Chapter 9

Statistical Inference: Hypothesis Testing for

Single Proportions
A foremost statistical mechanism for decision making is the hypothesis test. The concept of hypothesis
testing lies at the heart of inferential statistics, and the use of statistics to “prove” or “disprove” claims
hinges on it. With hypothesis testing, business researchers are able to structure problems in such a way
that they can use statistical evidence to test various theories about business phenomena.

9.1 Introduction to Hypothesis Testing

In searching for answers to questions and in attempting to find explanations for business phenomena,
business researchers often develop “hypotheses” that can be studied and explored. Hypotheses are
tentative explanations of a principle operating in nature. This text explores various types of hypotheses,
how to test them, and how to interpret the results of such tests so that useful information can be brought to
bear on the business decisionmaking process.

Statistical Hypotheses
In order to scientifically test research hypotheses, a more formal hypothesis structure needs to be set up
using statistical hypotheses. Instead of attempting to prove or disprove research hypotheses directly in
this manner, business researchers convert their research hypotheses to statistical hypotheses and then test
the statistical hypotheses using standard procedures. All statistical hypotheses consist of two parts, a null
hypothesis and an alternative hypothesis. These two parts are constructed to contain all possible
outcomes of the experiment or study. Generally, the null hypothesis states that the “null” condition exists;
that is, there is nothing new happening, the old theory is still true, the old standard is correct, and the
system is in control. The alternative hypothesis, on the other hand, states that the new theory is true, there
are new standards, the system is out of control, and/or something is happening. It is common symbolism
to represent the null hypothesis as H0 and the alternative hypothesis as Ha.

Statistical hypotheses are written so that they will produce either a one-tailed or a two-tailed test. Two-
tailed tests always use = and in the statistical hypotheses and are directionless in that the alternative
hypothesis allows for either the greater than or less than a possibility. One-tailed tests are always
directional, and the alternative hypothesis uses either the greater than or the less than a particular value.

Rejection and Nonrejection Regions

Using the critical values established of the hypothesis testing process, the possible statistical outcomes of
a study can be divided into two groups:
1. Those that cause the rejection of the null hypothesis
2. Those that do not cause the rejection of the null hypothesis.
89
90 Statistical Inference: Hypothesis Testing for Single Proportions

Conceptually and graphically, statistical outcomes that result in the rejection of the null hypothesis lie in
what is termed the rejection region. Statistical outcomes that fail to result in the rejection of the null
hypothesis lie in what is termed the nonrejection region.

Type I and Type II Errors

Because the hypothesis testing process uses sample statistics calculated from random data to reach
conclusions about population parameters, it is possible to make an incorrect decision about the null
hypothesis. In particular, two types of errors can be made in testing hypotheses: Type I error and Type II
errors. A Type I error is committed by rejecting a true null hypothesis. A Type II error is committed
when a business researcher fails to reject a false null hypothesis. In this case, the null hypothesis is false,
but a decision is made to not reject it. Note that each action alternative contains only one of the errors
along with the possibility that a correct decision has been made. Power, which is equal to 1 − β , is the
probability of a statistical test rejecting the null hypothesis when the null hypothesis is false.

9.2 Testing Hypotheses About a Population Mean Using the z

Statistic (s Known)
In searching for answers to questions and in attempting to find explanations for business phenomena,
business researchers often develop “hypotheses” that can be studied and explored. Hypotheses are
tentative explanations of a principle operating in nature. This text explores various types of hypotheses,
how to test them, and how to interpret the results of such tests so that useful information can be brought to
bear on the business decision making process. One of the most basic hypothesis tests is a test about a

population mean. The z test formula for a single mean is .

