Introduction to R

Adrian Rohit Dass

Institute of Health Policy, Management, and
Evaluation
Canadian Centre for Health Economics
University of Toronto

September 17th, 2021

 Outline
• Why use R?
• R Basics
• R for Database Management
– Reading-in data, merging datasets, reshaping, recoding variables, sub-
setting data, etc.
• R for Statistical Analysis
– Descriptive and Regression Analysis
• Other topics in R
– Tidyverse
– Parallel Processing
– R Studio
• R Markdown
• Applied Example
• R Resources
Learning Curves of Various Software
Packages

Source: https://sites.google.com/a/nyu.edu/statistical-software-guide/summary
Summary of Various Statistical
Software Packages

Source: https://sites.google.com/a/nyu.edu/statistical-software-guide/summary
 Goals of Today’s Talk
• Provide an overview of the use of R for database
management
– By doing so, we can hopefully lower the learning curve
of R, thereby allowing us to take advantage of its “very
strong” data manipulation capabilities
• Provide an overview of the use of R for statistical
analysis
– This includes descriptive analysis (means, standard
deviations, frequencies, etc.) as well as regression
analysis
– R contains a wide number of pre-canned routines that
we can use to implement the method we’d like to use
 Part I
R Basics
Command Window Syntax Window
 Programming Language
• Programming language in R is generally object
oriented
– Roughly speaking, this means that data, variables,
vectors, matrices, characters, arrays, etc. are treated
as “objects” of a certain “class” that are created
throughout the analysis and stored by name.
– We then apply “methods” for certain “generic
functions” to these objects
• Case sensitive (like most statistical software
packages), so be careful
 Classes in R
• In R, every object has a class
– For example, character variables are given the
class of factor or character, whereas numeric
variables are integer
• Classes determine how objects are handled by
generic functions. For example:
– the mean(x) function will work for integers but not
for factors or characters - which generally makes
sense for these types of variables
Packages available (and loaded) in R by
default
Package Description
base Base R functions (and datasets before R 2.0.0).
compiler R byte code compiler (added in R 2.13.0).
datasets Base R datasets (added in R 2.0.0).
grDevices Graphics devices for base and grid graphics (added in R 2.0.0).
graphics R functions for base graphics.
grid A rewrite of the graphics layout capabilities, plus some support for interaction.
Formally defined methods and classes for R objects, plus other programming tools, as
methods described in the Green Book.
Support for parallel computation, including by forking and by sockets, and random-
parallel number generation (added in R 2.14.0).
splines Regression spline functions and classes.
stats R statistical functions.
stats4 Statistical functions using S4 classes.
tcltk Interface and language bindings to Tcl/Tk GUI elements.
tools Tools for package development and administration.
utils R utility functions.
Source: https://cran.r-project.org/doc/FAQ/R-FAQ.html

For database management, we usually won’t need to load or install any additional packages,
although we might need the “foreign” package (available in R by default, but not initially loaded)
if we’re working with a dataset from another statistical program (SPSS, SAS, STATA, etc.)
 Packages in R
• Functions in R are stored in packages
– For example, the function for OLS (lm) is accessed via
the “stats” package, which is available in R by default
– Only when a package is loaded will its contents be
available. The full list of packages is not loaded by
default for computational efficiency
– Some packages in R are not installed (and thus
loaded) by default, meaning that we will have to
install packages that we will need beforehand, and
then load them later on
 Packages in R (Continued)
• To load a package, type library(packagename)
– Ex: To load the foreign package, I would type library(foreign) before
running any routines that require this package
• To install a package in R:
– Type install.packages(“packagename”) in command window
– For example, the package for panel data econometrics is plm in R. So, to
install the plm package, I would type install.packages(“plm”).
• Note that, although installed, a package will not be loaded by default
(i.e. when opening R). So, you’ll need library(package) at the top of
your code (or at least sometime before the package is invoked).
– Some packages will draw upon functions in other packages, so those
packages will need to be installed as well. By using install.packages(“ ”), it
will automatically install dependent packages
 Some Basic Operations in R
• Q: If x = 5, and y = 10, and z = x + y, what is the value of z?
• Let’s get R to do this for us:

