less than

3.5. Hypothesis Testing of Population Mean and Proportion

3.5.1. Hypothesis Testing For Population Mean (Large Samples)

Here the sample information is taken from a set of population were the population information is fully Reject
unknown or difficult to know. Then an assumption will be tested whether it is failed to accept or reject it.
The sample taken from the population is assumed to be large when > 30.
Accept Ho
If the standard deviation of the population is known, then based on the central limit theorem, then the Z

sampling distribution of the mean ̅ would follow the standard normal distribution for a large sample size. Ztab=1.64
0 5

The Z-statistics is given by: = =

√ .
= = = . = .
√ √
In this formula the numerator ( ̅ − ), measures how far the observed sample mean ̅ is from the = 1.645
hypothesized mean . The denominator ̅ is the standard error of the mean so the Z test statistics
represents how many standard errors . Decision: Accept Ho, i.e., the mean weight does not increase

If the population standard deviation is unknown, then a sample standard deviation S is used to estimate . 3.5.2. Hypothesis Testing for Population Mean (Small Samples)
The value of test statistics will be:
− In the previous discussion we employed or used Standard normal distribution (Z - test) because the sample
= size large. If the sample size is not large (n < 30), it is preferable to use student’s t - distribution. Thus, the
√ test statistics for determining the difference between the sample mean ̅ and population mean is given by:
Illustration 1: A packaging device is set to fill detergent powder packets with a mean weight of 5kg. The
standard deviation is known to be 0.01 kg. These are known to drift upwards over a period of time due to = =
machine fault, which is not tolerable. A random sample of 100 packets is taken and weighed. This sample √
has a mean weight of 5.03kg and a standard deviation of 0.21kg. Can we conclude that the mean weight Where s is unbiased estimation of unknown population standard deviation .This test statistics has a t-
produced by the machine has increased? Use a 5 percent level of significance. distribution with n-1 degrees of freedoms.

Solution: Illustration 2: Suppose the average breaking strength of steel rods is specified to be 18.5 thousand lbs. For
: = this a sample of 14 rods was tested. The mean and standard deviation obtained were 17.85 and 1.955,
respectively. Test the significance of deviation through 5% level of significance.
: >5
Solution: Let us take the null hypothesis that there is no significant deviation in the breaking strength of
( ) = 100, ̅ = 5.03 , = 0.21 ∝= 5%
the rods, that is,
: = 18.5 and : ≠ 18.5
Type of test: Two tail test

Department of Management Page 5 Department of Management Page 6

 WRU CoBE Statistics for Management II Handout Chapter 3 WRU CoBE Statistics for Management II Handout Chapter 3

= 14, ̅ = 17.85, = 1.955 To conduct a test of hypothesis, it is assumed that the sampling distribution of a proportion follows a
( ) = − 1 = 14 − 1 = 13 standardized normal distribution .Then using the value of the sample proportion ̅ and its standard
deviation ( ̅ ), we compute the value of Z-statistics as follows:
∝= 0.05
Since the tail is two tail tests, the given alpha has to be divided in to two equal parts as − −
∝/2 = 0.025 = =
The sample size is smaller and the population standard deviation is given as unknown (estimated using
sample deviation). Hence the appropriate test statistics to be used will be t - test. Illustration 3: Suppose a manufacturer claims that at least 95% of the equipments which he supplied to a
factory conformed to the specification. An examination of the sample of 200 pieces of equipment revealed
Decision rule: if the value of is between -2.16 and 2.16, accept the hypothesis else reject it. that 18 were faulty. Test the claim of the manufacturer.

Solution:
: ≥ . : < 0.95
Percent of pieces conforming the specification ( ̅ ) = 1 − = 0.91
Reject
Reject
= 200 (∝) = 5%

Accept Ho Decision rule: Accept Ho when Z cal is less than Z tab

Ztab=-2.16 0 Ztab=2.16

Reject Ho
. .
= = . =-1.24
√
Accept Ho
√

= ∝
, =− .
Z
Zcal = -2.67 Ztab=1.96
0
Decision: There is no significant deviation of sample mean from the population mean, i.e., accept Ho.

