Course Syllabus

Designing Social Inquiry: An Introduction to Causal Inference

Professor Vanessa Baird Office Ketchum 114E; Email Vanessa.Baird@colorado.edu

https://vbaird.com/

Please come to my office hours! Tuesdays 2-3:00; Wednesdays 1:30-3:30 (also by appointment)

I often will keep office hours open for other students to listen; if you would like to speak to me privately, please let me know.

 

DSI_image1.png

There is some but very little math. You will be using a statistics program (R or Stata), but the emphasis is on avoiding the common logical fallacies in interpreting the analysis, not on learning the program itself.DSI_image2.png

 

DSI_image3.png

 

All class periods, quizzes, and midterms are meant to help teach you what you need to know to create a nearly publishable paper. In other words, your paper should contribute to human knowledge.

 

Students have presented their work from this class at national conferences (those who did got into fancy law schools as a result: Berkeley and Michigan!)

 

A student from fall 2021 and I worked on revising her paper and we will submit our article for publication in a few weeks. We are presenting it on Friday, Aug. 26th.

 

Some choices for how to do the class project

 

 

#1 choice: You can work in a group or individually

DSI_image4.png

 

# 2 choice: There are several “tracks” to take

 

B track: You can choose a relatively easy “canned” paper, where I guide you on the variables you choose. This is an option for those who just want to pass the class with a B or a C. If you choose this relatively easy way to do the final project, even with the extra credit I sometimes offer, you cannot earn more than a B. I will calculate your grade just like the other students and if you earn at least an 82 average, you earn a B. If you earn lower than that, then you receive that grade.

 

Why choose B track? You are overloaded with other time commitments. I created this track because too many students have had to drop the class because doing the final project the way I intend was too overwhelming for some.

 

A+ track: You work with me in a group; no individual option. We will meet quasi-weekly, and we have the option of using larger data sets with more variables to navigate. There will be assignments due during those meetings (as any other group should do). To stay in good standing with this group, you must hold up your part. This option is very work intensive.

 

Why choose A+ track? This is your chance to see how scholarly co-authorships work. The goal is to create a publishable paper that we can present at a national conference, with me as your co-author if you choose.

 

A track: This is the most common track. You create and critique your own hypotheses.

 

All in all, there are five options:

  1. A track with group
  2. A track individual
  3. B track with group
  4. B track individual
  5. A+ track group

 

# 3 option: You can choose Stata or R

 

I recommend Stata because it is less irritating and super easy to learn. I will provide tons of informational videos on both languages as we will not spend much time teaching how to program. If you choose Stata, then $36 for the semester rental is the cost (there are free options). No text, or other costs.

 

Grade breakdown

Paper

 

Total paper

50%

Two paper milestones

10 x2

Final paper

25

Final paper presentation (Final Exam session)

5

 

 

Exams / quizzes (all open note)

 

Total exams / quizzes

35

Two midterms

10x2

Three quizzes

5x3

 

 

 

Participation

 

Total participation

15

In-class assignments, pop quizzes, etc

15

 

 

 

Extra credit available

 

Total extra credit

13

Turning in milestones on time

2 each, 4 max

Tutoring (or being tutored)

1 point per part, 5 max

Peer reviews of final rough drafts

1-2 each based on quality

2 peer reviews, 4 max

 

Major due dates

Note that the course has a cycle where there are more exams and midterms, and other times when there are milestones due, so that it is not overwhelming all at once. Most overwhelming are milestone due dates and the end of the semester when the main project is due.

 

Quizzes and midterms run from Thursday to Friday and are taken through Canvas, open note, open book, open internet.

