Statistical Analysis in Finance

Session 4:
Hypothesis Testing

Dr. Nemanja Radić

www.cranfield.ac.uk/som

Statistical Analysis in Finance

Content :
Sessions 1 & 2: Probability and Probability Distributions
Session 3: Sampling and Estimations
SESSION 4: HYPOTHESIS TESTING
Session 5: Problem Solving
Sessions 6 & 7: Regression Analysis
Session 8: Regression Models with Dummy Variables
Sessions 9 &10: Problem Solving and Exam Revision
2
 Statistical Analysis in Finance

Reading:
Statistical Techniques in Business and Economics
(17/E) by Douglas A. Lind, William G. Marchal and
Samuel A. Wathen 2017. McGraw-Hill. Chapters
Chapters 10, 11 and 12.

Intended Learning Outcomes

• Understand how to structure null and alternative hypotheses

• Be aware of type I and type II errors

• Be able to interpret the results of such tests

• Conduct a test of hypothesis about a population mean as well

as a population proportion

• Understand the basic rationale for testing differences among

several population means.
4
 What is a Hypothesis?

HYPOTHESIS A statement about the value of one or more

population parameters developed for the purpose of testing.

• Is the underlying mean return on this mutual fund different

from the underlying mean return on its benchmark?

• Did the volatility of returns on this stock change after the

stock was added to a stock market index?

• Are security’s bid ask spreads related to the number of

dealers making a stock market in the security?

What is a Hypothesis?

HYPOTHESIS TESTING A procedure based on sample

evidence and probability theory to determine whether the
hypothesis is a reasonable statement.

STEPS OF HYPOTHESIS TESTING:

1. Stating the hypothesis.
2. Specifying the significance level.
3. Identifying the appropriate test statistics and its
probability distribution.
4. Stating the decision rule.
5. Collecting data and calculating the test statistics.
6. Interpret the results (make economic or investment
decision).
6
 Step 1: Stating the Hypothesis

• Hypothesis tests resolve conflicts between two competing

opinions (hypotheses).
• In a hypothesis test, define:

• NULL HYPOTHESIS (H0)

The presumed default state of nature or status quo.
For example, we hypothesize that the population mean risk
premium for UK equities is less than or equal to zero.

• ALTERNATE HYPOTHESIS (H1)

A contradiction of the default state of nature or status quo.
The H1 to the above example is that population mean risk
premium for UK equities is greater than zero.
7

Step1: Stating the Hypothesis (cont’d)

• We conduct hypothesis tests to determine if

sample evidence contradicts H0.

• On the basis of sample information, we either

• “Reject the null hypothesis”
• Sample evidence is inconsistent with H0.
• “Do not reject the null hypothesis”
• Sample evidence is not inconsistent with H0.
• We do not have enough evidence to “accept” H0.

8
 Step1: Stating the Hypothesis (cont’d)

Important Things to Remember about H0 and H1

• Equality is always part of H0 (e.g. “=” , “≥” , “≤”). Because

H0 is the statement being tested, and we need a specific
value to include in our conclusions.

• “≠” “<” and “>” always part of H1

• H0 and H1 are mutually exclusive and collectively

exhaustive.

Step1: Stating the Hypothesis (cont’d)

Some guidelines Keywords

Inequality Part
Symbol of:
Larger (or more) than > H1
• In problem solving, look
for key words and Smaller (or less) < H1
convert them into No more than £ H0
symbols. Some key At least ≥ H0
words include: Has increased > H1
“improved, better Is there difference? ≠ H1
than, as effective as, Has not changed = H0
different from, has Has “improved”, “is See left H1
changed, etc.” better than”. “is more text
effective”

10
 Step1: Stating the Hypothesis (cont’d)

• Suppose our question concerns the value of a population

parameter, say the mean of the population μ, in relation to
one possible value of the parameter, μ0.

