Author: Michael Lo, Assistant Chair of the Department of Mathematics and Statistics

Practice Questions
1. Are male and female shoppers different regarding reading labels of packaged food? A random
sample of 100 shoppers was drawn in a nearby supermarket during a three-day period. They
were asked about their gender and if they regularly read the labels in packaged food or not. In
particular, they were asked about the gender in the following way:
I identify my gender as: ____________________ (Please fill in the blank).
All results are summarized in the following two-way table.
Gender
Male Female Other
Yes 20 15 1 36
Reading
Sometimes 5 2 1 8
Labels?
Not At All 35 21 0 56
60 38 2 100
a) Find expected frequency of the cell “Other” under Gender and “Sometimes” under
Reading Labels. Provide an interpretation of the value. [2 marks]
b) Notice that the calculated expected frequency is under 5 and we typically cannot
proceed with chi-square analysis with this. Suggest a way to make sure all expected
frequencies are bigger than 5 and be able to proceed. [3 marks]

2. Are male and female shoppers different regarding reading labels of packaged food? A random
sample of 100 shoppers was drawn in a nearby supermarket during a three-day period. They
were asked about their gender and if they regularly read the labels in packaged food or not.
Results are summarized in the following two-way table.
Gender
Male Female
Reading Yes 25 19 44
Labels No 35 21 56
60 40 100
a) Find the percentage of shoppers who are both female and reading labels regularly. [1 mark]
b) Find the percentage of male shoppers who read labels regularly. [1 mark]
c) Find the percentage of female shoppers who read labels regularly. [1 mark]
d) Among all shoppers who do not read labels regularly, find the percentage of them who are
male shoppers. [1 mark]
e) Draw a side-by-side bar graph with Gender on the x-axis. Provide a description of it. [2+1
marks]
f) Find the expected frequency of the cell “Male and Yes”. Provide an interpretation of the
value. [2 marks]
g) Can we conclude that male and female shoppers are different, in terms of reading labels
from the packaged food regularly? Make sure you use the three-step process and an
appropriate decision point to answer this question. [2+2+2 marks]
h) Now that we have the following conclusion. What it really means? In other words, rewrite
the conclusion in layperson’s term. [2 marks]
3. Suppose we have a similar setting but the following two-way table is from a different sample.
Gender
Men Women
Reading Yes 24 16 40
Labels? No 36 24 60
60 40 100
a) Find the percentage of male shoppers who read labels regularly. [1 mark]
b) Find the percentage of female shoppers who read labels regularly. [1 mark]
c) Draw a side-by-side bar graph, with the Reading Labels on the x-axis. [2 marks]
d) Calculate the c statistic. [2 marks]
2

e) What conclusion could you generalize from the answers above.

4. Suppose we have a similar setting but the following two-way table is from a different sample.
Gender
Men Women
Reading Yes 6 36 42
Labels? No 54 4 58
60 40 100
a) Find the percentage of male shoppers who read labels regularly. [1 mark]
b) Find the percentage of female shoppers who read labels regularly. [1 mark]
c) Draw a side-by-side bar graph, with the Reading Labels on the x-axis. [2 marks]
d) Calculate the c statistic. [2 marks]
2

e) What conclusion could you generalize from the answers above.

Answer to Practice Questions

1. Are male and female shoppers different regarding reading labels of packaged food? A random
sample of 100 shoppers was drawn in a nearby supermarket during a three-day period. They
were asked about their gender and if they regularly read the labels in packaged food or not. In
particular, they were asked about the gender in the following way:
I identify my gender as: ____________________ (Please fill in the blank).
All results are summarized in the following two-way table.
Gender
Male Female Other
Yes 20 15 1 36
Reading
Sometimes 5 2 1 8
Labels?
Not At All 35 21 0 56
60 38 2 100
a) Find expected frequency of the cell “Other” under Gender and “Sometimes” under
Reading Labels. Provide an interpretation of the value. [2 marks]
2´8
e= = 0.16
100
 This value means that there are about 0.16 shoppers who are categorized as Other in
Gender and only read labels “sometimes”, when the Gender and Reading Labels are
assumed to be independent.

b) Notice that the calculated expected frequency is under 5 and we typically cannot
proceed with chi-square analysis with this. Suggest a way to make sure all expected
frequencies are bigger than 5 and be able to proceed. [3 marks]
We will need to combine adjacent columns and/or rows to make sure all expected
frequencies bigger than 5. We usually start with the lowest value in either row totals
or column totals, i.e. “2” under “Other”.
Gender
Male Female Other
Yes 20 15 1 36
Reading
Sometimes 5 2 1 8
Labels?
Not At All 35 21 0 56
60 38 2 100
Since Gender is in nominal scale, it doesn’t matter if “Other” is grouped to “Male” or
“Female”. I’ll just group “Other” with “Female”, for convenience sake.
Gender
Male Female/Other
Yes 20 16 36
Reading
Sometimes 5 3 8
Labels?
Not At All 35 21 56
60 40 100
But you’ll notice right away that the cell “Female/Other and Sometimes” still have an expected
frequency under 5 (40*8/100=3.2). Therefore, we will need to combine an adjacent row. Here, we
have a choice to combine “Sometimes” to either “Yes” or “Not at all”. We typically would pick the
one with smallest row total (here it’s 36 with Yes).
Gender
Male Female/Other
Reading Yes/Sometimes 25 19 44
Labels? Not At All 35 21 56
60 40 100
Therefore, we wind up with a 2x2 table (with 2 rows and 2 columns).