 

Friday, September 2

Quiz 1

 

Friday, September 23

 

Milestone 1 rough draft, optional

Friday, September 30

 

Milestone 1 due

Friday, October 14

Quiz 2

 

Friday, October 21

Midterm 1

 

Friday, October 28

 

Milestone 2 rough draft, optional

Friday, November 4

 

Milestone 2 due

Friday, November 18

Midterm 2

 

Friday, December 2

Quiz 3

 

Tuesday, December 6

 

Rough draft, optional

 

 

 

Course Outline

 

Date

Subject

Exams, quizzes, due dates

Part 1: Introduction to the logic of various kinds of inferences

Tuesday, August 23

Overview of class

 

Thursday, August 25

The logic of inference

 

Tuesday, August 30

Exogeneity

 

Thursday, September 1

Introducing the data and codebooks

Review for quiz

Quiz 1

Any hour between 2pm Thursday to noon on Friday

Friday, September 2

 

Part 2: Introduction to simple (linear) causality

Tuesday, September 6

Why we care about “y”: using ethics responsibly in social science

Topic 2: How to read an abstract, table

Assignment Bring to class: paragraph: why does your Y matter?

Thursday, September 8

Three kinds of literature reviews

Assignment: Make a causal argument, using an independent variable in your data and bring it to class. Bonus if you use a scholarly article to bolster your argument.

Tuesday, September 13

How to write a literature review efficiently

Assignment: Bring a copy of three abstracts of papers related to the three kinds of literature reviews. 

Thursday, September 15

Concepts and units: Venn diagrams, DAGs, crosstabs, and scatterplots 

For the next few class periods, readings and assignments are light.

Tuesday, September 20

Descriptive inference: From variation to co-variation and correlation: Venn diagrams, DAGs, crosstabs, and scatterplots

In-class: Lecture and practice

Thursday, September 22

Bivariate correlation

In class: lecture and practice

Friday, September 23

 

Milestone 1 rough draft optional

due on Canvas at 5pm

 

Part 3: Diagnosing

 

1) additive independence, 2) confounding, 3) mediating, and 4) moderating relationships

 

Using crosstabs and scatterplots

 

Purpose: to understand controls conceptually and prepare for Part 5

 

Tuesday, September 27

Confounding versus independent effects

 

Thursday, September 29

Mediating versus confounding

Suppression (Simpson’s paradox)

 

Friday, September 30

 

Milestone 1 due

due on Canvas at 5pm

2 points extra if on time

Tuesday, October 4

Moderators: understanding conditional relationships

 

 

Date

Subject

Exams, quizzes, due dates

Part 4: Measurement and missing data

Thursday, October 6

Conceptualization and validity

 

Tuesday, October 11

Reliability

 

Thursday, October 13

Review for quiz

Quiz 2

Any hour between 2pm Thursday to noon on Friday

Friday, October 14

 

Tuesday, October 18

Scaling: mean indexes and missing data

 

Thursday, October 20

Go over quiz, review for midterm

Midterm 1

Any two hours between 2pm Thursday to noon on Friday

Friday, October 21

 

 

Part 5: Diagnosing

 

1) additive independent effects, 2) confounding, 3) mediating, and 4) moderating relationships

 

Using multiple regression

 

Tuesday, October 25

Regression with three variables, introduction to overlapping variation

 

Thursday, October 27

Regression: confounding versus additive

 

Friday, October 28

 

Milestone 2 rough draft optional

due on Canvas at 5pm

Tuesday, November 1

Regression: mediating relationships, suppression

 

Thursday, November 3

Go over quiz 3: using regression to correct measurement error and missing data

 

Friday, November 4

 

Milestone 2 due

due on Canvas at 5pm

2 points extra if on time

Tuesday, November 8

Regression: interactions

 

Thursday, November 10

Review for midterm: Three variable regression, diagnosing the relationship

 

Friday, November 11

 

 

Tuesday, November 15

Three variable regression, diagnosing the relationship

 

Thursday, November 17

Multiple variable regression

Midterm 2

Any two hours between 2pm Thursday to noon on Friday

Friday, November 18

 

Happy Thanksgiving!

 

Date

Subject

Exams, quizzes, due dates

Tuesday, November 29

Multiple variable regression, diagnosing the relationships

 

Thursday, December 1

Review for quiz

Quiz 3

Any hour between 2pm Thursday to noon on Friday

Friday, December 2

 

Tuesday, December 6

Optional presentation: workshopping puzzles

Rough draft

optional

due on Canvas at 5pm

Thursday, December 8

Optional presentation: workshopping puzzles

 

Friday, December 9

 

 

Final exam period, TBA

 

CU Administration will announce day and time around November

Presentations

 

 

Paper due: 24 hours after final presentation (TBA)