• In this example, we really only used the ‘+’ operator,

but note that ‘-’, ‘/’, ‘*’, ‘^’, etc. work the way they
usually do for scalar operations
 Some Basic Operations in R
• Now suppose we created the following vectors:
1 2
A= 2 B= 4
3 6

In R, c() is used to combine

• What is A + B? values into a vector or list. Since
we have multiple values, we
need to use it here

• Note that with vectors, ‘+’, ‘-’, ‘/’, ‘*’, ‘^’ perform element-wise
calculations when applied to vectors. So, vectors need to be
the same length.
 Working with Matrices in R
• A matrix with typical element (i,j) takes the following form:

(1,1) (1,2) (1,3)

(2,1) (2,2) (2,3)
(3,1) (3,2) (3,3)

• Where i = row number and j = column number

• In R, the general formula for extracting elements (i.e. single
entry, rows, columns) is as follows:
– matrixname[row #, column #]
• If we leave the terms in the brackets blank (or leave out the
whole bracket term) R will spit out the whole matrix
 Working with Matrices in R (Continued)
• Example: Suppose we had the following matrix:
1 4 7
2 5 8
3 6 9

• To create this matrix in R, type:

> matrix = matrix(c(1, 2, 3, 4, 5, 6, 7, 8, 9), nrow=3, ncol=3)
• Extract the element in row #2, column #3
> matrix[2,3]
8
• Extract the second row
> matrix[2,]
258
Since we require
• Extract the last two columns
multiple
> matrix[,c(2,3)]
columns, we
47
need to use c()
58 here
69
 Working with Matrices in R (Continued)
• Example: Suppose now we had the following vector, with typical element
‘i':
1
2
3

• Extract the third element of the vector

> vector[3]
3
• Suppose the 2nd element should be 5, not 2. How do we correct this
value?
> vector[2] = 5
> vector
1
5
3
 But wait a minute…
• Q: If this is a tutorial on the use of R for database
management/statistical analysis, then why are we
learning about vectors/matrices?
• A: The way we work with data in R is very
similar/identical to how we work with
vectors/matrices
– This is different from other statistical software
packages, which may be a contributing factor to the
“high” learning curve in R
• The importance of vector/matrices operations
will become more clear as we move
 Part II
R for Database Management
 Reading Data into R
What format is the data in?
• Data from Comma Separated Values File (.csv)
– Package: utils
– Formula: read.csv(file, header = TRUE, sep = ",", quote = "\"", dec = ".", fill =
TRUE, comment.char = "", ...)
• Data from Excel File (.xlsx)
– Package: xlsx
– Formula: read.xlsx(file, sheetIndex, sheetName=NULL, rowIndex=NULL,
startRow=NULL, endRow=NULL, colIndex=NULL, as.data.frame=TRUE,
header=TRUE, colClasses=NA, keepFormulas=FALSE, encoding="unknown", ...)
• Data from STATA (.dta)
– Package: foreign
– read.dta(file, convert.dates = TRUE, convert.factors = TRUE, missing.type =
FALSE, convert.underscore = FALSE, warn.missing.labels = TRUE)

Other Formats: See package “foreign”

https://cran.r-project.org/web/packages/foreign/foreign.pdf
 Reading Data into R
Examples:
• CSV file with variable names at top
– data = read.csv(“C:/Users/adrianrohitdass/Documents/R
Tutorial/data.csv”)
• CSV file with no variable names at top
– data = read.csv(“C:/Users/adrianrohitdass/Documents/R
Tutorial/data.csv”, header=F)
• STATA data file (12 or older)
– library(foreign)
– data = read.dta(“C:/Users/adrianrohitdass/Documents/R
Tutorial/data.dta”)
• STATA data file (13 or newer)
– library(readstata13)
– data = read.dta13(“C:/Users/adrianrohitdass/Documents/R
Tutorial/data.dta”)
Comparison and Logical Operators