• We can formulate the H0 and H1 in three different ways:

1. H0 : μ = μ0 versus H1 : μ ≠ μ0 (Two-tail test)

2. H0 : μ < μ0 versus H1 : μ > μ0 (Right-tail test)
3. H0 : μ > μ0 versus H1 : μ < μ0 (Left-tail test)

11

Step1: Stating the Hypothesis (cont’d):

Examples

1) A trade group predicts that back-to-school spending will average

£606.40 per family this year. A different economic model is
needed if the prediction is wrong. State the null hypothesis and
the alternative if a different economic model may be needed.

2) A television research analyst wishes to test a claim that no more

than 50% of the households will tune in for a TV episode. Specify
the null and the alternative hypotheses to test the claim.

3) It is generally believed that at least 0.6 of the residents of

Cranfield town are happy with their lives. A sociologist is
concerned about the lingering economic crisis and wants to
determine whether the crisis has adversely affected the happiness
level in Cranfield town. Specify the null and the alternative
hypotheses to determine if the sociologist‘s concern is valid.
12
 Step1: Stating the Hypothesis (cont’d):
Example 1

• A trade group predicts that back-to-school spending will

average £606.40 per family this year. A different
economic model is needed if the prediction is wrong.
State the null hypothesis and the alternative if a different
economic model may be needed.

ü Parameter of interest is m since we are interested in the

average back-to-school spending.

ü Since we want to determine if the population mean

differs from £606.4 (i.e, ≠), it is a two-tail test.

ü H0: m = 606.4
H1: m ≠ 606.4

13

Step1: Stating the Hypothesis (cont’d):

Example 2

• A television research analyst wishes to test a claim that no

more than 50% of the households will tune in for a TV
episode. Specify the null and the alternative hypotheses to
test the claim.

ü Parameter of interest is π since we are interested in the

proportion of households.

ü Since the analyst wants to determine whether π is no more

than 0.50, it is a one-tail test.

ü H0: π < 0.50

H1: π > 0.50

14
 Step1: Stating the Hypothesis (cont’d):
Example 3
• It is generally believed that at least 0.6 of the residents
of Cranfield town are happy with their lives. A sociologist
is concerned about the lingering economic crisis and
wants to determine whether the crisis has adversely
affected the happiness level in Cranfield town. Specify
the null and the alternative hypotheses to determine if
the sociologist ‘s concern is valid.

ü Parameter of interest is p since we are interested in the

proportion of Cranfield town population.

ü The population p has been at least 0.6.

ü H0: p ≥ 0.6
H1: p < 0.6

15

Step 2: Specifying the significance level

(cont’d)

Decision Null hypothesis is true Null hypothesis is false

Reject the null hypothesis Type I error Correct decision

Do not reject the null hypothesis Correct decision Type II error

16
 Example 4: Door Delivery

• Assume that there is a quality controller for a building company, a

supplier has just delivered 1000 doors. You sampled 50 doors, if
more than 6% of doors are faulty you reject the delivery.

• H0 is that the delivery contains 6% or less substandard doors.

• Of the 50 doors inspected, 4 doors are substandard, or 8%.

• The delivery is rejected because it exceeds the maximum of 6%.
• If the doors were substandard than the decision to return the
delivered doors was correct.

• However, suppose the 4 substandard doors selected in the sample

of 50 doors were the only substandard doors in the delivery of
1000 doors. 17

Example 4: Door Delivery (cont.)

• Then 0.004 were substandard doors. This is less than 6%. In

terms of hypothesis testing, we rejected a null hypothesis when
we should have failed to reject the null hypothesis. By rejecting
the H0 we committed a type I error.

• Suppose now in the sampled 50 doors there were 2

substandard (4%). Then the delivery is taken.

• It could be by chance the 48 doors in the sample were the only

acceptable doors of the entire delivery of 1000 doors!

• In this case the builder has accepted more than 6% faulty doors.
He committed type II error not rejecting when it is false.

