A political scientist has data on individuals who ran for U.S. Congress in 1990. The data indicates which individuals won their elections, and which individuals lost, and what their vote shares were. He also has data on the personal wealth in 2010 of all the individuals (winners and losers) in this dataset. He is interested in the causal effect of holding elected political office in the U.S. on personal wealth, e.g. because political office might provide connections and access that can be exploited for financial gain.
(a) Suppose that the political scientist finds that winners in 1990 Congressional races were richer in 2010 than losers in 1990 Congressional races. Does this prove that holding elected office leads to an increase in personal wealth?
(b) The political scientist decides to compare only those candidates that barely won office in 1990, to those candidates that barely 3lost in 1990. He ignores the rest of the observations in the data set. While visiting Stanford to present his findings in a research seminar, a Stanford faculty member says it is crazy to throw away any information, and that the research must include all the observations. How should the presenter respond to this point?
(c) What is the name of the type of analysis carried out in the previous sub-part?