Operator Description Example

= Assign a value x=5
== Equal to sex ==1
!= Not equal to LHIN != 5
> Greater than income >5000
< Less than healthcost < 5000
>= or <= Greater than or equal to income >= 5000
Less than or equal to healthcost <= 5000
& And sex==1 & age>50
| Or LHIN==1 | LHIN ==5
 Referring to Variables in a Dataset
• Suppose I had data stored in “mydata” (i.e an
object created to store the data read-in from a
.csv by R). To refer to a specific variable in the
dataset, I could type
mydata$varname

Name of Dataset Name of Variable in dataset

‘$’ used in R to extract named

elements from a list
 Creating a new variable/object
• No specific command to generate new
variables (in contrast to STATA’s “gen” and
“egen” commands)
– x = 5 generates a 1x1 scalar called “x” that is equal
to 5
– data$age = year – data$dob creates a new
variable “age” in the dataset “data” that is equal
to the year – the person’s date of birth (let’s say in
years)
 Looking at Data
• Display the first or last few entries of a dataset:
– Package: utils
– View entire dataset in separate window
• View(x, title)
– First few elements of dataset (default is 5):
• head(x, n, …)
– Last few elements of dataset (default is 5):
• tail(x, n, …)
• List of column names in dataset
– Package: base
– Formula: colnames(x)
 Missing Values
Missing Values are listed as “NA” in R
• Count number of NA’s in column
sum(is.na(x))
• Recode Certain Values as NA (i.e. non
responses coded as -1)
x[x==-1] = NA
 Renaming Variables (Columns)
A few different ways to do this:
• To rename the ‘ith’ column in a dataset
– colnames(data)[i] = “My Column Name”
• Can be cumbersome – especially if don’t know column # of the
column you want to rename (just it’s original name)
• Alternative:
– colnames(data)[which(colnames(data) == “R1482600”)] = “race”

Grabs column names Look-up that returns New column name

from specified the column #
dataset
 Subsetting Data
• Subsetting can be used to restrict the sample in the dataset, create
a smaller data with fewer variables, or both
• Recall: extracting elements from a matrix in R
• matrixname[row #, column #]
• What’s the difference between a matrix and a dataset?
– Both have row elements
• Typically the individual records in a dataset
– Both have column elements
• Typically the different variables in the dataset
• If we think of our dataset as a matrix, then the concept of
subsetting in R becomes a lot easier to digest
 Subsetting Data (Continued)
Examples:
• Restrict sample to those with age >=50
> datas1 = data[data$age >=50,]
• Create a smaller dataset with just ID, age, and
height
>datas2 = data[, c(“ID”, “age”, “height”)]
• Create a smaller dataset with just ID, age, and
height; with age >=50
>datas3 = data[data$age>=50, c(“ID”, “age”, “height”)]
 Recoding Variables in R
• Usually done with a few lines of code using comparison
and logical operators
• Ex: Suppose we had the following for age:
> data$age = [19, 20, 25, 30, 45, 55]
• If we wanted to create a categorical variable for age
(say, <20, 20-39, 40-59), we could do the following:
> data$agecat[data$age <20] = 1
> data$agecat[data$age >=20 & data$age <40] = 2
> data$agecat[data$age >=40 & data$age <60] = 3
> data$agecat
> [1, 2, 2, 2, 3, 3]
 Merging Datasets
Suppose we had the following 2 datasets:
Data1 Data2
Id Age Income Id Health Care Cost
1 55 49841.65 1 188.1965
2 63 46884.78 2 172.2420
3 65 45550.87 3 102.8355
4 69 26254.15 4 150.2247
5 52 22044.73

Our first dataset contains some data on age and income, but not health care
costs to the public system. Dataset 2 contains this data, but was not initially
available to us. It also doesn’t have age or income.

