Problem sets

The qss.student package contains the questions and data sets from the problem sets in

K. Imai (2017) ‘Quantitative Social Science: An Introduction’, Princeton University Press

plus a few extra problem sets not in the book.

Unpacking a problem set

Imagine you are assigned a problem set from the book called bias-in-turnout. To complete the problem set make sure you have the package loaded


and unpack the problem set into your current working directory


The current working directory should now look like this:

- bias-in-turnout
  ↳ bias-in-turnout.pdf
  ↳ bias-in-turnout.Rmd
  ↳ bias-in-turnout.tex
  ↳ data
     ↳ turnout.csv

that is, a folder called bias-in-turnout and inside it

  • bias-in-turnout.pdf: a short description of the problem and the data
  • bias-in-turnout.Rmd: the source file for the problem set
  • data another folder containing all the data needed to complete the problem set
  • data/turnout.csv: the data set

As get_pset messages suggest, let’s make the problem set folder our working directory


and open up the Rmd file so we can start filling in the answers to each question


If you’re in RStudio the file should now be editor tab.

Answering the questions

Each question has a header that looks like

## Question 1

To provide an answer to this question, move your cursor to just above the line that says ## Question 2 and start answering the question. When you need to compute something in R, start a ‘code block’. Here’s a bit of text you might write followed by a code block that loads the data set we need and counts the number of observations in it:

We load the data set and count the rows
turnout <- read.csv("data/turnout.csv")

You can have as much text and as many code blocks as you like. (The variables you make in one code block are still there in the next one.)

You can either type the beginning and end lines of the code block or you can press the green Insert button in the Editor pane’s button menu, choose ‘R’, and have it done for you. Just don’t forget to put the code inside the backticks and don’t nest code blocks inside one another or everything will get confused.

Compiling the problem set

Yo can see what your answers look like at any point by pressing the ‘Knit’ button. This will try to run all your code, read all your answers and knit the two together into a document. What kind of a document is controllable from that button. If you chose ‘Knit to HTML’ (the default) your document will be called ‘bias-in-turnout.html’. If you chose ‘Knit to PDF’ or ‘Knit to Word’ it will have a different suffix.

You can compile whenever you like. Somee people like to work in the Rmd file all the time and Knit intermittently to check their code works. Others, particularly when starting off, prefer to do their thinking and coding in a different document and then paste their working code and answers into the Rmd file at the last minute. Whatever works for you.

Getting into (and out of) trouble

Sometimes you might feel it’s best to start again with a fresh copy of the problem set. However, if you


then you’ll get a new problem set inside the one you’re abandoning. This makes things quite confusing. It’s better to go to top level of your files and grab a new copy, calling it something else. You can move your working directory up a level with


In the language of file paths .. means go up into the parent folder i.e. up one level. (In case you’re curious, . indicates the current folder, and if you wanted to go up two levels then ../.. would be what you wanted)

Now you’re back to where you unpacked bias-in-turnout. To grab a fresh copy with another name:

get_pset("bias-in-turnout", newname = "bias-in-turnout-the-sequel")

(You may have a better idea for a name.) Now you can do it all again, only better.

All the other problem sets

If you’re curious about the other problem sets you can get a list of their names with

#>  [1] "betting-markets"             "betting-markets-with-lm"    
#>  [3] "bias-in-turnout"             "bias-newspapers"            
#>  [5] "conditional-cash-transfers"  "constitutions"              
#>  [7] "criminal-record"             "election-clinton-trump"     
#>  [9] "election-US2016"             "empathy-in-courts"          
#> [11] "ethnic-voting"               "exclusionary-attitudes"     
#> [13] "fox-news-effect"             "gay-marriage"               
#> [15] "gay-marriage-revisited"      "gov-corruption"             
#> [17] "gov-transfer-brazil"         "health-savings"             
#> [19] "ideology-of-justices"        "immigration-threat"         
#> [21] "indiscriminate-violence"     "leader-assassination"       
#> [23] "leader-assassination-graphs" "mapping-elections"          
#> [25] "migration-mechanism"         "music-markets"              
#> [27] "oil-democracy"               "partner-violence"           
#> [29] "political-efficacy"          "population-dynamics"        
#> [31] "poverty-decision"            "reducing-transphobia"       
#> [33] "small-class-size"            "supreme-court-citations"    
#> [35] "trade-networks"              "un-voting"

and you can take a peek at one of them (constitutions) without unpacking any files using