Mystery

This exercise examines the relationships between deaths from a disaster and economic status (I = high, II = middle, III = low, Other = ?), Sex, and Age (Adult, Child). Click on Data under Chapter 1 in the left panel and then select Open remote ... under the File menu in the right panel. Select mysterybysex.xml from the Open dialog and then click the Open button. Note that EconStatus is assigned the X role and Sex is assigned the Y role by default, whereas Freq is assigned the F role. Select xy|z from the Graph menu. You will also need to use mysterybyage.xml for the last few questions.

  1. Select Contingency table from the Options menu item in the Plot menu. Note that this contingency table gives the population exposed to risk since it sums over the Survive? variable. Select Total % from the Display menu of the Contingency Table. What percentage of the population exposed to risk (the Col %) is Other?
  2. Select Row % from the Display menu of the Contingency Table. Notice that the Female percentages correspond to the blue bars in the Mosaic Plot and the Male percentages correspond to the dark pink bars. What is the percentage of females exposed to risk among those in EconStatus I? II? III? Other? What trend do you see? Do not close the mosaic plot or the contingency table.
  3. Assign Survive? the Z role (by double clicking the Survive? column header and selecting Z from the Column Role menu). Then select xy|z plot from the Graph menu. Click on the right of the Survive? conditioning tool (below the Plot menu) so that we expose the Dead rather than the Alive. Select Contingency table from the Plot menu and then select Row % from the Display menu. What is the percentage of females killed among those in EconStatus I? II? III? Other? What trend do you see?
  4. You now have two contingency tables called Contingency Table (from Question 1) and Contingency Table for Dead (from Question 3). Select Count from the Display menu in both of these tables. Compute the deaths per 100 exposed to risk for each economic status group and sex by dividing the numbers in the Contingency Table for Dead by the corresponding numbers in the Contingency Table (including the margins) and multiplying by 100. For example, the male deaths per 100 exposed to risk in economic status group I is: (118/180) x 100 = 65.6. The corresponding female death rate is: (4/145) x 100 = 2.8. Which status group had the lowest overall death rate (right table margin rates)? Which had the highest?
  5. Is the overall death rate higher among males or females (bottom table margin rates)?
  6. Compare the death rates of males and females for economic group I; for economic group II; for economic group III. What difference do you see among the groups when comparing these male/female rates?
  7. Close all open windows in the right-hand panel and open the mysterybyage.xml file from the Open Remote... menu item of the File menu. Repeat the directions in Questions 1 and 2 noting that age substitutes for sex. What is the percentage of children exposed to risk among those in EconStatus I? II? III? Other? What trend do you see? Do not close the mosaic plot or the contingency table.
  8. Follow the directions in Questions 3 and 4 again replacing sex by age. Look at the death rates for children in groups I, II, and III (no children are in Other). What do you notice.
  9. Guess what caused all these deaths. Explain what evidence you used to make your guess.
  10. Who do you think the Others were? Why?