These are some questions we reviewed in class and similar questions for you to practice.
Let’s say the only context given is this table: We are interested in knowing the effect of alcohol consumption on grades, and that grades are measured as GPA, beer tax ranges from 0 to 1 where 1 represents 100%, and standardized test scores (from SAT) from 400-1600 points. Given that information, answer the following questions:
- Where in the table does one find results that represent the first stage?
- Where in the table does one find results that represent the second stage?
- Where in the table does one find results representing the reduced form stage?
- What is the conclusion regarding the causal effect and its interpretation?
- What is the effect of an increase of 1 percentage point of beer taxes on GPA?
- What is the average drinks per week in the UVA community, knowing that we have a 6% alcohol tax in Charlottesville and an SAT score of 760?
The following table presents an example of the effect of the number of children on educational investment. In this example, we discussed using second-born twins as a potential IV. Another IV that is proposed here is same-sex sibships. The idea is that families that have their first two kids of the same sex may be more likely to have a third child (and therefore an increasing number of children) relative to families with the first two kids being of different sexes.
- Write down the regressions that represent the results from this table
- Which estimate is statistically significant and which is not?
- These tables represent what:
- The first stage
- Second stage
- reduced form way
- my fears
- Who are the compliers of the “same-sex” instrument?
- How many regressions are represented here?
- What would be the overall conclusion if only the results were based on OLS?
- What would be the overall conclusion considering all of the 2SLS results?
- Use information from Tables 3.5 and 3.4 to back out the coefficient of the instrument on one of the outcomes for each instrument setting.