2. Are male and female shoppers different regarding reading labels of packaged food? A random
sample of 100 shoppers was drawn in a nearby supermarket during a three-day period. They
were asked about their gender and if they regularly read the labels in packaged food or not.
Results are summarized in the following two-way table.
Gender
Male Female
Reading Yes 25 19 44
Labels No 35 21 56
60 40 100
 a) Find the percentage of shoppers who are both female and reading labels regularly. [1 mark]
Answer: 19/100*100% = 19%

b) Find the percentage of male shoppers who read labels regularly. [1 mark]
Answer: 25/60*100% = 41.7%

c) Find the percentage of female shoppers who read labels regularly. [1 mark]
Answer: 19/40*100% = 47.5%

d) Among all shoppers who do not read labels regularly, find the percentage of them who are
male shoppers. [1 mark]
Answer: 35/56*100% = 62.5%

e) Draw a side-by-side bar graph with Gender on the x-axis. Provide a description of it. [2+1
marks]

The two groups of bars do look alike.

f) Find the expected frequency of the cell “Male and Yes”. Provide an interpretation of the
value. [2 marks]
44 ´ 60
e= = 26.4 . It is the number of shoppers (from the random sample of 100
100
shoppers) who are male and reading labels regularly, when it is assumed that the two
variables Gender and Reading Labels are independent to each other.

g) Can we conclude that male and female shoppers are different, in terms of reading labels
from the packaged food regularly? Make sure you use the three-step process and an
appropriate decision point to answer this question. [2+2+2 marks]

Null hypothesis: Gender and Reading Labels are independent to each other among all
shoppers.
Alternative hypothesis: Gender and Reading Labels are not independent to each other among
all shoppers.

(25 - 26.4)2 (19 - 17.6)2 (35 - 33.6)2 (21 - 22.4)2

c2 = + + + = 0.07 + 0.11 + 0.06 + 0.09 = 0.33
26.4 17.6 33.6 22.4
DF = (r-1)(c-1) = (2-1)(2-1) = 1 à DP = 3.84
 Since the chi-square statistic is less than the decision point, we do not have enough
evidence to reject the null hypothesis and conclude that Gender and Reading Labels are
significantly independent to each other among all shoppers.

h) Now that we have the following conclusion. What it really means? In other words, rewrite
the conclusion in layperson’s term. [2 marks]
Note: Make sure you understand that the word “independent” or “not independent” is too
technical and most people have no idea what it really means. Our job is to make sure
everyone reading this conclusion can understand.
And the phrase “in layperson’s term” refers to the situation where a very technical piece of
writing is rewritten in such a way that everybody is able to understand.
Here, we could say that “male and female shoppers are not (significantly) different, when it
comes to reading labels of pre-packaged food.

3. Suppose we have a similar setting but the following two-way table is from a different sample.
Gender
Men Women
Reading Yes 24 16 40
Labels? No 36 24 60
60 40 100
a) Find the percentage of male shoppers who read labels regularly. [1 mark]
Answer: 24/60*100% = 40%

b) Find the percentage of female shoppers who read labels regularly. [1 mark]
Answer: 16/40*100% = 40%

c) Draw a side-by-side bar graph, with the Reading Labels on the x-axis. [2 marks]

d) Calculate the c2 statistic. [2 marks]

c =0
2
(Please let me know if you have trouble doing this calculation.)

e) What conclusion could you generalize from the answers above.

The following four statements are equivalent:
 1) Percentages from the subgroups of independent variables are similar (from the two-way
table).
2) Groups of bars look alike (from the side-by-side bar graph).
3) Relatively small chi-square statistic.
4) Two variables are independent.
4. Suppose we have a similar setting but the following two-way table is from a different sample.
Gender
Men Women
Reading Yes 6 36 42
Labels? No 54 4 58
60 40 100
a) Find the percentage of male shoppers who read labels regularly. [1 mark]
Answer: 6/60*100% = 10%

b) Find the percentage of female shoppers who read labels regularly. [1 mark]
Answer: 36/40*100% = 90%

c) Draw a side-by-side bar graph, with the Reading Labels on the x-axis. [2 marks]

d) Calculate the c2 statistic. [2 marks]

c = 63.05
2
(Please let me know if you have trouble doing this calculation.)

e) What conclusion could you generalize from the answers above.

The following four statements are equivalent:
1) Percentages from the subgroups of independent variables are different (from the two-
way table).
2) Groups of bars do not look alike (from the side-by-side bar graph).
3) Relatively big chi-square statistic.
4) Two variables are not independent.