10.

1 REVIEW (for quiz) NAME: ______ANSWER KEY____

For #1, show probability notation and the answer. The answer can be in fraction or % form.
1. Use the following table about people who are in doctoral programs:
Men Women
Completed 423 98 521
Still Enrolled 134 33 167
Dropped Out 238 98 336
795 229 1024

a. Add in the totals rows/columns

b. How many rows? How many columns? 3 rows and 2 columns
c. How many variables? 2 What are they? Doctoral program status & gender
d. How many cells? 6
e. What percent of people surveyed are men? P(M) = 795/1024 = 77.64%
f. What percent of people surveyed are women? P(W) = 229/1024 = 22.36%
g. What percent of people surveyed completed the program? P(C) = 521/1024 = 50.88%
h. What percent of people surveyed are still enrolled in the program? P(E) = 167/1024 = 16.31%
i. What percent of people surveyed dropped out of the program? P(D) = 336/1024 = 32.82%
j. What percent of people surveyed are men and completed the program? P(M and C) = 423/1024
= 41.31%
k. What percent of people surveyed are women and dropped out of the program? P(W and D) =
98/1024 = 9.57%
l. What percent of those that dropped out were men? P(M|D) = 238/336 = 70.83%
m. What percent of those that dropped out were women? P(W|D) = 98/336 = 29.17%
n. What percent of those that completed the program are men? P(M|C) = 423/521 = 81.19%
o. What percent of those that completed the program are women? P(W|C) = 98/521 = 18.81%
p. What percent of men are still enrolled? P(E|M) = 134/795 = 16.86%
q. What percent of women have dropped out? P(D|W) = 98/229 = 42.79%

2. For the table in problem 1,

a. find the expected cell counts BY HAND (show work):
Men Women
(521 x 795)/1024 = (521 x 229)/1024 =
Completed 404.49 116.51 521
(795 x 167)/1024 = (167 x 229)/1024 =
Still Enrolled 129.65 37.35 167
(795 x 336)/1024 = (336 x 229)/1024 =
Dropped Out 260.86 75.14 336
795 229 1024
 423  404.49   98  75.14 
2 2

b. What the test statistic?  

2
2  ...   13.398
404.49 75.14
c. What is the P-Value? P  2  13.398   0.0012
2

d. Is there sufficient evidence at the 5% level? Yes. Since the P-Value is less than 0.05
e. Is there sufficient evidence at the 1% level? Yes. Since the P-Value is less than 0.01

1
3. A large value of  2 tells us what about the expected and observed counts? What does a small value
tell us?
A large value of χ2 tells us the expected and observed counts are very different.
A small value means that the expected and observed counts are very similar.

4. Draw a  2 distribution and describe the shape.

They are all skewed to the right.

5. Is a χ2 value of -12.3 possible? Why or why not?

No, because all χ2 values are positive numbers.

6. Can you use the  2 distribution for a 2 sided test of

significance?
No. We always test that if it greater. We are interested if the differences between observed and
expected are really big or not so big. So we always test greater than.

7. Suppose you run a test and get a P-Value of 0.025. Which of the following is true?
a. The results are significant at the 5% and 1% level. NO.
b. The results are significant at the 5% level, but not at the 1% level. YES.
c. The results are significant at the 1% level, but not the 5% level. NO.
d. The results are not significant at the 5% or 1% level. NO.

P-Value of 0.025 is less than 0.05 but greater than 0.01.

2
Show all appropriate work on the following problems. All tests of significance must have the following to
be considered complete:
- Conditions stated and checked - Expected values shown in a table
- Hypotheses (written out) - Test statistic formula and value
- P-value (probability notation and value) - Conclusion (full 2 sentences)

8. “Looking Up to Athletes” (USA Today, May 7, 1991) reported “Here’s how sports team members say
athletes do as role models for children: Excellent – 16%, Good – 38%, Fair – 41%, Poor – 5%.” Suppose
you took a random poll of 350 members within your community and obtained the following results (in
the same order): 44, 145, 133, 28. Do your results show that your community has a significantly
different idea about athletes as role models than the sports team members?
χ2 Goodness of Fit test (because there is no table)
Opinion Excellent Good Fair Poor
Observed 44 145 133 28
Expected 56 133 143.5 17.5

Conditions:
1) SRS - Stated as random
2) All expected values > 5 – Yes

Ho: the observed distribution of community views of sports team members fits the expected
distribution.
Ha: the observed distribution of community views of sports team members doesn’t fit the expected
distribution.

df = 3
 44  56   28  17.5
2 2

 
2
3  ...   10.72
56 17.5
P  3  10.72   0.013
2

We reject Ho because p-value of 0.0133 less than alpha of 0.05.

