CHAPTER THREE
HYPOTHESIS TESTING
Introduction
A research worker or an experimenter has always some fixed ideas about certain
population(s) vis-à-vis population parameter(s) based on prior experiments, surveys or
experience. Sometimes these ideas might have been fixed in the mind vicariously. There
is a need to ascertain whether these ideas or claims are correct or not by collecting
information in the form of data. In this way, we come across two types of problems, first
is to draw inferences about the population on the basis of sample data and the other is to
decide whether our sample observations have come from a postulated population or not.
The first type of problem, the problem of estimation, has almost been covered in the
previous unit. In this unit, we would be dealing with the second type of problem, the
problem of hypothesis testing.
Generally, a hypothesis is established before hand. By hypothesis we mean to give
postulated or stipulated value(s) of a parameter. Also, instead of giving values, some
relationship between parameters is postulated in the case of two or more populations. On
the basis of observational data, a test is performed to decide whether the postulated
hypothesis be accepted or not and this involves certain amount of risk. This amount of
risk is termed as a level of significance.
significance. When the hypothesis is accepted, we consider it a
non-significant result and if the revenue situation occurs, it is called a significant result.
A test is defined as a statistical procedure governed by certain rules, which leads to take a
decision about the hypothesis for its acceptance or rejection on the basis of sample
values.
Statistical tests of hypothesis play an important role in industry, biological sciences,
social sciences, economics, etc. The use of tests has been made clear through a number of
practical problems.
1
5.1. BASIC CONCEPTS IN HYPOTHESIS TESTING
Basic concepts in the context of testing of hypothesis need to be explained.
a) Null Hypothesis and Alternative Hypothesis
A hypothesis is an assertion or conjecture about the parameter(s) of population
distribution(s). Alternative hypothesis is usually the one which one wishes to prove and
the null hypothesis is the one which one wished to disprove. Thus, a null hypothesis
represents the hypothesis we are trying to reject, and alternative hypothesis represents all
other possibilities.
A hypothesis which is to be actually tested for acceptance or rejection is termed as null
hypothesis. It is denoted by Ho and alternative hypothesis which is denoted by H 1 or HA
is a statement about the population parameter or parameters, which gives an alternative to
the null hypothesis within the range of pertinent values of the parameter.
b) Level of Significance:
Significance: -
Is the maximum value of the probability of rejecting Ho when it is true and is usually
determined in advance before testing the hypothesis or level of significance is the
quantity of risk of the type I error which we are ready to tolerate in making a decision
about Ho. In other words, it is the probability of type I error, which is tolerable. The level
of significance is denoted by and is conventionally chosen as 0.05 or 0.01. = 0.01 is
used for high precision and = 0.05 for moderate precision.
c) Decision rule or test of hypothesis: -
Given a hypothesis Ho and an alternative hypothesis Ha, we make a rule which is known
as decision rule according to which we accept Ho (i.e., reject Ha) or reject Ho (i.e., accept
Ha). For instance, if Ho is that a certain lot is good (there are very few defective items in
it) against Ha that the lot is not good (there are too many defective items in it), then we
must decide the number of items to be tested and the criterion for accepting or rejecting
the hypothesis. We might test 10 items in the lot and plan our decision saying that if there
2
are none or only 1 defective item among the 10, we will accept Ho otherwise we will
reject Ho (or accept Ha). This sort of basis is known as decision rule.
5.2. Type I and Type II errors: -
In the context of testing of hypothesis, there are basically two types of errors we can
make. We may reject Ho when Ho is actually true and we may accept Ho when in fact
Ho is false. The former is known as Type I error and the latter as type II error. In other
words, Type I error means rejection of hypothesis which should have been accepted and
Type II error means accepting the hypothesis which should have been rejected. Type I
error is denoted by (alpha) known as error, also called the level of significance of the
test; and Type II error is denoted by (beta) known as error. The probability of Type I
error is usually determined in advance and is understood as the level of significance of
testing the hypothesis.
