BUSINESS
THE HOLLAND SENTINEL
Wednesday, December 8, 2004
Gregordog:
No Girls Allowed?
Statistical study
exposes lack of gender diversity at local restaurant
HOLLAND, MI - People in American society today are well aware of gender
inequity and the potential problems associated with the issue. Female
rights enthusiasts across the country attempt to bridge gender gaps on
a daily basis, including those that may exist in the workplace and the
education system. We would like to believe that our generation is more
sympathetic to female rights than that of our grandparents.
Yet as women strive
for professional independence and recognition, their male counterparts
have quietly continued to dominate some other, more personal, aspects
of life. One such aspect is the fast food industry. This phenomenon can
clearly be seen in our own community. Gregordog, an eatery located
between Hope College and boisterous downtown Holland, was targeted by a
female college student, who asked for anonymity, for further
observation. To examine her claim that Gregordog's customer base is
predominantly male, she chose to call in the renowned experts at Grand
Rapids Extremely Accurate Testing (GREAT) Incorporated. GREAT promptly
sent its top young field statistician, Mark Dondero, to the lakeshore
to investigate.
Gregordog was founded in 2000 by Greg Vanden Berg, who sought to cater
to college students and employers rushed through their lunch hour at
work. The cheap, simple menu has helped develop a devoted following in
a relatively short time. It is, according to one satisfied male
customer, a "nice addition to the restaurants downtown."
Dondero decided to test three stereotypes of the fast food industry at
Gregordog: that a larger number of people eat fast food for lunch on a
Friday than on a Monday; that a larger proportion of males than females
eat fast food for lunch; that the average amount of food (in this case,
hot dogs) ordered by males would be larger than the average amount of
food ordered by females. Assuming that the stereotypes are incorrect,
no significant difference should be found in any of the three studies.
To test his hypotheses, Dondero opted to sit at the counter of the
eatery between 11:00 am and 1:00 pm and record the gender and order of
the customers. On Monday, November 29 and Friday, December 3 of last
week, he conducted data collection as described.
A chi-square test was utilized for analysis of the differences between
Monday and Friday. The observed number of customers, 23 for Monday and
29 for Friday, were quite close to the expected number of 26 customers
each. A P-value of 0.4054 was
obtained. This is well above the P-value
of 0.05 required to establish a significant difference. Therefore,
Dondero could not conclude that a larger number of people eat lunch at
Gregordog on a Friday than on a Monday.
A two-way table was
constructed to effectively organize the data by gender. The
proportion of male customers was 0.6538, meaning that about 65% of
Gregordog's customers during the observation period were male. The
remaining 0.3462, or about 35%, of the customer base was female.
Although this study was not specifically interested in the age of the
customers, Dondero was quick to point out that almost a third of the 18
female customers were children. In contrast, only two of the 34 male
customers were children.
A single proportion test was used to determine if a majority of
Gregordog customers are male. The test yielded a P-value of 0.018. With a P-value so far below the desired
value of 0.05, Dondero could safely conclude that a majority of the
lunch customers at Gregordog are male. In fact, he is 95% confident
that the proportion of customers that are male is between 50.9% and
78.0% on any given day.
A two-sample t-test was
performed on the hot dog data to determine if males ordered a
significantly larger number of hot dogs. The mean (which is much like
an average) number of hot dogs ordered by male customers was 1.912,
with a very small standard deviation (average approximate distance of
each data point from the mean) of 0.452. Female customers ordered only
a mean number of 1.556 hot dogs with a larger standard deviation of
0.984. The t-statistic given
by the test was 1.46, with a P-value
of 0.080. This P-value was
slightly higher than the desired 0.05, so Dondero could not conclude that the average
male customer orders more hot dogs than the average female customer.
That said, there were intricate details in the hot dog data that lead
Dondero to believe that there is, in fact, a significant trend. "One
particular outlier stands out in my mind," he explained. "A female
college student came in after class on Friday and ordered 5 hot dogs.
Until then, the maximum order I had observed was 3. This could further
help us understand why the standard deviation was over twice as large
for the females compared to the males." He also contends that another
study, one with a larger sample size than his study of 52, would
produce better results. "Just because the data did not prove the trend
doesn't mean the trend is not there," Dondero insists. "Had I had
enough time and money to continue the project for another week or two,
I'm confident I would have proved that point. Unfortunately, the data
didn't cooperate, as data so often doesn't," he finished with a
good-natured chuckle.
