An “Unusual Episode”
Purpose:
This activity is intended to illustrate the use of contingency tables (two-way tables).
Background:
The tables below contain a data set giving counts for a population at risk and fatalities for an “unusual episode.” This mortality episode is well known historically. There were 2201 people at risk. Those at risk are categorized by economic status (I -- high or rich, II -- medium, III -- low or poor, and Other), age (child or adult), gender (female or male), and survival status (survived or did not survive). More information cannot be specified about the Other economic status group. This would give away the answer to what the “unusual episode” describes.
Instructions:
Examine the data carefully. Look for any interesting features. To help get a more complete picture of how each of the explanatory variables (economic status, age, and gender) is related to the response variable (survival status), complete the Questions.
After examining the data and completing the Questions, give your best “educated guess” as to what historical episode this data set describes.
Data:
Population at Risk and Deaths for an “Unusual Episode”:
|
|
|
|
By Economic Status and Gender |
|
|||||
|
|
Population Exposed to Risk |
|
Number of Deaths |
|
|||||
|
Economic status |
Male |
Female |
Both |
|
Male |
Female |
Both |
||
|
I (rich) |
180 |
145 |
325 |
|
118 |
4 |
122 |
||
|
II |
179 |
106 |
285 |
|
154 |
13 |
167 |
||
|
III (poor) |
510 |
196 |
706 |
|
422 |
106 |
528 |
||
|
Other |
862 |
23 |
885 |
|
670 |
3 |
673 |
||
|
Total |
1731 |
470 |
2201 |
|
1364 |
126 |
1490 |
||
|
|
|
|
By Economic Status and Gender |
|
|||
|
|
Population Exposed to Risk |
|
Number of Deaths |
|
|||
|
Economic status |
Male |
Female |
Both |
|
Male |
Female |
Both |
|
I (rich) |
180 |
145 |
325 |
|
118 |
4 |
122 |
|
II |
179 |
106 |
285 |
|
154 |
13 |
167 |
|
III (poor) |
510 |
196 |
706 |
|
422 |
106 |
528 |
|
Other |
862 |
23 |
885 |
|
670 |
3 |
673 |
|
Total |
1731 |
470 |
2201 |
|
1364 |
126 |
1490 |
Questions:
1.
(a) Calculate the overall percentage of deaths: _____% of those at risk died.
(b) Calculate the percentage of the overall deaths that were male and female.
Deaths
Male Female
% %
2. Do you think that there is an association between Gender and Number of Deaths for the “unusual episode?” Explain.
3. Construct a two-way table to display the variables Economic Status and Survival Status:
Economic Status
I II III Other
Survival Died 122
Status
Survived
4. Calculate the percentage of deaths in each of the four economic status groups. You have calculated the conditional distributions of Survival Status given Economic Status.
5. Do you think that there is an association between Economic Status and Survival Status for the “unusual episode?” Explain.
6. Construct a two-way table to display the variables Economic Status and Male Survival Status:
Economic Status
I II III Other
Male Died 118
Survival
Status Survived
7. Calculate the conditional distributions of Male Survival Status given Economic Status.
8. Do you think that there is an association between Economic Status and Male Survival Status for the “unusual episode?” Explain.
9. Construct a two-way table to display the variables Economic Status and Female Survival Status:
Economic Status
I II III Other
Female Died 4
Survival
Status Survived
10. Calculate the conditional distributions of Female Survival Status given Economic Status.
11. Do you think that there is an association between Economic Status and Female Survival Status for the “unusual episode?” Explain.
12. Construct a two-way table to display the variables Economic Status (I, II, III) and Child Survival Status:
Economic Status
I II III
Child Died
Survival
Status Survived 6
13. Calculate the conditional distributions of Child Survival Status given Economic Status (I, II, III).
14. Do you think that there is an association between Economic Status and Child Survival Status for the “unusual episode?” Explain.
·What “unusual episode” in history do you think this data set describes? Explain.
Answers to Activity Questions and Assessment Question:
Activity Questions.
1.
(a) 1490/2201 = 68% of those at risk died
(b) Deaths
Male Female
1364/1490 = 92% 126/1490=8%
2. It appears that there is an association between Gender and Number of Deaths. The deaths were overwhelmingly male. (The high mortality rate among males might be the result of the rule of ‘women and children first’.)
3. Economic Status
I II III Other
Survival Died 122 167 528 673
Status
Survived 203 118 178 212
325 285 706 885
4. Economic Status
I II III Other
Died 
Survival
Status
Survived 
5. It appears that there is an association between Economic Status and Survival Status. The lower economic status groups (and the Other group) have higher mortality rates. (The high mortality rate among those in Class III might be explained by the more vulnerable position of the third-class cabins, lower in the hull of the ship.)
6. Economic Status
I II III Other
Male Died 118 154 422 670
Survival
Status Survived 62 25 88 192
180 179 510 862
7. Economic Status
I II III Other
Male Died 
Survival
Status
Survived 
8. It appears that there is an association between Economic Status and Male Survival Status. The highest economic status group had the lowest mortality rate.
9. Economic Status
I II III Other
Female Died 4 13 106 3
Survival
Status Survived 141 93 90 20
145 106 196 23
10. Economic Status
I II III Other
Female Died 
Survival
Status
Survived 
11. It appears that there is an association between Economic Status and Female Survival Status. The lowest economic status group had the highest mortality rate.
12. Economic Status
I II III
Child Died 0 0 52
Survival
Status Survived 6 24 27
6 24 79
13. Economic Status
I II III
Child Died 0 0 
Survival
Status
Survived 100%
100% 
14. It appears that there is an association between Economic Status and Child Survival Status. The lowest economic status group had the lowest child survival rate.
·The “unusual episode” is the sinking of the ocean liner Titanic after colliding with an iceberg on April 15th, 1912.
Assessment Question.
(a) The company should calculate the percentage of each claim amount for each type of vehicle. That is, the company should calculate the conditional distributions of claim amount given type of vehicle.
(b) Type of Vehicle
Car Truck Sport Utility
Claim Amount 
$10,000 
$10,000 
(c) It appears that there is an association between the type of vehicle and the claim amount. Sport utility vehicles have a much higher percentage of claims over $10,000 than either cars or trucks.