We have sufficient evidence that the observed distribution of community views of sports team
members doesn’t fit the expected distribution.

3
9. Alcohol and nicotine consumption during pregnancy may harm the fetus permanently. Because
drinking and smoking behaviors may be related, it is important to understand the nature of this
relationship when assessing the possible effects on children. One study classified 452 randomly
selected mothers according to their alcohol intake prior to pregnancy recognition and their nicotine
intake during pregnancy. Carry out a complete analysis of the association between alcohol and
nicotine consumption.
Nicotine (mg/day)
None 1-15 16 or more
Alcohol None 105 7 11
(oz/day) 0.01-0.10 58 5 13
0.11-0.99 84 37 42
1.0 or more 57 16 17

χ2 test for association (because there is a table given)

Nicotine (mg/day)
None 1-15 16 or more
Alcohol None 82.73 17.69 22.59
(oz/day) 0.01-0.10 51.12 10.93 13.96
0.11-0.99 109.63 23.44 29.93
1.0 or more 60.53 12.94 16.53

Conditions:
1) SRS - Stated
2) All expected values > 5 – Yes

H0: There is no association between alcohol and nicotine consumption.

HA: There is an association between alcohol and nicotine consumption.

df = 6
105  82.37  17  16.53
2 2

 
2
6  ...   42.257
82.73 16.53
P  6  42.257   1.64 10
2 7

We reject Ho because p-value of 1.64 x 10-7 less than alpha of 0.05.

We have sufficient evidence that there is an association between alcohol and nicotine consumption.

4
10. A city expressway utilizing four lanes in each direction was studied to see whether drivers preferred to
drive on the inside lanes. A total of 1000 automobiles were observed during the heavy early-morning
traffic, and their respective lanes were recorded. Do the data present sufficient evidence to indicate
that some lanes are preferred, or that they are EQUALLY preferred? State your hypotheses and carry
out the test.

Lane 1 2 3 4
Observed Count 294 276 238 192
EXPECTED 250 250 250 250

χ2 Goodness of Fit test (because there is no table)

Conditions:
1) SRS – assumed
2) All expected values > 5 – yes

Ho: the observed distribution of lane preference fits the expected distribution.
Ha: the observed distribution of lane preference doesn’t fit the expected distribution.

df = 3
 294  250  192  250 
2 2

 
2
3  ...   24.48
250 250
P  3  24.48  1.983 10
2 5

We reject Ho because p-value of 1.983 x 10-5 is less than alpha of 0.05.

We have sufficient evidence that the observed distribution of community views of sports team
members doesn’t fit the expected distribution.
Therefore some lanes are preferred.

5
11. A psychologist is investigating how a person reacts to a certain situation. He feels the reaction may be
influenced by how ethnically pure the person’s neighborhood is. He collects data on 500 people. Does
there appear to be a relationship between neighborhood and reaction at the 0.10 level of significance?
OBSERVED EXPECTED
Pure? Mild Medium Strong Mild Medium Strong
Yes 170 100 30 144 120 36
No 70 100 30 96 80 24

χ2 test for association (because there is a table given)

Conditions:
1) SRS 1) assumed
2) All expected values > 5 2) see matrix B in calculator

Ho: There is no association between purity and reaction.

Ha: There is an association between purity and reaction.
df = 2
170  144   30  24 
2 2

 
2
2  ...   22.569
144 24
P   2  22.569   1.26 10
2 5

We reject Ho because p-value of 1.26 x 10-5 is less than alpha of 0.10.

We have sufficient evidence that there is an association between purity and reaction.