18
 Step 2: Specifying the significance level
(cont’d)

• Type I Error
• Defined as the probability of rejecting the
null hypothesis when it is actually true.
• This is denoted by the Greek letter “a”
• Also known as the significance level of a
test

• Type II Error
• Defined as the probability of failing to reject
the null hypothesis when it is actually false.
• This is denoted by the Greek letter “β”
19

Step 2: Specifying the significance level

(cont’d)

• Controlling the probabilities of the two types of errors involves a

trade-off.
• Decreasing the probability of type I error by specifying smaller
significance (0.01 rather than 0.05), lead to an increase in the
probability of making type II error.
• The only way of reducing both probabilities is to increase the
sample size.
• Example: Consider the following competing hypotheses that relate
to the court of law.
H0: An accused person is innocent
HA: An accused person is guilty
• Consequences of Type I and Type II errors:
• Type I error: Conclude that the accused is guilty when in
reality, he/she is innocent.
• Type II error: Conclude that the accused is innocent
when in reality, he/she is guilty. 20
 Step 2: Specifying the significance level
(cont’d)

• A Type I error is considered more important than a Type II error,

so we protect against a Type I error to the detriment of a Type II
error. This is the same as in statistics.

• If we make it so hard to send an innocent man to jail (make α

small) by requiring massive amounts of evidence to convict
anybody, then a lot of guilty people will not be convicted (β
becomes bigger).

• On the other hand, if we make it really easy to send a guilty

person to prison by requiring only a small amount of evidence,
then a lot of innocent people will also go to prison (β is made
small, α gets big).

• In general, scientists like α to be 0.10, 0.05, or 0.01. 21

Step 3: Definition of Test Statistics

Ø TEST STATISTIC: a value, determined from sample

information, used to determine whether to reject the
null hypothesis.

Ø Then we determine what probability distribution the test

follows.

Ø In this session we will encounter the following

distributions:

Z-distribution
T-distribution
22
 Step 4: Stating the decision rule

Critical value approach:

makes the comparison in
terms of value of the test
statistics
Two approaches
based on the
level of
significance, a
p-value approach: makes the
comparison in terms of
probabilities

23

Step 4: Stating the decision rule: Critical

value approach

• The critical value approach specifies a region of

values, also called the rejection region, such that if
the test statistics falls into this region, then we reject
the null hypothesis.

• In the case of Z distribution Reject H0 if Z > Za

• In the case of t distribution Reject H0 if t > ta

24
 Step 4: Stating the decision rule: Critical
value approach (cont.)

The critical values for a Z distribution with a 0.05 level significance

ONE-TAIL, RIGHT TAIL TEST

25

Step 4: Stating the decision rule: Critical

value approach (cont.)

The critical values for a Z distribution with a 0.05 level significance

TWO-TAIL, NON-DIRECTIONAL TEST ONE-TAIL, LEFT TAIL TEST

26
 Step 4: Stating the decision rule:
p-Value approach

• The value of the test statistics is used to compute the p-

value.
• P-value is the probability of getting the observed value
of the test statistic, or a value with even greater
evidence against H0, if the null hypothesis is true.

• The p-value is compared to the significance level (a).

• Decision rule using the p-value:

Reject H0 if p-value < significance level

27

Step 5: Make a decision

Ø First, select a sample and compute the value of the test

statistic
Ø Compare the value of the test statistic to the critical value
Ø Then, make the decision regarding the null hypothesis

What does it mean when p-value <a?

