You are on page 1of 9

MBA-622 Marketing Research

VIU Fall 2009

Summary of chapter 13

Hypothesis testing

Submitted to:
Dr. Michael Keith Simms

Submitted by:
Laila Yassine

1
MBA-622 Marketing Research
VIU Fall 2009

Synopsis:
This chapter entitled hypothesis testing, describes the general procedure
involved in hypothesis testing, discusses the role of hypothesis testing in data
analysis, and outlines several hypothesis tests frequently encountered in
marketing.
First part distinguishes between descriptive and inferential analysis. Descriptive
analysis consists of computing measures of central tendency and dispersion, as
well as constructing one-way tables; it helps the researcher summarize the
general nature of the study variables (p401). Inferential analysis is data analysis
aimed at testing specific hypotheses (p402).
Second part states the null and alternative hypotheses pertaining to a variety of
decision situations requiring formal hypothesis testing.
We will use the symbols H0 and Ha, to denote the null and alternative
hypotheses, respectively. The null and alternative hypotheses complement each
other; they are mutually exclusive and collectively exhaustive. In other word, they
are stated in such a way that Ha will be not accepted if the sample evidence
strongly supports H0. Similarly, Ha will be accepted if the sample evidence is
strong enough to reject H0.
Third section in this chapter defines type I and type II errors and state the
relationship between them, and define the significance level and power of
hypothesis test.
Two kinds of errors are possible in testing hypotheses about population
characteristics based on sample data: a type I error, which refers to rejecting the
null hypothesis when it is actually true (p404), and a type II error, which refers to
failing to reject the null hypothesis when it is false. Traditional hypothesis-testing
procedures place an upper bound on the probability of a type I error called the
significance level α, but do not explicitly control type II error.

2
MBA-622 Marketing Research
VIU Fall 2009
When α is lowered, however, the probability of a type II error (called β) will
increase, and power (1- β), which represents protection against a type II error,
will decrease.
The general procedure for testing hypotheses consists of the following steps
(p409):
Step 1: set up H0 and Ha
Step 2: identify the nature of the sampling distribution curve and specify the
appropriate test statistic.
Step 3: determine whether the hypothesis test is one-tailed or two-tailed.
Step4: taking into account the specified significance level, determine the critical
value (two critical values for a two-tailed test) for the test statistic from the
appropriate statistical table.
Step 5: state the decision rule for rejecting H0
Step 6: compute the value for the test statistic from the sample data.
Step 7: using the decision rule specified in step 5, reject either H0 or Ha.
Next section is the role of hypothesis testing in data analysis. Two factors are
crucial in choosing an appropriate analysis procedure: the number of variables to
be analyzed and the nature of the data collected on each variable. Analysis
procedure are classified as being univariate analysis which is appropriate when
just one variable is the focus of the analysis (p410), and multivariate analysis
which is appropriate when two or more variables are to be analyzed
simultaneously (p410).
Nonparametric tests are used when the data are non metric (nominal or ordinal),
and parametric procedures are analysis techniques suitable for metric data
(interval and ratio data).
Last part in this chapter discussed several hypothesis tests, the first (the Chi-
square), and the rest are parametric.
The objective of most research studies includes an examination of relationships
among key variables. Two-way tabulation is a useful preliminary step in
understanding the nature of the association between a pair of variables.

3
MBA-622 Marketing Research
VIU Fall 2009

For two-way tabulation to be meaningful, the data on each variable must be


coded into a fixed set of categories, and the number of categories should not be
large.
The specific hypothesis tests are:
1- Cross-tabulations: Chi-square contingency test: is widely used for
determining whether there is a statistically significant relation between two
categorical (nominal or ordinal) variables.
2- The z-test for a single mean when the sample size is large. To test
hypotheses that compares the population mean of variable to a
prespecified value.
3- The t-test for a single mean when the sample size is small.
4- The z-test for a single proportion when the sample size is large. To test
hypotheses that compare population proportion of a variable to a
prespecified value.
5- The z-test for two means when both samples are large.
6- The t-test for two means when either sample is small.
7- The t-test for two means when the samples are dependent. To test
hypotheses that compare two population means of the same variable
when the data for the test are collected from the same set of sample units.
8- The z-test for two proportions when both samples are large. To test
hypotheses that compares the population proportion of variable for two
separate populations.

The article:

4
MBA-622 Marketing Research
VIU Fall 2009
"Black Friday" deals may not signal retail comeback
Sun Nov 22, 2009 2:00pm EST

By Jessica Wohl
CHICAGO (Reuters) - When the U.S. holiday shopping season kicks off on the day
after Thanksgiving, retailers can expect to see millions of less frightened, but even
more bargain-hungry customers cross their thresholds.

Industry experts expect a strong turnout on Black Friday, which falls on November
27 this year, as deep discounts lure shoppers after more than a year of subdued
spending. But they caution it will not mean a bumper holiday season in the weeks
leading up to Christmas since consumers still remain cautious.

"Given what we know about consumer shopping patterns, even this month, I would
suspect it will turn out to be a very strong performance," said Michael Niemira, chief
economist of the International Council of Shopping Centers.

Special promotion days have been big drivers of sales, he said, pointing to the lift
retailers saw on the November 11 Veteran's Day holiday.

Retailers and websites dedicated to Black Friday deals have leaked sales plans earlier
than usual, in the hopes of sparking demand for flat-panel televisions, toys and other
goods after 2008's worst holiday season in decades.

While the economy remains weak and unemployment has risen, U.S. shoppers have
had more than a year to adjust their spending and digest the bad news. In 2008,
holiday shopping started just weeks after the global financial crisis erupted.

"Certainly last year was a year of tremendous uncertainty going into Black Friday
because we were right in the middle of the storm," said Chris Donnelly, a partner in
Accenture's retail practice. "There is much less panic, I would say, or much less
uncertainty, as we go into the season."

