Threats to internal and external validity. A researcher wants to measure the income gains from migration. She has data on wages from a random sample of people who recently migrated to the Bay Area and she also has data on wages from a random sample of people living in six countries that are the source of many migrants to the Bay Area: Mexico, China, the Philippines, Vietnam, India, and South Korea. She pools together all of the data and runs the following regression: log(wage_i) = beta_0 + beta_1 migrant_i + u_i, where migrant_i is a dummy variable that equals 1 if the person is a migrant to the Bay Area and equals 0 if the person is still living in his or her home country. (a) For each potential threat to internal validity, assess whether or not it poses a problem in the regression and explain why or why not. List each threat clearly. If necessary, make further assumptions about the nature of the data and state them clearly.
(b) In your opinion, what are the two most important threats to the study's internal validity? Explain why.
(c) Provide two examples of potential populations-of-interest:
i. One for which you think the study is externally valid. Explain.
ii. One for which you think the study is not externally valid. Explain.