class: center, middle, inverse, title-slide # 400 Level Info Night ## Econ 474 - Fall 2021 ### Marcelino Guerra ### April 14, 2021 --- # Beyond correlation on social sciences
--- # The road not taken & Random assignment .pull-left[ Naive comparisons lead us to biased results. The goal is to have a strategy that makes the groups choosing different paths as similar as possible The best way to find answers to causal questions is through randomized trials. However, that can be expensive, time consuming and (sometimes) unethical **The causal effect of CoronaVac on** 1. Number of symptomatic Covid-19 Cases 2. Disease severity * 9,823 participants (healthcare professionals) received two shots: 4,953 in the vaccinated group, and 4,870 in the placebo group * The random assignment and the large number of participants make the vaccinated and the placebo group, on average, very similar - you are comparing apples to apples! ] .pull-right[ ![](figs/vax.png) **Source**: Palacios et al. * 253 confirmed COVID-19 cases in the cohort: 85 among the vaccinated group, 168 among the placebo group. The efficacy is 50.7% * 6 severe COVID-19 cases in the placebo group and zero in the vaccinated group ] --- # The furious five <style type="text/css"> .pull-left2 { float: left; width: 59%; } .pull-right2 { float: right; width: 40%; } .pull-right2 ~ p { clear: both; } </style> .pull-left2[ ### You will learn... * Random assignment * Regression * Instrumental variables * Regression discontinuity designs * Differences-in-differences ### ...to answer causal questions * What are the economic returns to attendance at elite private colleges? * Do police reduce crime? * Does the minimum legal drinking age reduce mortality? * What is the effect of minimum wage on unemployment? ] .pull-right2[![](https://media.giphy.com/media/MBVemoHuyw9Ik/giphy.gif) ### Evaluation (tentative) * 5 Homework assingments (40%) * One midterm (20%), one Lab Quiz (20%) * Final project (20%) * You will use
and `Rmarkdown` ]