3.5.3. Hypothesis Testing for Population Proportion The appropriate test statistics for the distribution is:
− −
= =
We have seen how to conduct hypothesis tests for a mean. We now turn to proportions. The process is
completely analogous, although we will need to use the standard deviation formula for a proportion.
0.91 − 0.95
= = −2.67
Sometimes instead of testing a hypothesis pertaining to a population mean, a population proportion (P) of ( . ) .

values in a particular category is considered. For this random sample of size (n) is selected to compute the
proportion of success in a particular sample as follows: = 1.96

Decision: Reject Ho because is less than which is within the area of rejection. Hence we
= = conclude that the proportion of equipments conforming to specifications is not 95 percent.

Department of Management Page 7 Department of Management Page 8

 WRU CoBE Statistics for Management II Handout Chapter 3 WRU CoBE Statistics for Management II Handout Chapter 3

3.5.4. Hypothesis Testing of the Difference between Two Means ( ̅ − ̅ )−( )

=
Testing the difference implies checking the presence or absence of difference and their direction +
comparison of population parameter based on sample information taken from two different target
populations. This test will check the presence and type of difference between two independent population . . ) ( )
means based on sample mean values difference. = = 1.01
. .

Let and be the sample means obtained in large samples of sizes N1 and N2 drawn from respective
populations having means and and standard deviations and . Consider the null hypothesis that
there is no difference between the population means (i.e., = ), which is to say that the samples are
drawn from two populations having the same mean.
Reject Ho
The sampling distribution of differences in means is approximately normally distributed, with its mean and
standard deviation given by:
Accept
= 0 and = + Z
Zcal =1.01 Ztab=1.96
0
Illustration 4: Do employees perform better at work with music playing? The music was turned on during
the working hours of a business with 45 employees. Their productivity level averaged 5.2 with a standard
deviation of 2.4. On a different day the music was turned off and there were 40 workers. The workers' To decide whether to accept or reject the set null hypothesis, it is mandatory to determine both
productivity level averaged 4.8 with a standard deviation of 1.2. What can we conclude at the 0.05 level? and and conduct comparison.

Solution Z is the value of Z score obtained from table considering level of significance ∝ . Z at the table
with 95% confidence level is 1.96 and Z calculated is 1.01 which is less than Z tabulated.
We first develop the hypotheses:
:− ≤ 0 1.5.4. Hypothesis Test for the Difference between Two Population Proportions
: − > 0
= 1 − 2 and the standard deviation is In the previous sub-section we have discussed hypothesis testing of the difference between two population
means. Now we will discuss hypothesis testing of the difference between two population proportions. Let
= + two independent populations each having proportion and standard deviation of an attribute be as follows:

= 1 − 2 = 5.2 − 4.8 = 0.4 Population Proportion Standard deviation

. . 1
= + = + = 0.404
and 2
Now we can calculate the Z-score. We have
The sampling distribution of difference in sample proportions ( − ) is based on the assumption that the
difference between two population proportions ( − ) is normally distributed. The standard deviation of
sampling distribution of − is given by:
= +

The Z statistics for the difference between two population proportions is stated as:

Department of Management Page 9 Department of Management Page 10

 WRU CoBE Statistics for Management II Handout Chapter 3 WRU CoBE Statistics for Management II Handout Chapter 3

( − )−( − )
=

In variably, the standard error ̅ ̅ of the difference between sample proportions is not known. Thus
when a null hypothesis states that there is no difference between the population proportions, we combine
two sample proportions ( ) to get one unbiased estimates of population proportion as follows:
Reject

Pooled estimate =
The Z test statistics is then restated as: Accept Ho
−
= Z
Zcal=1.13 ZtabZ=1.96
=1.96
0 tab

Illustration 5: Suppose that a company is considering two different television advertisements for
promotion of a new product. Management believes that advertisement A is more effective than Substituting values in Z statistics, we have
advertisement B. Two test market areas with virtually identical consumer characteristics are selected:
advertisement A is used in one area and advertisement B in another area. In a random sample of 60 0.30 − 0.22
customers who saw advertisement A, 18 had tried the product. In a random sample 100 customers who saw = = 1.131
0.0707
advertisement B, 22 had tried the product. Does this indicate that advertisement A is more effective than
advertisement B, if a 5 percent level of significance is used? ∝= 0.05 = 1.96

Solution: : = : > Decision: There is no significance difference in the effectiveness of the two advertisements.
18 22
= 60, = = 0.30; = 100, = = 0.22
60 100
Level of significance ∝= 5
Test statistics:
( − )−( − )
=

Where, = ( + ); = −

. . . .
= ( + )= 0.0707

Decision rule: Accept Ho if is less than

Department of Management Page 11 Department of Management Page 12