(a) .10, we have some evidence that H0 is not true.
(b) .05, we have strong evidence that H0 is not true.
(c) .01, we have very strong evidence that H0 is not true.
(d) .001, we have extremely strong evidence that H0 is not true.
28
 Step 6: Interpret the results
• The process does not end with the value of a sample statistic or the decision
to reject or not reject the null hypothesis.
• What can we say or report based on the results of the statistical test?
• In a recent speech to students, the dean of the Business School reported that
the mean debt for students is £20,000. You decide to conduct a test of the
dean’s statement or hypothesis to investigate the statement’s truth.
H0: μ = £20,000
H1: μ ≠ £20,000
• A random sample of college students provides a sample mean and standard
deviation, and you compute a z-statistic. The hypothesis test results in a
decision to reject the null hypothesis. How do you interpret the result? The
sample evidence does not support the dean’s statement. Based on the
sample data, the mean amount of student debt is different from £20,000. You
have disproved the null hypothesis with a stated probability of a Type I error,
α. That is, there is a small probability that the decision to reject the null
29
hypothesis was an error due to random sampling.

Hypothesis tests concerning the mean (µ)

• Test that the mean of a population is equal to or

differs from some hypothesized value.

• Test if the sample means from two different

populations differ.

30
 Hypothesis Setups for Testing a single
Mean, µ

31

Tests concerning a single mean, μ:

σ is known

• The test is based on the sampling distribution of the

sample mean.

• It is essential that the sampling distribution of the sample

mean is normally distributed.

• Recall that the sampling distribution of the sample mean is

normally distributed when the underlying population is
normally distributed.

• If the underlying distribution is not normally distributed,

then by the C.L.T. , sampling distribution of the sample
mean is approximately normally distributed if the sample
size is sufficiently large, n ≥ 30.
32
 Tests concerning a single mean, μ:
σ is known (cont’d)

• The value of the test statistics for the hypothesis test of the
population mean, μ, when the population standard deviation
σ is known is computed as:

33

Example 5: Two-Tailed Test, σ Known

The waiting time for customers at Burger King Restaurants follows a normal
distribution with a population standard deviation of 1 minute. At the Victoria Road
Burger King, the quality-assurance department sampled 50 customers and found
that the mean waiting time was 2.75 minutes. At the .05 significance level, can we
conclude that the mean waiting time is less than 3 minutes?

Step 1: State the null hypothesis and alternate hypothesis.

Ho: µ ³ 3
H1: µ < 3

Step 2: Select the level of significance. Here α = .05

Step 3: Select the test statistic. In this example, we’ll use z
34
 Example 5 (Cont.)

Step 4: Formulate the decision rule by first determining the critical values of z.

Decision Rule: Reject Ho if z < -1.645

35

Example 5 (Cont.)

Step 5: Take sample, compute the test statistic, make decision.

Compute z:
x# − µ 2.75 - 3
z= = = -1.768
σ/ n 1/ 50

Decision: Reject Ho, because -1.768 does fall in the rejection region, we
decide to reject H0.

Step 6: Interpret the result.

p = 0.0384, found by (0.5000 – 0.4616). We conclude that the mean

waiting time is less than three minutes. When Ho is true, the probability
of obtaining a value smaller than -1.77 is 0.0384.
36
 Testing for the population mean μ:
unknown σ

• When the population standard deviation (σ) is

unknown, the test statistics for testing the population
mean, μ, is assumed to follow the tdf distribution with
n-1 degrees of freedom and its value is computed as:

X -µ
t n -1 =
s/ n
where,
t n -1 = t - statistic with n - 1 degrees of freedom
X = sample mean
µ = the hypothesized value of the population mean
s = sample standard deviation
37

Example 6

• You are analysing Etihad Equity fund, a midcap growth fund, that has
been in existence for 24 moths. During this period, it has achieved a
mean monthly return of 1.5% with a sample standard deviation of monthly
returns of 3.6%.

• Given its level of its systematic (market) risk and according to a pricing
model, this equity fund was expected to have earned a 1.10% mean
monthly return during that time period. Assuming that returns are
normally distributed, are the actual results consistent with a population
mean monthly return of 1.10%.