5.3. Two-tailed and one-tailed tests
These two terms are quite important and must be clearly understood. A two-tailed test
rejects the null hypothesis if, say, the sample mean is significantly higher or lower than
the hypothesized value of the mean of the population. Such a test is appropriate when the
null hypothesis is some specified value and the alternative hypothesis is a value not equal
to the specified value of the null hypothesis. Symbolically, the two-tailed test is
appropriate when we have Ho: = o and Ha: o which mean > o or <o. Thus,
in a two-tailed test, there are two rejection regions, one on each tail of the curve which
can be illustrated as under:
3
Acceptance and rejection regions in case of two-
tailed test with 5% significance level in Z-
statistic
acceptance region
Accept Ho if falls in this region
0.475 0.475
0.025 of area area
0.025
of area of area
Both taken together
Equals 0.95 or
95% of area
Z = -1.96 = o Z = 1.96
Reject Ho if the sample
mean falls in either of
these two regions
Mathematically we can state:
Acceptance Region A: 1.96
Rejection Region R: > 1.96
A one-tailed test would be used when we are to test, say, whether the population mean is
either lower than or higher than some hypothesized value. For instance, if our Ho: = o
and Ha: < o, then we are interested in what is known as Left-tailed test (wherein there
is one rejection rejoin only on the left tail) which can be illustrated as below
Rejection Acceptance region
region Mathematically
acceptance region
A: Z>-1.645
Rejection region
0.45 0.5 of
R: Z -1.645
of area area
both 0.95 of
the area
Z=-1.64
Reject Ho if the sample Z = -1.64 =0
Mean falls here
4
In case our Ho: = o and Ha: > o, we are then interested in what is known as one-
tailed test (right-tail) and the rejection region will be on the right tail of the curve as
shown below.
Acceptance and rejection regions in case of one-
tailed test (right tail) with 5% significance level
in Z –statistic
Rejection
acceptance region
region
0.5 0.45
area area
[both 0.95
area] 0.05 of area
= o Z = 1.645
Reject Ho if the sample mean
Mathematically we can state: falls in this region
Acceptance Region A: Z 1.645
Rejection Region R: Z > 1.645
5.4. TESTING A HYPOTHESIS ABOUT A SINGLE POPULATION MEAN
Mean of the population can be tested presuming different situation such as the population
may be normal or other than normal, it may be finite or infinite, sample size may be large
or small, variance of the population may be known or unknown and the alternative
hypothesis may be two-sided or one sided. Our testing technique will differ in different
situations. We may consider some of the important situations.
1. Population normal, population infinite, sample size may be large or small but
variance of the population is known, Ha may be one-sided or two-sided:
In such a situation, the test statistic is taken to be the Z-test, which is worked out as
under:
Z=
Where - is the sample mean
o – is the hypothesized population mean under Ho.
5
p – is the population variance
n – is the sample size.
Example: - A sample of 400 male students is found to have a mean height 67.47 inches.
Can it be reasonably regarded as a sample from a large population with mean height
67.39 inches and standard deviation 1.30 inches? Test at 5% level of significance.
Solution: - Taking the null hypothesis that the mean height of the population is equal to
67.39 inches, we can write:
Ho: = 67.39”
Ha: 67.39”
And the given information as = 67.47”, = 1.30”, n = 400. Assuming the
population to be normal, we can work out the test statistic Z as under:
Z= = 1.231
As Ha is two-sided in the given question, we shall be applying a two-tailed test for
determining the rejection regions at 5% level of significance, which comes to as under.
Using normal curve area table:
R: |z| > 1.96
The observed value of Z/2 is 1.231 which is the acceptance region since R: |z| > 1.96 and
thus Ho is accepted. We may conclude that the given sample (with mean height = 67.47”)
can be regarded to have been taken from a population with mean height 67.39 and
standard deviation 1.30 at 5% level of significance.
If the alternative hypothesis is
Ha: < o = 67.39”
Then R: Z < -Z
Zcal < -1.645
i.e compare 1.231 with –1.645 as Zcal > -1.645 then Ho is accepted
If the alternative hypothesis is
Ha: > o = 67.39” then compare Zcal = 1.231 with Z = 1.645 in this case the
rejection region R: Z > Z but as 1.231 < 1.645, again Ho remains accepted.
