title |
author |
date |
output |
Lecture 14 Zimmerman 2014 |
Nick Huntington-Klein |
`r Sys.Date()` |
revealjs::revealjs_presentation |
theme |
transition |
self_contained |
smart |
fig_caption |
reveal_options |
solarized |
slide |
true |
true |
true |
|
|
|
```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = FALSE, warning=FALSE, message=FALSE)
library(tidyverse)
theme_set(theme_gray(base_size = 15))
```
## Zimmerman 2014
- Zimmerman 2014 uses a cutoff in the admissions process to estimate the returns to education for academically marginal students
- Today we will be discussing that paper
## Zimmerman 2014
First off:
- What does he look for and what does he find?
- Why might we be particularly interested in the returns to education for marginal students?
- How do we know that RDD gives us the return for *those* students?
- What kind of RDD is this?
- How can we characterize his results and any strengths/weaknesses?
## Zimmerman 2014
- Why does he check for *manipulation of the running variable* in Section V.A?
- Why might this be important?
- What does manipulation mean and why might it mess up an RDD result?
- How does he do this check?
## Running Variable Notes
- We can do these sorts of tests ourselves for manipulation using the `rddensity()` and `rdplotdensity()` functions in the **rddensity** package
- Other potential issues with running variables: *granularity*
- Why might it be difficult to do an RDD if the running variable is very *coarsely defined*?
## Zimmerman 2014
- What other tests does he do?
- What does Figure 3 show us?
- How can we get the results from the graphs and from the regression tables?
- Is there anything we might want to do differently in this study?