1. Formulate H0 and H1 consistent with the verbal description of the

research goal.
2. Identify the test statistic for conducting a test of the hypothesis in Part 1.
3. Using the critical approach, identify the rejection point or points for the
hypothesis tested in part 1 at 0.10 level of significance.
4. Determine whether the H0 is rejected at the 0.10 level of significance. 38
 Example 6: Solution

1. H0 : μ = 1.10 versus H1 : μ ≠ 1.10.

2. Because the population variance is not known,

we use t-test with 24 -1 = 23 degrees of freedom.

3. This is a two-tail test, we have a rejection point

t0.1, 23 . The two rejection points (from the table)
are 1.714 and -1.714. We reject the null
hypothesis if we find t > 1.714.

1.50 - 1.10 0.40

4. t23 = = = 0.544331 or 0.544
3.6 / 24 0.734847

39

Example 6: Solution (Cont.)

• We do not reject H0 because 0.544 falls outside

the rejection area.

• At the 10% level of significance, we conclude

that a population mean return of 1.10% level is
consistent with 24th - month observed data
series.

40
 Comparing two populations
– Some Examples

1. Is there any difference between the mean return on FTSE

100 index in the 1980s and the mean return in the 1990s?

2. Is there a difference in the mean value of residential real

estate sold by male agents and female agents in
Cranfield?

3. Is there a difference in the mean number of defects

produced on the day and the afternoon shifts at Kimble
Products?

4. Is there a difference in the mean number of days absent

between young workers (under 21 years of age) and older
workers (more than 60 years of age) in the fast-food
industry?
41

Comparing two population means

• The tests that we are going to discuss depend on the

assumptions that we make about :

• Whether the samples are independent (i.e. unrelated to

each other)

• The distribution of the two populations

• Standard deviation of the populations: known, unknown

(equal, unequal)

42
 Comparing two population means (Cont.)

If σ1 and σ 2 are known

X1 - X 2
z =
s 12 s 22
+
n1 n2

ØThe samples are from independent populations

ØSampling distribution of is assumed to be normally distributed.
ØA linear combination of normally distributed random variables is also
normally distributed.
ØIf the underlying distribution is not normal, then by the C.L.T., the
sampling distribution of is approximately normal only if both n1 ≥ 30
and n2 ≥ 30.
43

Comparing Two Population Means –

Example 7

• An economist claims that the average weekly food expenditure of

households in Bedford is more than that of households in Milton
Keynes. She surveys 35 households in Bedford and obtains an
average weekly food expenditure of £164. A sample of 30
households in Milton Keynes yields an average weekly food
expenditure of £159.

• Historical data reveals that the population standard deviation for

Bedford and Milton Keynes are £12.50 and £9.25 respectively.

a) Specify the competing hypotheses to test the economist’s claim.

b) Calculate the value of the test statistics and its associated p-
value.
c) At a = 5% significance level, is the economist’s claim supported
by the data? 44
 Example 7: Solution

a) The economist wishes to determine if the mean weekly

food expenditure in Bedford (µB) is more than that of
Milton Keynes (µM), or µB > µM. The hypotheses for
this right-tail test are:
H0: µB < µM,
H1: µB > µM

b) Since the population standard deviations are known,

we assume the Z test statistic.

164 - 159
z= = 1.85
2 2
12.5 9.25
+
35 30 45

Example 7: Solution

b) The p-value of the right-tailed test is computed as

p-value = P(Z > 1.85) = 0.5 - 0.4678 = 0.0322.

Since the p-value of 0.0322 < a = 0.05 (the

significance level) we reject the null hypothesis.

c) Therefore, at the 5% significance level, the economist

concludes that average weekly food expenditure in
Bedford is more than that of Milton Keynes.

46
 Comparing population means:
Equal but unknown population standard
deviations (the Pooled t-test)

The t distribution is used as the test statistic. The

required assumptions are:

1. The samples are from independent populations.

2. Sampling distribution of X - X is assumed to be normally

1 2

distributed. If the underlying distribution is not normal,

then by the C.L.T., the sampling distribution of X - X is1 2

approximately normal only if both n1 ≥ 30 and n2 ≥ 30.