The common element between the two datasets is “Id”, which uniquely identifies
the same individuals across the two datasets.

Note that, for some reason, individual 5 does not have a reported health care
cost
 Merging Datasets (Continued)
• Command: merge
– Package: base
• For our example:
Optional, but
– Datam = merge(Data1, Data2, by=“Id”, all=T) default is F,
meaning those
who can’t be
Unique identifier matched will be
across datasets excluded
– Resulting Dataset
Datam
Id Age Income Health Care Cost
1 55 49841.65 188.1965
2 63 46884.78 172.2420
3 65 45550.87 102.8355
4 69 26254.15 150.2247
5 52 22044.73 NA
 Part II

R for Statistical Analysis

 Descriptive Statistics in R
• Mean
– Package: base
– Formula: mean(x, trim = 0, na.rm = FALSE, ...)
• Standard Deviation
– Package: stats
– Formula: sd(x, na.rm = FALSE)
• Correlation
– Package: stats
– Formula: cor(x, y = NULL, use = "everything”, method
= c("pearson", "kendall", "spearman"))
 Descriptive Statistics (Example)
• Suppose we had the following data column in
R (transposed to fit on slide):
– Vector = [5,5,6,4]
• What is the mean of the vector?
• In R, I would type
> mean(Vector)
>5
 Descriptive Statistics (Example)
• Suppose now we had the following:
– Vector = [5,5,6,4,NA]
• What is the mean of the vector?
• In R, I would type
> mean(Vector)
> NA
• Why did I get a mean of NA?
– Our vector included a missing value, so R couldn’t
compute the mean as is.
• To remedy this, I would type
> mean(Vector, na.rm=T)
>5
 Tabulations R
• Tabulations of categorical/ordinal variables can be done with
R’s table command:
– Package: base
– Formula: table(..., exclude = if (useNA == "no") c(NA, NaN), useNA =
c("no”, "ifany", "always"), dnn = list.names(...), deparse.level = 1)
Ex: Table Sex Variable, with extra column for missing values (if
any)
 Graphing Data in R
• Generic X-Y Plotting
– Package: graphics
– Formula: plot(x, y, ...)
Example:
plot(cost.data$income,cost.data$totexp)

• Plotting with ggplot() function

– Package: ggplot2
– Formula: ggplot(data = NULL, mapping = aes(), ..., environment =
parent.frame())
Example:
ggplot(cost.data, aes(x=income, y=totexp)) + geom_point()
Resulting Graph (Generic)
 Resulting Graph (ggplot2)

See https://github.com/rstudio/cheatsheets/raw/master/data-visualization.pdf for

ggplot cheatsheet
 Ordinary Least Squares
• The estimator of the regression intercept and slope(s) that minimizes the sum of
squared residuals (Stock and Watson, 2007).

– Package: stats
– Formula: lm(formula, data, subset, weights, na.action, method = "qr", model =
TRUE, x = FALSE, y = FALSE, qr = TRUE, singular.ok = TRUE, contrasts = NULL,
offset, ...)

Examples:

Regression of “total health care expenditure” on “age, gender, household income,

supplementary insurance status (insurance beyond Medicare), physical and activity
limitations and the total number of chronic conditions” using dataset “cost.data” from
Medical Expenditure Panel Survey (65+)

ols.costdata = lm(totexp ~ age + female + income + suppins + phylim + actlim + totchr,

data = cost.data)

Online Help File

https://stat.ethz.ch/R-manual/R-devel/library/stats/html/lm.html
 Ordinary Least Squares