Even so, more than 172 million shoppers visited stores and websites from
Thanksgiving Day through Sunday last year, up from 147 million in 2007, according
to the National Retail Federation. The average amount of money spent by shoppers
over that weekend rose 7.2 percent to $372.57 per person.

Those numbers, however, did not prevent a sales slide of 3.4 percent for the entire
shopping season last year, marking the first decline since the NRF began tracking
such data.

5
MBA-622 Marketing Research
VIU Fall 2009
While the NRF has not issued a Black Friday forecast, it expects 2009 holiday season
sales to decline 1 percent. The ICSC forecast a 1 percent to 2 percent rise.

"Retail sales have been, while not stellar, somewhat stabilizing over the past few
months and there is every reason to believe that as we go into the holiday season
that we are going to see some stability as well," Donnelly said.

BARGAIN FRIDAY

The term "Black Friday" is said to have originated in Philadelphia during the 1960s to
describe the difficulty of police and drivers to deal with exceptionally heavy traffic on
that day as shoppers flooded the city's commercial center.

The phrase was later co-opted by retailers to refer to the holiday shopping period as
a time of year when their business moves into the black, or turns a profit.

Niemira, for one, refers to Black Friday as "Bargain Friday" since it is known for
deals. "If Black Friday is flattish to slightly positive, that would be encouraging, but I
don't think we'd be ready to kind of write the story of the season yet," Donnelly said.

Sixty-one percent of chief marketing officers at leading U.S. retailers surveyed by


BDO Seidman expect Black Friday sales to be flat, while 33 percent predicted an
increase.

One factor that could help spur consumer appetite is a prediction for good weather,
with mild temperatures and slim chance of precipitation, though the Pacific
Northwest could see strong storms, according to tracking firm Planalytics.

Market research firm IBISWorld expects total retail sales over Black Friday weekend
to rise 2.8 percent to $42.9 billion. It expects 76.9 million people to swarm into retail
stores on Black Friday alone.

ShopperTrak, which measures customer traffic, expects Black Friday to again be the
busiest day in stores after accounting for 6 percent of traffic in the 2008 holiday
season.

Target Corp (TGT.N), Best Buy (BBY.N) and others are opening at 5 a.m. on
November 27, while Chelsea Premium Outlets locations are opening as early as 9
p.m. on Thanksgiving Day.

Wal-Mart Stores Inc (WMT.N) is keeping most of its discount stores open for 24
hours a day to help control crowds after a worker was trampled to death at one of its
stores during last year's Black Friday rush.

SEARCHING FOR DISCOUNTS

After last winter's bevy of deep discounts, many consumers will only open their
wallets if they spot a bargain. Almost 70 percent of consumers surveyed by
America's Research Group said they wanted to see discounts of at least 50 percent
before they would buy something for the holidays.

6
MBA-622 Marketing Research
VIU Fall 2009
"The range of the promotions will be different from what we've seen in recent years,
because it has to be," said NRF spokeswoman Kathy Grannis. "Retailers know that
consumers are so bargain focused and bargain conscious."

BDO Seidman's poll found that 96 percent of retailers planned to increase


promotions and discounts this year.

"They are going to have to be creative" said Ted Vaughan, a partner in BDO
Seidman's retail and consumer product practice.

Of course, fewer retailers are vying for that business. According to Bain & Co, 27
retailers went bankrupt in 2008 and 18 more have done so to date in 2009.
Together, those chains used to account for about $25 billion to $30 billion in sales.

Earlier holiday-themed sales expanded the season this year and may make it easier
for consumers who have less credit available, Vaughan said. Sears, Kmart and Toys
R Us have also been touting layaway plans to help shoppers spread out spending.

Consumers flocking to stores may not be buying gifts. A Consumer Reports survey
found that 66 percent of shoppers heading out over the weekend will be shopping for
themselves.

CONNECTION

7
MBA-622 Marketing Research
VIU Fall 2009

The article provided a clear example of hypothesis and the use inferential
data analysis in drawing conclusion based on the hypothesis. Z-test was used to
test data collected because the writer used a single proportion of the population
to test the hypothesis.

The writer hypothesizes that “the Black Friday” deals may not signal retail
come back. In order to prove that this null hypothesis is true, the writer compared
data from 2007 black Friday and 2008 black Friday. In 2008 when the financial
crisis erupted and unemployment skyrocketed, more than 172 million shoppers
visited stores and websites from Thanksgiving Day through Sunday last year, up
from 147 million in 2007, according to the National Retail Federation. The
average amount of money spent by shoppers over that weekend rose 7.2
percent to $372.57 per person. Those numbers, however, did not prevent a sales
slide of 3.4 percent for the entire shopping season last year.

The writer used the above argument to prove that retail sales in 2009 will also
drop because the economy remains weak and unemployment has risen, and the
postulated that the retail sale in Black Friday 2009 will decline by 1 percent. In
order to further test her hypothesis that the 2009 retail sale of black Friday will
decline and does not signal a market comeback, the writer used Z-test as she
collected data from major marketing officers in the United States. The result of
data analysis shows that 63% of the officers predicted a decline in the retail sale
on Black Friday, while 33% predicted an increase in the retail sale on Black
Friday 2009. Based on the result figure, the writer proved that her hypothesis that
black Friday deals may not signal retail comeback was proved to be a true
hypothesis.

8
MBA-622 Marketing Research
VIU Fall 2009
References:

Parasuram.A, Grewal, and D.Krishnan. Marketing Research.


New York: Houghton Mifflin Company, 2007.

Jessica Wohl “black Friday” deals may not signal retail


comeback

CHICAGO (Reuters)
November 22, 2009 |