This is the course website for BUEC 333, Summer 2017.

- Instructor: Chris Muris
- Office: WMC 3639
- Office hour: Thursday 11:00-12:00
- No email: talk to me before or after class, in the breaks, or during my office hour.
- Syllabus
- Regrade request form
- Open labs: In WMC 2506, with Olivia:
- Fridays 10:30-13:30 (starting in week 3)
- Tuesday 14:00-17:00 (starting in week 4)

Week 1 starts on Monday, May 8, 2017. Tests are during lectures (Wednesdays 09:30-12:30). Assignment deadlines are at the start of your tutorial of that week.

Week | Topic | Event |
---|---|---|

1 | Probability Theory | |

2 | Sampling | |

3 | Randomized control trials | Hand-in 1 |

4 | Linear regression: Mechanics | |

5 | Linear regression: Inference | Hand-in 2 |

6 | Test 1 (June 14, 09:30-12:30) | Test 1 |

7 | Multiple regression: Mechanics | Hand-in 3 |

8 | No lecture | |

9 | Multiple regression: Inference | Hand-in 4 |

10 | Nonlinear regression | |

11 | Instrumental variables | Hand-in 5 |

12 | Test 2 (July 26, 09:30-12:30) | Test 2 |

13 | TBD | Hand-in 6 |

Readings and exercises are from Stock and Watson, “Introduction to Econometrics”, Third Edition (Updated). Tests will be about the readings below, and on anything discussed during the lectures or tutorials. I may post additional materials, such as slides, to Canvas.

- Topic: Introduction; Probability theory
- Concepts: Random variables and distributions; Expectation and variance; Joint distributions
- Readings: 1, 2.1, 2.2 (skip “Other measures…”), 2.3 (“Joint and marginal distributions” and “Covariance and Correlation”)
- In-class exercises: None this week.
- Tutorial exercises: No tutorials, see “Week 1” under Week 2’s tutorials.
- Lab: No labs, but see next week’s if you want a head start.

- Topic: Sampling
- Concepts: Conditional distributions and independence; Transformations of random variables; Distribution of the sample mean
- Readings: 2.3 (remainder), 2.4 (“Normal distribution”, “F distribution”), 2.5
- In-class exercises: 2.6, 2.10 (a)+(c), 2.22
- Tutorial exercises:
- Related to Week 1: 2.1, 2.2, 2.7 (a-c)
- Related to week 2: 2.3, 2.5, 2.7 (d), 2.9 (a-c)

- Lab: Follow steps 1-5 (or 6!) at swirl and go through the course “R Programming” (no TA in the lab yet)

- Topic: Statistics; Randomized control trials
- Readings: 2.6, 3.1, 3.2, 3.3, 3.4, 3.5
- In-class exercises: 2.17, 3.5, 3.10
- Tutorial exercises: 3.1, 3.2, 3.3, 3.4, 3.6, 3.7, 3.8, 3.12, 3.16, 3.17
- Lab: Finish the “R Programming” course in Swirl.

- Topic: Linear regression: Mechanics
- Readings: 2.3 (reread “Conditional distributions”), 4.1, 4.2, 4.3
- In-class exercises: 4.4, 4.5
- Tutorial exercises: 2.23, 2.27, 4.1, 4.2, 4.3, 4.9, 4.14
- Lab: Work on HI2

–> –>

- Topic: Linear regression: Inference
- Readings: 4.4, 4.5, 5.1, 5.2
- In-class exercises: 5.1, 5.3
- Tutorial exercises: 4.10, 5.2, 5.4, 5.5, 5.7

–>

- Topic: Multiple linear regression
- Readings: 6.1-6.3
- In-class exercises: 6.5 and 6.6
- Tutorial exercises: 6.1, 6.2, 6.3, 6.4, 6.9

- Topic: Inference in multiple regression
- Readings: 6.4-6.7, 7.1, 7.6
- In-class exercises: 7.7, 7.11
- Tutorial exercises: 7.1, 7.2, 7.3, 7.4, 7.5, 7.8

- Topic: Nonlinear regression; dummy variables
- Readings: Chapter 8, plus section 5.3
- In-class exercises: 8.7
- Tutorial exercises: 8.1, 8.2

- Topic: Instrumental variables
- Readings: Chapter 12
- In-class exercises: Wooldridge 15.2 and 15.7
- Tutorial exercises: 12.2, 12.9

- Topic: TBD.