Example adapted from Jones (2013) Applied Health Economics

 Post-Estimation
Package: lmtest
• Breusch-Pagan test for heteroskedasticity.
bptest(formula, varformula = NULL, studentize = TRUE, data = list())
• Ramsey’s RESET test for functional form.
resettest(formula, power = 2:3, type = c("fitted", "regressor", "princomp"),
data = list())
Package: car
• Variance Inflation Factor (VIF)
vif(model)
Package: sandwich
• Heteroskedasticity-Consistent Covariance Matrix Estimation
coeftest(ols.costdata, vcovHC(ols costdata, type = "HC1"))
Notes: need to combine with lmtest coeftest() command, and use type =
“HC1” to get the same results as STATA’s “robust” command
 Extracting Beta coefficients, standard
errors, etc. from model
• A couple of ways to do this, but most of the information we’re after is stored in the
coefficients object returned from summary:

• The above is a matrix, so we can get the information we need through column
extractions:
– Beta coefficients: summary(ols.costdata)$coefficients[,1]
– Standard errors: summary(ols.costdata)$coefficients[,2]
– T-value: summary(ols.costdata)$coefficients[,3]
– P-value: summary(ols.costdata)$coefficients[,4]
 Residuals vs Fitted Values
• For Residuals vs Fitted Values (RVFV) Plot, use generic plot() function on
regression object. First plot is RVFV
• Formula: plot(ols.costdata, 1)

*The other 5 plots are: Normal Q-Q, Scale-Location, Cook’s distance, Residuals vs
Leverage, and Cook’s distance vs Leverage
 Models for Binary Outcomes
• R does not come with different programs for binary outcomes. Instead, it
utilizes a unifying framework of generalized linear models (GLMs) and a
single fitting function, glm() (Kleiber & Zeileis (2008))

Package: stats
Formula: glm(formula, family = gaussian, data, weights, subset, na.action,
start = NULL, etastart, mustart, offset, control = list(...), model = TRUE,
method = "glm.fit”, x = FALSE, y = TRUE, contrasts = NULL, ...)

• For binary outcomes, we specify family=“binomial” and link= “logit” or

“probit”
• Can be extended to count data as well (family=“poisson”)

Online help: https://stat.ethz.ch/R-manual/R-

devel/library/stats/html/glm.html
 Models for Binary Outcomes
Example: Probit Analysis: factors associated with being arrested
 Instrumental Variables
A way to obtain a consistent estimator of the unknown co-
efficiencts of the population regression function when the
regressor, X, is correlated with the error term, u. (Stock and
Watson, 2007).

Package: AER
Formula: ivreg(formula, instruments, data, subset, na.action,
weights, offset, contrasts = NULL, model = TRUE, y = TRUE, x =
FALSE, ...)

Online documentation: https://cran.r-

project.org/web/packages/AER/AER.pdf
 IV Example
Example: Determinants of Income (As a function of Health)

Prints out F-test for

Weak Instruments,
Hausman Test
Statistic (vs ols) and
Sargan’s Test for
Over-identifying
Restrictions (if more
than one instrument
use)
 Other Regression Models
• Panel Data Econometrics
– Package: plm
– https://cran.r-
project.org/web/packages/plm/vignettes/plm.pdf
• Linear and Generalized Linear Mixed Effects Models
– Package: lme4
– https://cran.r-project.org/web/packages/lme4/lme4.pdf
• Quantile Regression
– Package: quantreg
– https://cran.r-
project.org/web/packages/quantreg/quantreg.pdf
 Part III
Other topics in R
Tidyverse
 Tidyverse
From Tidyverse website:
“The tidyverse is an opinionated collection of R packages
designed for data science. All packages share an underlying
design philosophy, grammar, and data structures…tidyverse
makes data science faster, easier and more fun”

