Week 05 - Measurement

# Week 05 - Measurement
## Survey Biases and List Experiments
<html>
<div style="float:left">

</div>
<hr color='#EB811B' size=1px width=800px>
</html>
### Danilo Freire
### 20th February 2019

---

<style>

.remark-slide-number {
  position: inherit;
}

.remark-slide-number .progress-bar-container {
  position: absolute;
  bottom: 0;
  height: 6px;
  display: block;
  left: 0;
  right: 0;
}

.remark-slide-number .progress-bar {
  height: 100%;
  background-color: #EB811B;
}

.orange {
  color: #EB811B;
}
</style>

# Today's Agenda

* Social desirability bias

* List experiments

* Final project
]

---

# Descriptive Statistics

* Standard deviation: average distance from points to the mean
	
	- `$SD = \sqrt{\frac{1}{n} \sum^{n}_{i = 1} (x_{i} - \bar{x})^{2}}$`

* Outliers: 1.5x IQR

* Median is robust to outliers

* When `$n$` is large, mean and median are roughly the same
]

---

# Bar Plots

* Bar plots:

- Describe factor variables
	- `barplot(prop.table(table(dataset$variable)))`
]

---

# Bar Plots

```r
afghan <- read.csv("https://raw.githubusercontent.com/pols1600/pols1600.github.io/master/datasets/measurement/afghan.csv")
barplot(prop.table(table(afghan$violent.exp.taliban)), main = "Taliban Victimisation Rate")
```

---

# Histograms

- Describe numeric variables
	- Show the shape of the variable distribution
	- Densities are .orange[not] percentiles
	- Densities are the proportion of _each bin_
	- `hist(dataset$variable, freq = FALSE)`
]

---

# Histograms

```r
hist(afghan$age, main = "Age Distribution", freq = FALSE)
```

---

# Box Plots

- Distribution of a numeric variable
	- Quartiles, median, IQR, outliers
	- `boxplot(variable ~ group)` (where `~ group` is optional)
]
---

# Box Plots

```r
boxplot(afghan$age ~ afghan$province, main = "Age Distribution by province")
```

---

# Survey Biases

- Individual non-response
	- Item non-response
]

---

# Survey Biases

* Quote sampling doesn't control for .orange[unobservables]

* Random sampling produces groups identical on average

* Difficulties:

- Some individuals have more chances of being selected
	- Costs
]

---

# Social Desirability Bias

* Examples: support for drug use, abortion, etc

* Under- or overestimation of true proportion
]

---

# List Experiments

* Grant anonymity to respondents

* Control group sees a list of statements

* Treatment group sees the same list _plus a sensitive item_

* _Assuming that respondents don't lie and that both groups would answer the same number of non-sensitive items_, we can infer their true preferences
]
---

# Racial Attitudes

.font150[
* Example based on Graeme Blair and Kosuke Imai, 2012, [Statistical Analysis of List Experiments](https://www.jstor.org/stable/41403738), _Political Analysis_ 20(1), 47-77. 
]
---

# Racial Attitudes

Now  I’m  going  to  read  you  three  things  that  sometimes  make  people  angryor upset. After I read all three, just tell me HOW MANY of them upset you. (I don’t want to know which ones, just how many.)

1) "the federal government increasing the tax on gasoline;"

2) "professional athletes getting million-dollar-plus salaries;"

3) "large corporations polluting the environment."

How many, if any, of these things upset you?
]
---

# List experiments

1) "the federal government increasing the tax on gasoline;"

2) "professional athletes getting million-dollar-plus salaries;"

3) "large corporations polluting the environment."

How many, if any, of these things upset you?
]

---

# Racial Attitudes

```r
library(list)
data(race)
summary(race)
```

```
##        y             treat             age            south      
##  Min.   :0.000   Min.   :0.0000   Min.   :1.800   Min.   :0.000  
##  1st Qu.:2.000   1st Qu.:0.0000   1st Qu.:3.000   1st Qu.:0.000  
##  Median :2.000   Median :1.0000   Median :4.000   Median :0.000  
##  Mean   :2.169   Mean   :0.5144   Mean   :4.273   Mean   :0.235  
##  3rd Qu.:3.000   3rd Qu.:1.0000   3rd Qu.:5.200   3rd Qu.:0.000  
##  Max.   :4.000   Max.   :1.0000   Max.   :8.800   Max.   :1.000  
##                                                                  
##       male           college           state    
##  Min.   :0.0000   Min.   :0.0000   CA     :130  
##  1st Qu.:0.0000   1st Qu.:0.0000   TX     : 67  
##  Median :0.0000   Median :1.0000   PA     : 64  
##  Mean   :0.4402   Mean   :0.5804   FL     : 63  
##  3rd Qu.:1.0000   3rd Qu.:1.0000   IL     : 62  
##  Max.   :1.0000   Max.   :1.0000   NY     : 62  
##                                    (Other):765
```

---

# Racial Attitudes

```r
round(mean(race$y[race$treat == 1]) - mean(race$y[race$treat == 0]), 4)
```

```
## [1] 0.0678
```

```r
diff.in.means <- ictreg(y ~ 1, data = race, treat = "treat",
                        J = 3, method = "lm")
summary(diff.in.means)
```

```
## 
## Item Count Technique Regression 
## 
## Call: ictreg(formula = y ~ 1, data = race, treat = "treat", J = 3, 
##     method = "lm")
## 
## Sensitive item 
##               Est.    S.E.
## (Intercept) 0.0678 0.04962
## 
## Control items 
##                Est.    S.E.
## (Intercept) 2.13413 0.03317
## 
## Residual standard error: 0.866365 with 1211 degrees of freedom
## 
## Number of control items J set to 3. Treatment groups were indicated by '' and '' and the control group by ''.
```

---

# Shy Trump Supporter

Coppock, Alexander. 2017. [Did Shy Trump Supporters Bias the 2016 Polls? Evidence from a Nationally-representative List Experiment]( https://alexandercoppock.com/papers/Coppock_shytrump.pdf). Statistics, Politics, and Policy. 8(1):29-40
---

# Shy Trump Supporter

---

# Shy Trump Supporter

---

# Shy Trump Supporter

---

# Questions?

<html><div style='float:left'></div><hr color='#EB811B' size=1px width=720px></html> 
---

# Final Project

<html><div style='float:left'></div><hr color='#EB811B' size=1px width=720px></html> 
---

# Final Project

* 50% of final grade

* Groups of 3 or 4. Please send me your groups until .orange[this Friday]

* 15-page essay (_not_ including R code and graphs) and a 10-minute presentation 
]

---

# Final Project

* Topics of your choosing

* If you can find an article about it, the better

* Deadline?
]

---

# Final Project

- Introduction: relevance, brief summary of results
	- (Short) Literature Review: theoretical puzzle 
	- Data: descriptive statistics and graphs
	- Results: methods and results
	- Discussion: conclusion, shortcomings, future extensions
	- References
]

---

# Final Project

- Same format, about 10 minutes long
	- Please send me a pdf with the slides before the due date
	- Bonus point if you can make [your presentation in Rmarkdown](https://bookdown.org/yihui/rmarkdown/beamer-presentation.html) (_you don't have to_)
]

---

# See you on Friday!