Testing the Mean with a Finite Population

If the hypothesis test for the population mean is being conducted with a known finite population, the
population information can be incorporated into the hypothesis-testing formula. Doing so can increase the
potential for rejecting the null hypothesis. However, remember from Chapter 7 that if the sample size is
less than 5% of the population, the finite correction factor does not significantly alter the solution.
Formula 9.1 can be amended to include the population information. The formula to test hypotheses about

µ with a finite population is:

Using the p-Value to Test Hypotheses

Another way to reach a statistical conclusion in hypothesis testing problems is by using the p-value,
sometimes referred to as observed significance level. The p-value is growing in importance with the
increasing use of statistical computer packages to test hypotheses. No preset value of is given in the p-
value method. Instead, the probability of getting a test statistic at least as extreme as the observed test
statistic (computed from the data) is computed under the assumption that the null hypothesis is true.
Chapter 9 91

Virtually every statistical computer program yields this probability (p-value). The p-value defines the
smallest value of alpha for which the null hypothesis can be rejected. For example, if the p-value of a test
is 0.038, the null hypothesis cannot be rejected at α = 0.01 because 0.038 is the smallest value of alpha for
which the null hypothesis can be rejected. However, the null hypothesis can be rejected for α = 0.05.

Demonstration Problem 9.1

In an attempt to determine why customer service is important to managers in the United Kingdom,
researchers surveyed managing directors of manufacturing plants in Scotland. One of the reasons
proposed was that customer service is a means of retaining customers. On a scale from 1 to 5, with 1
being low and 5 being high, the survey respondents rated this reason more highly than any of the others,
with a mean response of 4.30. Suppose U.S. researchers believe American manufacturing managers
would not rate this reason as highly and conduct a hypothesis test to prove their theory. Alpha is set at
0.05. Data are gathered and the following results are obtained. Use these data and the eight steps of
hypothesis testing to determine whether U.S. managers rate this reason significantly lower than the 4.30
mean ascertained in the United Kingdom. Assume from previous studies that the population standard
deviation is 0.574.
34554554444444454443444354454445

Step 1: State the hypotheses: H0: µ = 4.30; Ha: µ < 4.30.

Step 2: Determine the appropriate test. The z test statistic is defined as .

Step 3: Specify the Type I error rate: α = 0.05.

Step 4: State the decision rule. This test is a one-tailed test and the observed test statistic must be less than
the critical z value to be rejected. This value is calculated by using the function "= NORM.S.INV(0.05)".
The value is -1.645. You can compare this z-value with the critical z-value for α = 0.05, which is -1.96.
The calculated z-value is greater than the critical z-value so the result is not significant.

Step 5: Gather the sample data. In this case, input the data into Excel in one column.

Step 6: Calculate the value of the test statistic and perform a z test. Go to an empty cell in an Excel
worksheet and from the Insert Function (fx), select the Statistical category and the function Z.TEST.
Select the observed values in Array, 4.3 as the X value (null hypothesis value, and Sigma as 0.574. The
result of the z test is 0.9217. The output is the right-tailed p-value for the test statistic. Subtract this value
from one to get the lower tail probability of 0.078.
92 Statistical Inference: Hypothesis Testing for Single Proportions

Step 7: State the statistical conclusion. Because the observed test statistic is not less than the critical value
and is not in the rejection region, the statistical conclusion is that the null hypothesis cannot be rejected.
The same result is obtained using the p-value method. The observed test statistic is z = -1.42. The
probability of getting a z value at least this extreme when the null hypothesis is true is 0.078. Hence, the
null hypothesis cannot be rejected at α = .05. If α = 0.10, the decision would be to reject the null
hypothesis.

STEP 8. Make a managerial decision. The test does not result in enough evidence to conclude that U.S.
managers think it is less important to use customer service as a means of retaining customers than do UK
managers

9.3 Testing Hypotheses About a Population Mean Using the t

Statistic (σ Known)
Very often when a business researcher is gathering data to test hypotheses about a single population
mean, the value of the population standard deviation is unknown and the researcher must use the sample
standard deviation as an estimate of it. In such cases, the z test cannot be used. Chapter 8 presented the t
distribution, which can be used to analyze hypotheses about a single population mean when σ is unknown
if the population is normally distributed for the measurement being studied. In this section, we will
examine the t test for a single population mean. In general, this t test is applicable whenever the
researcher is drawing a single random sample to test the value of a population mean (µ), the population
standard deviation is unknown, and the population is normally distributed for the measurement of

interest. The formula for a t test follows: , where df = n − 1.

