@Academic Giant Statistics help reseach methods
Michael262015
chapter_12_research_methods.docxEducation and Earnings Continued
12.1 Think of some other common causes of education and earnings in the real world—other than those mentioned in this chapter. When estimation the casual effect of education on earnings, what bias might result from their omission?
Thinking about important control variables to include
12.2 Following are some possible relationships that we might find in observational date that come, say, from a sample survey of U.S. adults:
a. Computer skills → earnings
b. Exercise → diabetes
c. marital status → Happiness
Given that these simple relationships come from observation date, and what are we interested in getting at the true casual effects suggested by the arrows, what control variables would we need? Think about likely common causes of both variables, and be careful not to pick intervening variables.
Country matchmaking
12.3 Freedom House, an advocacy organization, classifies countries as “free” versus “partly free” or “not free.” Say you are interested in whether freedom causes countries to be more prosperous (lower poverty). Below are fine “free” countries in 2013, according to Freedom House:
a. Costa Rica, Uruguay, Ghana, India, South Korea
b. First, find the Map of Freedom on the organization’s website: www.freedomhouse.org
c. For each of the five “free” countries above, find a matching “partly free” or “not free” country.
d. What criteria (variables) did you use to make your matches? Why do these make good control variables?
e. Would a comparison of poverty in the five “free” countries with the five matching countries you chose demonstrate that freedom is the cause? Why or why not?