6
Check Your Progress –1
1. The following information is available
Ho: = 50
H1= 50
The sample mean is 49, and the sample size is 36. The population standard
deviation is 5. Use the .05 significance level.
a) is this a one-tailed or a two-tailed test?
b) State a decision rule.
c) Compute the value of the test statistic
d) What is your decision regarding Ho?
1. Population normal, population finite, sample size may be large or small but
variance of the population is known, Ha may be one-sided or two-sided:
In such a situation Z-test is used and the test statistic Z is worked out as under (using a
finite population correction):
Z=
Example: - Suppose we are interested in a population of 20 industrial units of the same
size, all of which are experiencing excessive labor turnover problems. The past records
show that the mean of the distribution of annual turnover is 320 employees, with a
standard deviation of 75 employees. A sample of 5 of these industrial units is taken at
random which gives a mean of annual turnover as 300 employees. Is the sample mean
consistent with the population mean? Test at 5% level.
Solution: - Taking the null hypothesis that the population mean is 320 employees, we
can write:
Ho: = 320 employees
Ha: 320 employees
And the given information as under:
= 300 employees, p = 75 employees
n = 5; N = 20
7
Assuming the population to be normal, we can workout the test statistic Z as
under:
Z=
= = -0.67.
As Ha is two-sided in the given question, we shall apply a two-tailed test for determining
the rejection regions at 5% level of significance which comes to as under, using normal
curve area table:
R: |z| > 1.96
The observed value of Z is –0.67 which is in the acceptance region since R: |z| > 1.96 and
thus, Ho is accepted and we may conclude that the sample mean is consistent with
population mean i.e., the population mean 320 is supported by sample results.
2. Population normal, population infinite, sample size small and variance of the
population unknown, Ha may be one-sided or two-sided.
In such a situation, t-test is used and the test statistic t is worked out as under:
t= with degree of freedom = (n – l)
and s = sample variance =
Example: - From past records it is known that the mean life of a battery used in a digital
clock is 305 days. The lives of the batteries are normally distributed. The battery was
recently modified to last longer. A sample of 20 modified batteries were tested. It was
discovered that the mean life was 311 days. And the sample standard deviation was 12
days. At the 0.05 level of significance, did the modification increase the mean life of the
battery?
Solution: - Sample size n = 20 is small.
Ho: o = 305
8
H1: > o
Sample variance s =12, sample mean = 311
Test statistics: t =
= 2.236
and since the test is one-sided (right-tailed) the rejection region R: t > t ( = 0.05)
t-tabulated for (n – 1) = 19 degree of freedom from the t -table is t0.05 (1 9) = 1.729
as t = 2.236 > 1.729 is in the rejection region, then reject Ho and accept H 1, that the mean
is greater than 305 days. It is concluded that the modification increased battery life.
Similarly: - for small sample with unknown population the two-tailed test is obtained by
comparing |t| with t/2(n-1). i.e. R: |t| > t/2(n-1) and for the alternative hypothesis < o,
(left tailed) test one should compare t calculated with -t (n-1) i.e. the rejection region is
R: t < -t(n-1)
3. Population normal, population finite, sample size small and variance of the
population unknown, and Ha may be one-sided or two-sided:
In this case, t- test is used and the test statistic ‘t’ is worked out as under (using finite
population multiplier):
t= with d.f = (n-1) and
s=
4. Population may not be normal but sample size is large, variance of the
population may be known or unknown, and Ha may be one – sided or two-
sided:
In such a situation, by central limit theorem, we can assume the normal distribution
and use Z-test and workout the test statistic Z as:
9
Z= in case of infinite
Population with known variance and replace by s (sample variance) if the population
variance p is not known.
And for a finite population, using the finite population correction for the standard error,
we use the modified statistic Z as:
Z= for known finite population variance p and replacing p by
s for if the population variance p is not known.
The decision rule is similar to those examples given earlier for one-tailed and two-tailed
tests using area under standard normal distribution.