Demonstration Problem 9.2

Figures released by the U.S. Department of Agriculture show that the average size of farms has increased
since 1940. In 1940, the mean size of a farm was 174 acres; by 1997, the average size was 471 acres.
Between those years, the number of farms decreased but the amount of tillable land remained relatively
constant, so now farms are bigger. This trend might be explained, in part, by the inability of small farms
to compete with the prices and costs of large-scale operations and to produce a level of income necessary
to support the farmers’ desired standard of living. Suppose an agribusiness researcher believes the
average size of farms has now increased from the 1997 mean figure of 471 acres. To test this notion, she
randomly sampled 23 farms across the United States and ascertained the size of each farm from county
records. The data she gathered follow. Use a 5% level of significance to test her hypothesis. Assume that
number of acres per farm is normally distributed in the population.
445 489 474 505 553 477 454 463 466 557 502 449 438 500 466 477 557 433 545 511 590 561 560

Step 1: State the hypotheses: H0: µ = 471; Ha: µ > 471.

Step 2: Determine the appropriate test. The t test statistic is defined as .

Chapter 9 93

Step 3: α = 0.05.

Step 4: Find the critical t value. This test is a one-tailed test and the observed test statistic must be less
than the critical t value to be rejected. This value is calculated by the function "= T.INV(0.95)". 0.95 is
used because we are interested in the upper value because the alternative hypothesis is >. The value is
1.717. The decision rule is to reject the null hypothesis if the observed test statistic is greater than 1.717.

Step 5: Input the data into Excel in one column.

Step 6: Calculate the value of the test statistic and perform a t test. To perform a t test of a single mean in
Excel, one needs to “fool” Excel by using a two-sample t test. To do this, enter the location of the single
sample observations as one of the two requested samples and enter the location of the hypothesized mean
repeated as many times as there are observations as the other sample.
Go to an empty cell in an Excel worksheet and from Insert Function (fx), select the Statistical category
and the function T.TEST. Select the observed values in Array1, select the column of hypothesized
means in Array2, 1 as the number of Tails, and 3 as the Type of two-samples with unequal variance.

445 471
489 471
474 471
505 471
553 471
477 471
454 471
463 471
466 471
557 471
502 471
449 471
The resulting p-value is 0.00478. The t value that corresponds to this probability
438 471 can be calculated by subtracting the p-value of 0.00478 from 1 and inserting the
500 471 following function into an empty cell: "= T.INV(0.99522)". The resulting t value
466 471 is 2.84 which is greater than the critical value of 1.717.
477 471
557 471 Step 7: State the statistical conclusion. The observed t value of 2.84 is greater than
the table t value of 1.717, so the business researcher rejects the null hypothesis.
433 471
She accepts the alternative hypothesis and concludes that the average size of a
545 471 U.S. farm is now more than 471 acres.
511 471
590 471 Step 8. Make a managerial decision. Agribusiness researchers can speculate about
561 471 what it means to have
larger farms. If the average size of a farm has increased from 471 acres to almost
560 471
500 acres, it may represent a substantive increase. It could mean that small farms
are not financially viable.

NOTE: Another way to solve this problem is by going to the Data tab and using the Data Analysis
menu, select t-Test: Two-Sample Assuming Unequal Variances from the dialog box. Enter the location
of the observations from the single sample of data in Variable 1 Range:. Enter the location of the
repeated hypothesized mean values in Variable 2 Range:. Enter the value of zero in Hypothesized Mean
94 Statistical Inference: Hypothesis Testing for Single Proportions

Difference. Check Labels if you have labels. Select Alpha. The output includes the observed t value, p-
values for both one- and two-tailed tests, and critical t values for both one- and two-tailed tests. The
output is shown below:

9.4 Testing Hypotheses About a Proportion

Data analysis used in business decision making often contains proportions to describe such aspects as
market share, consumer makeup, quality defects, on-time delivery rate, profitable stocks, and others.
Business surveys often produce information expressed in proportion form, such as 0.45 of all businesses
offer flexible hours to employees or 0.88 of all businesses have Web sites. Business researchers conduct
hypothesis tests about such proportions to determine whether they have changed in some way.

The formula for a z test of a population proportion is:

where = sample proportion, p = population proportion, and q = 1 − p.

