empirical-methods

Homepage for 17-803 "Empirical Methods" at Carnegie Mellon University


Project maintained by bvasiles Hosted on GitHub Pages — Theme by mattgraham

Empirical Methods (thanks for the promo, @JoshQuicksall!)

This is the Fall 2022 offering of this course. For older versions, see here: Spring 2021Fall 2018.

Overview

Empirical methods play a key role in the design and evaluation of tools and technologies, and in testing the social and technical theories they embody. No matter what your research area is, chances are you will be conducting some empirical studies as part of your work. Are you looking to evaluate a new algorithm? New tool? Analyze (big) data? Understand what challenges practitioners face in some domain?

This course is a survey of empirical methods designed primarily for computer science PhD students, that teaches you how to go about each of these activities in a principled and rigorous way. You will learn about and get hands-on experience with a core of qualitative and quantitative empirical research methods, including interviews, qualitative coding, survey design, and many of the most useful statistical analyses of (large-scale) data, such as various forms of regression, time series analysis, and causal inference. And you will learn how to design valid studies applying and combining these methods.

There will be extensive reading with occasional student presentations about the reading in class, homework assignments, and a semester-long research project for which students must prepare in-class kickoff and final presentations as well as a final report.

After completing this course, you will:

As a side effect, this course helps you develop a healthy dose of skepticism towards scientific results in general. Does the study design really allow the authors to make certain claims? Does the analysis technique? Is the evidence provided as strong as it could be? Are there fundamental flaws and threats to validity?

Coordinates

Course Syllabus and Policies

The syllabus covers course overview and objectives, evaluation, time management, late work policy, and collaboration policy.

Learning Goals

The learning goals describe what I want students to know or be able to do by the end of the semester. I evaluate whether learning goals have been achieved through assignments, written project reports, and in-class presentations.

Schedule

Below is a preliminary schedule for Fall 2022. Each link points to a dedicated page with materials and more details. All videos are published on this YouTube channel.

Note: The schedule is subject to change and will be updated as the semester progresses.

Date Topic Notes
Tue, Aug 30 Introduction slidesvideo
Thu, Sep 1 Formulating research questions slidesvideo
Tue, Sep 6 The role of theory slidesvideo
Thu, Sep 8 Literature review slidesvideo
Tue, Sep 13 Conducting interviews slidesvideo
Thu, Sep 15 Exemplar interview papers slidesvideo
Tue, Sep 20 Qualitative data analysis slides
Thu, Sep 22 Survey design (part I)  
Tue, Sep 27 Survey design (part II)  
Thu, Sep 29 Project proposal presentations  
Tue, Oct 4 Numbers and nonsense  
Thu, Oct 6 Causal relationships  
Tue, Oct 11 Experimental design  
Thu, Oct 13 Intro to regression modeling  
Tue, Oct 18 Fall break, no class  
Thu, Oct 20 Fall break, no class  
Tue, Oct 25 Linear regression diagnostics  
Thu, Oct 27 Standardized coefficients + Mixed-effects  
Tue, Nov 1 Exemplar regression papers  
Thu, Nov 3 Simpson’s paradox + Mixed-effects  
Tue, Nov 8 Interrupted time series design  
Thu, Nov 10 Diff-in-diff + CausalImpact  
Tue, Nov 15 Mixed-methods designs  
Thu, Nov 17 Social network analysis (part I)  
Tue, Nov 22 Social network analysis (part II)  
Thu, Nov 24 Thanksgiving, no class  
Tue, Nov 29 Research vs researcher  
Thu, Dec 1 Agree to disagree  
Tue, Dec 6 Final presentations (part I)  
Thu, Dec 8 Final presentations (part II)