The Gregordog study convinced GREAT of the validity of the female
student's concerns. Female customers are indeed significantly
underrepresented at Gregordog. Gregordog employees declined to comment
on the situation, or what will be done (if anything) to help target
women. Perhaps they could take an example from Subway, who introduced a
hefty beef sandwich in several Southern and Midwestern United States
markets in 2000. The sandwich was geared towards the eating habits of
working-class men, who had been neglected by Subway's low-fat,
heart-healthy stance.
Gregordog could potentially take steps in this direction to attract
more female customers. However, for the moment, it is what GREAT
affectionately termed a "testosterone-controlled operation." As long as
hungry and hurried men seek a fast meal at an affordable price,
Gregordog will always have a place in downtown Holland.
APPENDIX
A: Data Collected
Day Gender #
hot dogs
Monday
M 2
Monday
M 2
Monday
M* 1
Monday
M 1
Monday
M 2
Monday
F 1
Monday
M 2
Monday
M 2
Monday
F 2
Monday
F 1
Monday
M 2
Monday
M 2
Monday
M 2
Monday
M 2
Monday
F 2
Monday
F 2
Monday
M 3
Monday
M 2
Monday
M 2
Monday
M 2
Monday
M 2
Monday
M 2
Monday
M 1
Friday
M 3
Friday
F 5
Friday
M 2
Friday
M 2
Friday
M 2
Friday
M 2
Friday
M 2
Friday
M 2
Friday
F 2
Friday
F 1
Friday
M 2
Friday
M 2
Friday
M 2
Friday
F 2
Friday
M 2
Friday
F 1
Friday
F* 1
Friday
F 1
Friday
F* 1
Friday
M* 1
Friday
F 1
Friday
F* 1
Friday
F 1
Friday
F* 1
Friday
F 2
Friday
M 1
Friday
M 2
Friday
M 2
Friday
M 2
* Denotes Child
APPENDIX
B: Works Referenced
Anderson, Miranda K. "Hot Dog! New restaurant fills niche downtown." Holland Sentinel, Holland, MI. Dec.
17, 2000.
<http://www.hollandsentinel.com/stories/121700/bus _Dog.html>
Thompson, Stephanie. "Subway burger targets working-class men.
Introduction of Subway Steakburgers." Advertising
Age 71 (Mar. 62000): 3, 63.
APPENDIX C: Goals, Hypotheses, and
Results
A) Goals:
What day of the week do people tend to eat lunch (11:00 - 1:00) at
Gregordog?
What gender tends to eat lunch at Gregordog? Is there a significant
difference between the average numbers of hot dogs ordered by each
gender?
B) Hypotheses:
1) Separate by day of the week
H0: There is no significant difference between the
numbers of people that eat lunch at Gregordog on a Monday and on a
Friday.
Ha: A larger number of people eat lunch at Gregordog
on a Friday than on a Monday.
2) Separate by gender
H0: There is no significant difference between the
proportion of males and the proportion of females that eat lunch at
Gregordog.
Ha:A larger proportion of males eat lunch at Gregordog
compared to the proportion of females.
3) Separate by number of hot dogs ordered
H0: There is no significant difference between the average
number of hot dogs ordered by males and the average number of hot dogs
ordered by females.
Ha: Males, on average, order a larger number of hot dogs
than females.
C) Results:
1) Chi-square test: Comparing Day
Monday Observed: 23, Friday Observed: 29
Monday Expected: 26, Friday Expected: 26
chi-square = 0.6923 df= 1
P-value = 0.4054
Conclusion: We cannot conclude
that people are more likely to eat lunch at Gregordog on a Friday than
on a Monday.
2) Two-Way Table: Comparing Gender
|
Friday
|
Monday
|
Total
|
Female
|
13
|
5
|
18
|
Male
|
16
|
18
|
34
|
Total
|
29
|
23
|
52
|
Proportion of males: 34/52 = 0.6538
3) Single Proportion Test: Are a majority of customers male?
P-value = 0.018
95% Confidence Interval: (0.509144, 0.780342)
Conclusion: We can conclude
that a larger proportion of males eat lunch at Gregordog compared to
the proportion of females.
4) 2-Sample T - Test: Comparing Hot
Dogs Ordered
Gender
|
n
|
Mean
|
St. Dev
|
SE Mean
|
Male
|
34
|
1.912
|
0.452
|
0.077
|
Female
|
18
|
1.556
|
0.984
|
0.230
|
t-statistic = 1.46; P-value = 0.080
Conclusion: We cannot conclude
that males, on average, order a larger number of hot dogs than females.