Demonstration Problem 9.3

A survey of the morning beverage market shows that the primary breakfast beverage for 17% of
Americans is milk. A milk producer in Wisconsin, where milk is plentiful, believes the figure is higher
for Wisconsin. To test this idea, she contacts a random sample of 550 Wisconsin residents and asks which
primary beverage they consumed for breakfast that day. Suppose 115 replied that milk was the primary
beverage. Using a level of significance of .05, test the idea that the milk figure is higher for Wisconsin.

Step 1. State the hypotheses: H0: p = 0.17; Ha: p > 0.17

Step 2: Determine the appropriate test. The z test statistic for a proportion is defined above.

Step 3: α = 0.05.
Chapter 9 95

Step 4: This test is a one-tailed test and the observed test statistic must be less than the critical z value to
be rejected. This value is calculated by the function " =NORM.S.INV(0.95)". 0.95 is used because we are
interested in the upper value because the alternative hypothesis is >. The value is 1.645. The decision rule
is to reject the null hypothesis if the observed test statistic is greater than 1.645.

Step 5: The sample proportion is calculated by .

Step 6: There is no direct way to calculate the z value for proportions. You can input a formula into a cell
to calculate the value: "=(0.209-0.17)/SQRT(0.17*0.83/550)". The result is 2.435.

Step 7. Because is z = 2.435 is beyond critical z value of 1.645 in the rejection region, the milk producer
rejects the null hypothesis. The probability of obtaining a z > 2.435 by chance is 0.0073 (you can
determine this by inputting the function "=1-NORM.S.DIST(2.435,1)". Because this probability is less
than α = 0.05, the null hypothesis is also rejected with the p-value. On the basis of the random sample, the
producer is ready to conclude that the proportion of Wisconsin residents who drink milk as the primary
beverage for breakfast is higher than the national proportion.

Step 8. If the proportion of residents who drink milk for breakfast is higher in Wisconsin than in other
parts of the United States, milk producers might have a market opportunity in Wisconsin that is not
available in other parts of the country. Is a proportion of almost 0.21 really a substantive increase over
0.17? Certainly in a market of any size at all, an increase of almost 4% of the market share could be worth
millions of dollars and in such a case, would be substantive.

9.5 Testing Hypotheses About a Variance

At times a researcher needs to test hypotheses about a population variance. The procedure for testing
hypotheses about a population variance is similar to the techniques presented in Chapter 8 for estimating
a population variance from the sample variance. Formula 9.5 used to conduct these tests assumes a
normally distributed population.

where df = n − 1

Demonstration Problem 9.4

A small business has 37 employees. Because of the uncertain demand for its product, the company
usually pays overtime on any given week. The company assumed that about 50 total hours of overtime
per week is required and that the variance on this figure is about 25. Company officials want to know
whether the variance of overtime hours has changed. Given here is a sample of 16 weeks of overtime data
(in hours per week). Assume hours of overtime are normally distributed. Use these data to test the null
hypothesis that the variance of overtime data is 25. Let
57 56 52 44 46 53 44 44 48 51 55 48 63 53 51 50

Step 1. State the hypotheses: H0: σ2 = 25; Ha: σ2 25

96 Statistical Inference: Hypothesis Testing for Single Proportions

Step 2: Determine the appropriate test. The z test statistic for a proportion is defined as: .

Step 3: This test is two tailed, α = 0.10 is split as α/2 = 0.05.

Step 4: The degrees of freedom are 16 − 1 = 15. The two critical chi-square values can be calculated by
the function "=CHISQ.INV(0.95,15)" which results in the higher value of 24.9958, and by the function
"=CHISQ.INV(0.05,15)" which results in the lower value of 7.2609. The decision rule is to reject the null
hypothesis if the observed value of the test statistic is < 7.2609 or > 24.9958.

Step 5: The data is listed and can be input into Excel in a single column.

Step 6: The sample variance is s2 = 28.0625. The observed chi-square value is calculated by inputting a
formula into a cell: "=(15*28.0625)/25" The result is 16.8375.

STEP 7. The observed chi-square is in the nonrejection region because 16.85 is within the two values of
7.26 and 25.00. The company fails to reject the null hypothesis.