Source: https://www.tidyverse.org

• Packages within tidyverse: ggplot2, dplyr, tidyr, readr, purrr,

tibble, stringr, and forcats

• To get, type: install.packages(“tidyverse”) in R console

 Tidyverse (Continued)
Package: dplyr
• Description: provides a flexible grammar of data
manipulation.
• Example Commands:
– Restrict sample to those with age >=50
• subdata = filter(data, age>=50)
– Create a smaller dataset with just ID, age, and height
• subdata = select(data, ID, age, height)
– Create a smaller dataset with just ID, age, and height;
with age >=50
• subdata = data %>%
filter(age>=50) %>%
select(ID, age, height)
 Tidyverse (Continued)
Package: dplyr
• Example Commands (continued):
– Create new variable (age) in existing dataset
• data = mutate(data, age = year – dob)
– Rename a variable in a dataset (new name = old
name)
• data = rename(data, race = R1482600)
• https://cran.r-
project.org/web/packages/dplyr/dplyr.pdf
 Tidyverse (Continued)
Other (selected) packages in Tidyverse:
• Package: readr
– Description: The goal of 'readr' is to provide a fast and
friendly way to read rectangular data (like 'csv', 'tsv', and
'fwf’)
– https://cran.r-project.org/web/packages/readr/readr.pdf
• Package: tidyr
– Description: Tools for reshaping data, extracting values out
of string columns, and working with missing values
– https://cran.r-project.org/web/packages/tidyr/tidyr.pdf
Parallel Processing
 Parallel Processing in R
• Parallel computing: From Wikipedia: “Parallel computing is a type of
computation in which many calculations or the execution of
processes are carried out simultaneously. Large problems can often
be divided into smaller ones, which can then be solved at the same
time.”
– See here for more:
https://en.wikipedia.org/wiki/Parallel_computing
• Modern day computers typically contain:
– Single-core
– Multicore (Dual, Quad, Hexa, Octo, etc.)
• May also contain hyperthreading
Parallel Processing in R (Continued)
• Parallel processing can be used in many
situations, including:
– Bootstrapping
– Microsimulation models
– Monte Carlo experiments
– Probabilistic Sensitivity Analysis
• By utilizing parallel processing, we can
significantly speed up the processing time of
our calculations
Parallel Processing in R (Continued)
• There are many packages to perform parallel processing in R, including
• parallel
– Available in R by default
– Handles large chunks of computations in parallel
– https://stat.ethz.ch/R-manual/R-
devel/library/parallel/doc/parallel.pdf

• doParallel
– “parallel backend” for the “foreach” package
– provides a mechanism needed to execute foreach loops in parallel
– https://cran.r-
project.org/web/packages/doParallel/vignettes/gettingstartedParallel.
pdf
Example: Monte Carlo Experiment
Example: Monte Carlo Experiment
(Continued)

Notice we changed %dopar%

to %do% to run everything
through a single core
R Studio
 What is R Studio?
From R Studio Website:
• An integrated development environment (IDE) for R.
Includes:
– A console
– Syntax highlighting editor
– Tools for plotting, history, debugging, and workspace
history
• Can think of it as a more user friendly version of R
• A free version is available as well
• For more information, see https://www.rstudio.com
 List of datasets/variables

Syntax Window

Files, plots, packages, help, and viewer

Command/Results Window
R Markdown
(In R Studio)
 What is R Markdown?
From R Markdown website:
“R Markdown provides an authoring framework for data science. You
can use a single R Markdown file to both
• save and execute code
• generate high quality reports that can be shared with an audience”
Source: https://rmarkdown.rstudio.com/lesson-1.html

With R Markdown, you can render to a variety of formats, which

includes PDF (uses LaTeX) and Microsoft Word

To create a R Markdown file, go to File à New File à R Markdown

“Knit”, or generate document

Global options for document here (echoing of R code, loading

packages, etc.)