3. The populations must have equal standard deviations.

47

Comparing population means:

Equal but unknown population standard
deviations (the Pooled t-test)

Finding the value of the test

statistic requires two steps. (n1 - 1) s12 + (n2 - 1) s22
s =
2
p
n1 + n2 - 2
1. Pool the sample standard
deviations.
2. Use the pooled standard
deviation in the formula. X1 - X 2
t=
3. The number of degrees of æ1 1 ö
s 2p çç + ÷÷
freedom is n1 + n2 - 2 è n1 n2 ø

48
 Example 8

The realized mean monthly return on a sample the FTSE All Share index
in the 1990s appears to have been substantially different from the mean
return on a sample of the FTSE All Share index in the 1980s. Was the
difference statistically significant? Assuming equal population variance,
determine whether the null hypothesis is rejected or not at the 0.05 level
of significance.
1. Formulate the null and alternate hypothesis
2. Assuming equal population variance, determine whether the null
hypothesis is rejected or not at the 0.05 level of significance.

Decade # of Months Mean Monthly Standard

(n) Return (%) Deviation
1980s 120 0.580 4.598
1990s 120 1.470 4.738
49

Example 8: Solution

1) State the null and alternative hypotheses.

H0: µ80s = µ90s
H1 : µ80s ≠ µ90s

2) Determine the appropriate test statistic.

Because the two samples are drawn from different decades,
they are independent. Given the size of the sample of 120, then
by CLT we can assume that the distribution of difference in the
sample means of 80s and 90s is approximately normal. The
population variances are unknown but assumed to be equal. We
can use the t-test with degrees of freedom equal 120 + 120 -2 =
238 degrees of freedom.
50
 Example 8: Solution (Cont.)

Formulate a decision rule.

From the table the closest degree of freedom to 238 is 200. For a two
sided test, the rejection points are ± 1.972 for 0.05 level.

(n1 - 1) s12 + (n 2 - 1) s 22 (120 - 1)(4.598) 2 + (120 - 1)(4.738) 2

s =
2
p =
n1 + n 2 - 2 120 + 120 - 2
5,187.24
s 2p = = 21.80
238
µ 80 s - µ 90 s 0.580 - 1.470
t= 1/ 2
= 1/ 2
æ s 2p s 2p ö æ 21.80 21.80 ö
ç + ÷ ç + ÷
ç n1 n 2 ÷ è 120 120 ø
è ø
- 0.89
t= = -1.477
0.602704 51

Example 8: Solution (Cont.)

The t-value of -1.477 does not fall in the

rejection region (i.e below 1.972).
Therefore, we do not reject the null hypothesis.

52
 Comparing population means: Unequal and
unknown population standard deviations

Use the formula for the t-

statistic shown on the right if it
is not reasonable to assume
the population standard
deviations are equal.

The degrees of freedom are

adjusted downward by a rather
complex approximation
formula. The effect is to reduce
the number of degrees of
freedom in the test, which will
require a larger value of the
test statistic to reject the null
hypothesis.
53

Test for Equal Variances

• Inferences about the ratio of two population variances are

based on the ratio of the corresponding sample variances.
• These inferences are based on a new distribution: F
distribution.
• Reject H0 if F > Fa/2,df1,df2 ; df1 (n1-1) and df2 (n2-1)

54
 Characteristics of F-Distribution

1. There is a “family” of F
Distributions. A particular member
of the family is determined by two
parameters: the degrees of
freedom in the numerator and the
degrees of freedom in the
denominator.
2. The F distribution is continuous
3. F value cannot be negative.
4 . The F distribution is positively
skewed.
5. It is asymptotic. As F ® ¥ the
curve approaches the X-axis but
never touches it. 55

