I Scream,You Scream, We
All Scream for Ice Cream!
Sarah Beeson, Jill Hall, Sarah Regan
Cold Stone Creamery




Founded in 1988 in
Tempe, Arizona
Uses in-store made ice
cream that is
combined with mix-ins,
candy, and other
toppings to create a
new flavor
Open year round
Ice cream parlor chain
Maggie Moo’s




Founded in 1989 in
Kansas City, Kansas
136 Treateries in 43
US states, Singapore,
and Puerto Rico
Specializes in ice
cream and other
desserts
Independently owned
and operated franchise
stores
Our Topic
We set out to compare Cold Stone and
Maggie Moo’s
 We went to two Cold Stone’s
(Warrington and Willow Grove) and two
Maggie Moo’s’ (Warrington and Landsale)
 Observed and surveyed for subject’s
gender, order, price paid, size, preference,
and frequency visited

Procedure
Observed customers in two of each store
from May 27- 31
 Allowed customer to order and surveyed
every third customer to ensure
randomness
 Used data table to record information

Exploratory Data Analysis
Bar Graph of Frequency of Visits

Seasonally is the most
common
◦ With Sometimes close
behind

Frequently- 11.67%
Sometimes- 31.67%
Seasonally- 36.67%
Almost Never- 20%
Frequently is the least
common
◦ Possibly due to ice cream
being seen as a seasonal
treat
Gender and Flavor
With the exception of
coffee and chocolate,
flavors are fairly evenly
distributed between
gender
 Conclude gender has
no effect on flavor

Cake Batter- 50% female, 50% male
Chocolate- 70% female, 30% male
Coffee- 80% female, 20% male
Mint- 55.56% female, 44.44% male
Specialty- 62.5% female, 37.5% male
Strawberry- 40% female, 60% male
Vanilla- 54.54% female, 45.46% male
Store Preference
Cold Stone and Maggie
Moo’s preference is
approximately evenly
distributed between the
two
 We conclude that the
population is evenly
split on favorite store

Cold Stone- 51.67%
Maggie Moo’s- 48.33%
Size

Small and Medium are
approximately evenly
distributed
◦ Large has a much smaller
frequency

Small- 40%
Medium- 38.33%
Large- 21.67%
We conclude that the
majority of the
population orders small
or medium
Price


Histogram is right skewed. It spreads from
$3.25 to $6.75. The center is $4.81.
There are gaps between
$3.50 and slightly less
than $4.25, slightly above
$4.25 and $4.50, and
between approximately
$5.12 and $5.30.
We conclude that the
average price paid is
around $4.80
◦ Size and amount of toppings
affect price
Gender and Preference
Preference for store is
approximately evenly
distributed by each
gender
 We conclude that
gender and preference
are not dependent

Cold Stone: male- 38.71%
female- 61.29%
Maggie Moo’s: male- 41.38%
female- 58.62%
Data Analysis
GOF test for variety of flavors
 χ2 test of homogeneity to compare size
 χ2 test of independence for gender vs.
store preferred
 χ2
x2 GOF Test for Variety of Flavors

Hypotheses
◦ Ho: Observed distribution of flavors fits the expected
distribution of flavors
◦ Ha: Observed distribution of flavors does not fit the
expected distribution of flavors
Flavor
Frequency
Flavor
Frequency
cake
batter
chocolate coffee
4
10
mint
5
cake batter chocolate coffee
8.5714
8.5714
8.5714
specialty strawberry vanilla
9
mint
8.5714
16
5
total
11
specialty strawberry vanilla
8.5714
8.5714
8.5714
60
total
60
x2 GOF Test for Variety of
Flavors
State
Check
Categorical data
2. SRS
3. All expected counts ≥ 5
1.
1.
2.
3.
Flavor and frequency are
both categorical
Assumed representative
8.5714 ≥ 5
Conditions metx2 distribution x2 GOF test
x2 GOF Test for Variety of Flavors
2
2
2
(
obs

exp)
(
4

8
.
5714
)
(
10

8
.
5714
)
2
 


 ...
exp
8.5714
8.5714
 x2= 12.8
 p= .0463
 df=59
 We reject the Ho because the p-value of .0464 is
less than alpha=.05
 We have sufficient evidence that the observed
distribution of flavors does not fit the expected
distribution of flavors
x2 Test for Homogeneity

Hypotheses
◦ The distribution of sizes at Maggie Moo’s and Cold
Stone is equal
for Independence
distribution
of sizes at MaggieTest
Moo’s
and Cold
Test ◦of The
Combined
Data
First attribute
(categorical):
Size
Stone
is not equal
Second attribute (categorical): Store_At
Size
large
Store_At
medium
Row
Summary
small
cold stone
1 (6.5)
6 (11.5)
23 (12.0)
30
maggie moo's
12 (6.5)
17 (11.5)
1 (12.0)
30
Column Summary
First attribute: Size
Number of categories: 3
Second attribute: Store_At
13
23
24
60
2
x Test
for Homogeneity
State
1.
2.
3.
Check
Categorical Data
SRS
All expected counts ≥ 5
Size and the store at are
both categorical
2. Assumed representative
3. Counts are all above 5
1.
Conditions metx2 distribution x2 test for
homogeneity
x2 Test for Homogeneity
2
2
2
(
obs

exp)
(
1

6
.
5
)
(
6

11
.
5
)
2
 


 ...
exp
6.5
11.5
 x2= 34.74
 p= 2.8596x10-8
 df= 2
 We reject the Ho because the p-value of
2.8596x10-8 is less that alpha=.05.
 We have sufficient evidence that the distribution
of sizes ordered at Cold Stone and Maggie Moo’s
is not equal.
x2 Test for Independence

Hypotheses
◦ Gender and preference are independent of each
other
◦ Gender and preference are Test
dependent
on each
for
Independence
Test
of Combined Data
other
First attribute (categorical): Preference
Second attribute (categorical): Gender
Preference
cold stone
Gender
maggie moo's
Row
Summary
female
19 (18.6)
17 (17.4)
36
male
12 (12.4)
12 (11.6)
24
Column Summary
31
29
60
x2 Test for Independence
State
1.
2.
3.
Check
Categorical Data
SRS
All expected counts ≥ 5
1.
2.
3.
Gender and preference are
categorical
Assumed representative
Counts are all above 5
Conditions metx2 distribution x2 test for
independence
x2 Test for Independence
2
2
2
(
obs

exp)
(
19

18
.
6
)
(
17

17
.
4
)
2  


 ...
exp
18.6
17.4
x2= .04449
p= .83
df= 1
We fail to reject the Ho because the p-value of
.83 is greater than alpha=.05.
 We have sufficient evidence that gender and
preference for Cold Stone and Maggie Moo’s are
dependent variables.




Conclusions from Tests
 χ2
GOF test for variety of flavors
◦ Observed distribution of flavors does not fit
the expected distribution of flavors
 χ2 test
of homogeneity to compare size
◦ Distribution of sizes ordered at Cold Stone
and Maggie Moo’s is not equal
test of independence for gender vs.
store preferred
 χ2
◦ Gender and preference for Cold Stone and
Maggie Moo’s are dependent variables
Bias and Error

Undercoverage
◦ Only surveyed 4 stores in the area; other
areas may be different

Time of day/ year
◦ Different times of day would get different
responses

Number of toppings
◦ Large numbers of toppings affect price
Personal Opinions and Conclusions

Preference for store was basically evenly
distributed between subjects
◦ Preference broken down by gender was also
even

Project took a lot of time and effort
◦ More time would have allowed for better data
collection