This is a lecture with optional material for interested students. You can choose the topic of this lecture. Possible topics include nonparametric regression, regression discontinuity design, panel data, prediction using machine learning tools (LASSO, support vector machines, random forests), time series, …

There are 6 hand-in assignments in this course. The hand-in assignments are a mixture of theoretical questions and of empirical exercises for which you are to use a statistical computing environment called “R”.

We use R, with RStudio, because it is free, and because it works on all major operating systems (OSX, Linux, Windows).

I will occasionally demonstrate how to use R during the lectures. There are dedicated computer labs where you can get help. This page provides additional details.

Useful resources

- R can de downloaded from its official website
- A free, cross-platform, user-friendly frontend for R is RStudio. It can be downloaded here
- You will be learning R from within R, using swirl
- An excellent way to learn R is by studying “Introduction”, “Basics”, and “Linear regression” in “Applied econometrics with R” by Christian Kleiber and Achim Zeileis. It is available online through the SFU library through this link

Hand-in assignment 1 covers the material in weeks 1 and 2. You will not yet use R in this assignment. Solve the following exercises from Stock and Watson:

- 2.4 (a)
- 2.5
- 2.10(b)+(d)
- 2.19

For grading, one question will be selected at random. You will receive full points for the assignment if you provide the correct answer with a full explanation / derivation for that randomly selected question. If your assignment is not readable, you will receive 0 points.

Hand-in assignment 2 covers the material in weeks 3 and 4. Hand in your written answers to regular assignments at the start of your tutorial. Hand in the answer to the empirical exercise(s) via Canvas, at the Canvas deadline. For the empirical exercise(s), hand in an RMarkdown file `.Rmd`

and the resulting `.html`

. For more information about the hand-in format, ask your lab TA.

Exercises:

- 2.18
- 3.13
- 4.7
- Empirical exercise E3.2

Hand-in assignment 3 covers the material in weeks 4 and 5. These are both empirical exercises, so you should hand in your solutions via Canvas. Your solutions consist of an `.Rmd`

file and a `.html`

file.

Exercises:

- Empirical Exercise E4.2.
- Empirical Exercise E5.1, but skip (d) and (e)

Hand-in assignment 4 covers the material in weeks 7 and 8. Exercises: See the Canvas announcement for instructions and exercises.

Hand-in assignment 5 covers the material in weeks 9 and 10. Exercises: TBD.

Hand-in assignment 6 covers the material in week 11. Exercises: TBD.

Last year, the hand-in assignments were structured differently: there were two large hand-in assignments. You can find them below.

**Formatting instructions**

Your hand-in assignment is a written assignment.

In addition to the R code that is at the heart of your answers, you are expect to give answers to the questions in full sentences, in English.

*You are responsible for delivering a readable html document.*

Every subquestion starts with a sentence or two explaining what you are going to do. Then, there will be some code, and R output (print only what you need to answer the question). Finally, a sentence or two to interpret the findings, and answer the question. Whenever you use code from a source (internet, lab, fellow student), give proper reference.

An example of what we expect, using E4.1, follows. These instructions are in addition to the instructions in the hand-in assignment itself.

More importantly, here is a list of things *not* to do, some with an html file that demonstrates what went wrong. We will subtract points for violations. When in doubt, consult your lab TA.

- Do not print entire data frames, html. The output just looks ridiculous, and the TA has to scroll through hundreds of pages to read your answer.
- Do not use “###”" for text, only for headers html. In MS Word, you would not format your written text as “Heading 1” either. Regular text has no “#” in front of it. Use “#” only for headers for (sub-)questions. See the example above.
- Do not put you actual answer inside a code block as a comment. Try grading 10 pages of this
- Do not use a separate files for each question. You will receive 0 points. Hand in 1 (one) Rmd file, and 1 (one) html file.
- Do not hand in your assignment without checking your html file first. Your html file is the file that the TA uses to grade your assignment. If it is not readable, you will receive 0 points. Generating a readable html file is your responsibility.
- Print to screen only the variable you need for your answer. Do not print every single variable. If the question asks for a confidence interval, do not report know what the standard error and then 1.96 times the standard error is.
- Do not show a summary object with >10 components, and assume that we will find that one number that you were asked to report.
- Do not print data exploration (‘head(df)’, ‘tail(df)’, ‘str(df)’, etc.) to the screen, unless asked for by the question.

- Do not use tools you do not understand. Example: ‘plot(lm(…))’ produces a whole bunch of graphs that don’t answer the question)