STEP 8. This result indicates to the company managers that the variance of weekly overtime hours is
about what they expected.

9.6 Solving for Type II Errors

If a researcher reaches the statistical conclusion to fail to reject the null hypothesis, he makes either a
correct decision or a Type II error. If the null hypothesis is true, the researcher makes a correct decision.
If the null hypothesis is false, then the result is a Type II error.

There is no direct way to solve for Type II errors in Excel. You can solve for values using formulas but
not specifically to the Type II errors.

Demonstration Problem 9.5

Recompute the probability of committing a Type II error for the soft drink example if the alternative
mean is 11.96 ounces.

1. The hypotheses stay the same. We solve for the critical value by first solving for the critical z value

and its associate x value. .

2. From this observed value of z, the researcher does not reject the null hypothesis because the critical
value is -1.645. However, the distribution for the new z-value with µ1 = 11.96 ounces, to the right of
Chapter 9 97

11.979 is which corresponds to a probability of committing a Type II error is

only .0708.

Operating Characteristic and Power Curves

Because the probability of committing a Type II error changes for each different value of the alternative
parameter, it is best in managerial decision making to examine a series of possible alternative values. The
text shows Power Curves generated in Excel. You can create one of these curves by solving for the Type
II error (β) for several different possible alternative means for the soft drink example discussed in
Demonstration Problem 9.5, in which the null hypothesis was H 0: = 12 ounces and α = 0.05. As
previously mentioned, power is the probability of rejecting the null hypothesis when it is false and
represents the correct decision of selecting the alternative hypothesis when it is true. Power is equal to 1 −
β. Note that Table 9.2 also contains the power values for the alternative means and that the and power
probabilities sum to 1 in each case.

The table is filled in with different values of the alternative mean. The previous example calculated β for
µa = 11.96 as 0.07 and Power as 1 − β = 0.93 The process can be repeated for several µa values. A
scatterplot can be created by plotting the power values against the alternative mean values. (input the data
into 2 columns, select the columns and insert a scatterplot with a curved line, delete the legend and the
gridlines).

SUMMARY OF EXCEL COMMANDS USED IN CHAPTER 7

Excel has limited capability for conducting hypothesis testing with single samples. By piecing together
various Excel commands, it is possible to compute a z test of a single population mean and a t test of a
single population mean.

Conduct a z Test
To conduct a z test of a single population mean, begin with the Insert Function (fx). To access the Insert Function,
98 Statistical Inference: Hypothesis Testing for Single Proportions

go to the Formulas tab on an Excel worksheet. The Insert Function is on the far left of the menu bar. In the Insert
Function dialog box at the top, there is a pulldown menu where it says Or select a category. From the pulldown
menu associated with this command, select Statistical. Select ZTEST from the Insert Function’s Statistical menu.

 In the ZTEST dialog box, place the location of the observed values in Array.
 Place the hypothesized value of the mean in X.
 Record the value of the population standard deviation in Sigma.
 The output is the right-tailed p-value for the test statistic. If the z value is negative, subtract 1-Excel output
to obtain the p-value for the left tail.
Conduct a t Test
To perform a t test of a single mean in Excel, one needs to “fool” Excel by using a two-sample t test. To do this,
enter the location of the single sample observations as one of the two requested samples and enter the location of the
hypothesized mean repeated as many times as there are observations as the other sample.

Begin this t test by selecting the Data tab on the Excel worksheet. From the Analysis panel at the right top of the
Data tab worksheet, click on Data Analysis. If your Excel worksheet does not show the Data Analysis option, then
you can load it as an add-in following directions given in Chapter 2.
 From the Data Analysis pulldown menu, select t-Test: Two-Sample Assuming Unequal Variances from
the dialog box.
 Enter the location of the observations from the single sample of data in Variable 1 Range:.
 Enter the location of the repeated hypothesized mean values in Variable 2 Range:.
 Enter the value of zero in Hypothesized Mean Difference. Check Labels if you have labels.
 Select Alpha. The output includes the observed t value, p-values for both one- and two-tailed tests, and
critical t values for both one- and two-tailed tests.