Hypothesis testing involving dependent

samples

Dependent samples are samples that are paired or related in some

fashion
1. Dependent sample is characterized by a measurement followed by
an intervention of some kind and then another measurement. This
could be called a “before” and “after” study. Paired observations are
from the same unit.
For example:
• If you wished to compare the dividend policy of companies before
and after a change in the tax law. You would have pairs of “before”
and “after” observations for the same companies. So, you would be
testing a hypothesis about the mean of the differences (mean
differences) that you observe across companies. 56
 Hypothesis testing involving dependent
samples

2. Dependent sample is characterized by matching or pairing

observation. Paired observations are not form the same
unit.

For example:
• If you wished to test whether the mean returns earned by
two investment strategies were equal over a study period.
The observations here are dependent in the sense there is
one observation for each strategy in each month, and both
strategies are likely to be related to some common market
risk factors.
57

Hypothesis testing involving dependent

samples

Use the following test when the samples are dependent:

d
t=
sd / n
There are n - 1 degrees of freedom
d is the mean of the difference between the paired observations.
s d is the standard deviation of the differences between the paired observations.
n is the number of paired observations.

• The statistical test requires that both populations are normally distributed
or the sample size of the sample n ≥ 30.
58
 Example 9

A researcher examined the popular investment strategy of investing in the

10 stocks with highest yields in the FTSE, compared with a buy-and-hold
strategy of the FTSE-100. The study period was for 50 years, from 1961 to
2010. Based on the below data, determine whether the mean difference
between the FTSE-10 and the FTSE-100 strategies equal 0. Use 0.01 level
of significance. Discuss the choice of the test that you have used.

Strategy Mean Return (%) Standard

Deviation (%)
FTSE-10 16.77 19.10
FTSE-100 13.71 16.64
Difference 3.06 6.62*

* Sample standard deviation of differences. 59

Example 9: Solution

• With μd as the underlying mean difference between the FTSE-10 and

FTSE-100 strategies, we have H0 : μd = 0 versus H1 : μd ≠ 0 .

• The population variance is unknown, so

we use the following t-test with 49
degrees of freedom and a = 1%. From the table the critical value is 2.68

• t49 = (3.06)/[6.62/(50)1/2] = 3.2685 or 3.27.

• We reject the null hypothesis and we can conclude that the difference in
mean returns was clearly statistically significant.

• The FTSE-10 and the FTSE-100 are not independent samples; in general,
the correlation of returns on the FTSE-10 and FTSE-100 should be
positive. Because the samples are dependent, a paired comparison test
was appropriate.
60
 Test Statistic for testing a single population
proportion

• A Proportion is the fraction or percentage that indicates the part of the population
or sample having a particular trait of interest.
• The sample proportion is denoted by p and is found by x/n
• Assumptions in testing a population proportion using the z-distribution are met (in
the slides of the last lecture).
• The test statistic is computed as follows:
Hypothesized
Sample proportion population proportion

p -p
z =
p (1 - p )
n

Sample size 61

Two-sample tests about proportions

EXAMPLES
• The vice president of human resources wishes to know whether there
is a difference in the proportion of hourly employees who miss more
than 5 days of work per year at the Atlanta and the Houston plants.

• General Motors is considering a new design for the Pontiac Grand

Am. The design is shown to a group of potential buyers under 30
years of age and another group over 60 years of age. Pontiac wishes
to know whether there is a difference in the proportion of the two
groups who like the new design.

• A consultant to the airline industry is investigating the fear of flying

among adults. Specifically, the company wishes to know whether
there is a difference in the proportion of men versus women who are
fearful of flying.
62
 Two sample tests of proportions

• We investigate whether two samples came from populations with

an equal proportion of successes.

• The two samples are pooled using the following formula.

63

Two sample tests of proportions (cont’d)

The value of the test statistic is computed from the

following formula.

64