# for Document Sections

R code chunk for output (summary of “cars” data)

R code chunk for output (to insert a plot available in R memory)

Page 1 (of 2)
Page 2 (of 2)
 Tips for Outputting In MS Word
Output Option • The word_document2 (Bookdown) and rdocx_document (Officedown) formats are
generally superior to word_document (default in R Markdown), particularly for
automatic numbering of figures/tables, and cross-referencing of figures/tables.
• The rdocx_document lets you easily switch between landscape and portrait
Tables Default knitr::kable() function works, but flextable() function flextable creates “pretty”
tables with a large amount of flexibility (customize cell padding and column widths, table
footnotes, long tables, etc.)
Figures Use knitr::include_graphics(filepath) for previously saved figures to include in the
document
References • Default reference style is Chicago. Visit Zotero Style Repository to search for additional
Citation Style Language (CSL) files (Vancouver, APA, journal specific styles, etc.). Can
modify existing reference style, which may be necessary for certain journals
(https://editor.citationstyles.org/about/)
• Add citations with markdown syntax by typing [@cite] or @cite.
• Store references in plain text BibTeX database (*.bib)
• Can also look up and Insert Citations dialog in the Visual Editor by clicking the @
symbol in the toolbar or by clicking Insert > Citation
Document To modify font sizes, text alignment, etc., need to create a reference style document
formatting following these instructions: https://rmarkdown.rstudio.com/articles_docx.html

Please also see the R Markdown cheat sheet:

https://github.com/rstudio/cheatsheets/raw/master/rmarkdown-2.0.pdf
 Applied Example
• Analysis of Health Expenditure Data in Jones et al.
(2013) Chapter Three
• The data covers the medical expenditures of US citizens
aged 65 years and older who qualify for health care
under Medicare.
– Outcome of interest is total annual health care
expenditures (measured in US dollars).
– Other key variables are age, gender, household income,
supplementary insurance status (insurance beyond
Medicare), physical and activity limitations and the total
number of chronic conditions.
• Data can be downloaded from here (mus03data.dta):
https://www.stata-press.com/data/musr.html
 R Markdown Code From Example
---
title: "Untitled"
output: word_document
---

```{r setup, include=FALSE}

knitr::opts_chunk$set(echo = FALSE)
```

# Regression Results

```{r regresults}
load("cost.data.results.RData")
knitr::kable(cost.data.results)
```

# Plot

```{r plot}
knitr::include_graphics("RVFV.jpg")
```
 Conclusions
• R has extremely powerful database management
capabilities
– Is fully capable of performing the same sort of tasks as
commercial software programs
– Can be enhanced through Tidyverse package for a more user
friendly experience
• R is very capable of statistical analysis
– Is fully capable of calculating summary statistics and performing
regression analysis right out of the box
– Can install additional packages to perform other sorts of
analysis, depending on the research question of the user
– Performance can be improved by the use of parallel processing
• R, and the additional packages available to enhance the use
of R, are available free of charge
R Resources
 R Online Resources
• A list of R packages is contained here:
https://cran.r-
project.org/web/packages/available_packages_by_
date.html
• By clicking on a particular package, you’ll be
taken to a page with more details, as well as a link
to download the documation
• Typing help(topic) in R pulls up a brief help file
with synax and examples, but the online manuals
contain more detail
 R Online Resources
• UCLA Institute for Digital Research and
Education
– List of topics and R resources (getting started, data
examples, etc.) can be found here:
http://www.ats.ucla.edu/stat/r/
• RStudio Cheatsheets
– https://www.rstudio.com/resources/cheatsheets/
 Other R Resources
1. Kleiber, C., & Zeileis, A. (2008). Applied econometrics with
R. Springer Science & Business Media.
• Great reference for the applied researcher wanting to use R for
econometric analysis. Includes R basics, linear regression
model, panel data models, binary outcomes, etc.
2. Jones, A. M., Rice, N., d'Uva, T. B., & Balia, S.
(2013). Applied health economics. Routledge.
• Excellent reference for applied health economics. Examples
are all performed using STATA, but foreign package should help
here.
3. CRAN Task View: Econometrics
• A listing of the statistical models used in econometrics, as well
as the R package(s) needed to perform them. Available at:
https://cran.r-project.org/view=Econometrics
Thanks for Listening
Good luck with R!