5.5. TESTING A HYPOTHESIS ABOUT A POPULATION PROPORTION
Some assumptions must be made and conditions met before testing a population
proportion. To test a hypothesis about a population proportion, a random sample is
chosen from the population. This is called the experiment. It is assumed that the binomial
assumptions are met: (1) the sample data collected are result of counts; (2) an outcome of
an experiment is classified into one of two mutually exclusive categories – a “success” or
a “failure”; (3) the probability of a success stays the same for each trial; and (4) the trials
are independent, meaning that the outcome of one trial does not affect the outcome of any
other trial.
The test we will conduct shortly is appropriate when both np and n(1-p) are greater than
5. n is the sample size, and p is the population proportion. In addition, n should be large.
Researchers disagree on how large n should be; some say 30, some say 50, and others say
100. For the purpose of this book, the sample size should be at least 50. The distribution
of sample proportions is approximately Normally distributed by the central limit theorem
for sufficiently large sample size. Thus, we can use the Normal distribution table to find
the critical values for testing hypothesis about the population proportion (true
proportion). This test is a good example of the case where in the normal probability
10
distribution is applied to approximate a binomial probability distribution with a great deal
of accuracy.
If Po is the assumed or hypothesized population proportion, then the test statistic will be:
Where P is the sample
Zcal = proportion, and n is the
sample size
Example: - Suppose prior elections in a state indicated that it is necessary for a candidate
for governor to receive at least 50 percent of the vote in the northern section of the state
to be elected. The incumbent governor is interested in assessing his chances of returning
to office and plans to conduct a survey consisting of 2000 registered voters in the
northern section of the state. If the sample survey of the 2000 potential voters in the
northern part of the state revealed that 1, 500 planned to vote for the incumbent governor,
will the incumbent governor be re elected? Test at 5% level of significance.
Solution: - as n.p and n(1-p) exceed 5 and n is greater than 50, the following test of
hypothesis can be conducted.
Ho: P .80
H1: P < .80
As H1 states a direction, the test is one-tailed.
Z is an appropriate statistic, found by
Z= where P is the sample proportion P = = 0.75
and Po is the population proportion = 0.8
Thus Zcal = = -5.5
and -Z tabulated, -Z0.05 = -1.645
the rejection region R: Z < -Z and since Zcal = -5.5 < -1.645 which is in the rejection
region, the null hypothesis is rejected at 0.05 level. i.e. The difference of 5% points
11
between the sample percent (75%) and the hypothesized population percent 80% is
statistically significant.
It is probably not due to sampling variation. To put it another way, the evidence at this
point does not support the claim that the incumbent governor will return to the governor’s
mansion for another five years.
The two-tailed and right-tailed tests can be conducted accordingly for population
proportions.
Check Your Progress –2
1. The following hypotheses are given.
Ho: P = .40
H1: P .40
A sample of 120 observations revealed that the sample proportion P = .30. At the .05
significance level, can the null hypothesis be rejected?
SUMMARY
In this unit, some basic concepts in hypothesis testing have been discussed. A hypothesis
testing is a part of statistical inference in which an assumption or hypothesized idea about
the population parameter(s) is proved or disproved even though not absolutely but
probabilistically by the help of observed samples. We have seen whether the population
mean, proportion or variance equals the assumed value. In hypothesis testing procedures,
stating the null and alternative hypothesis clearly and fixing the level of significance and
selecting an appropriate test statistic are the crucial steps. Calculating the statistic
carefully and reading statistical tables are also important steps for accuracy.
ANSWERS TO SELF-REVIEW PROBLEMS
1. a) Two tailed
b) Reject Ho and accept H1. where Z does not fall in the region from –1.96 to 1.96
c) –1.2 found by
12
Z= = -1.2
d) Fail to reject Ho
2. Z= = -2.236
Rejection region:
R: |Z| > Z/2
|Z| > Z0.025
|Zcal| > 1.96
Zcal < -1.96 or Zcal > 1.96
as Zcal = -2.236 < -1.96 is in the rejection region, then Ho is rejected.
rejected.
13
