Week 1: Science from Particles to People

DSAN 5650: Causal Inference for Computational Social Science
Summer 2025, Georgetown University

Jeff Jacobs

jj1088@georgetown.edu

Wednesday, May 21, 2025

Schedule

Today’s Planned Schedule:

	Start	End	Topic
Lecture	6:30pm	7:00pm	Quick Hello Hi Everyone Style Intro →
	7:00pm	7:25pm	science \(\leadsto\) Social Science “Phase Transition” →
	7:25pm	7:50pm	Motivating Examples I: Social Science →
Break!	7:50pm	8:00pm
	8:00pm	8:30pm	Motivating Examples II: Causal Inference →
	8:30pm	9:00pm	Course Logistics →

\[ \DeclareMathOperator*{\argmax}{argmax} \DeclareMathOperator*{\argmin}{argmin} \newcommand{\bigexp}[1]{\exp\mkern-4mu\left[ #1 \right]} \newcommand{\bigexpect}[1]{\mathbb{E}\mkern-4mu \left[ #1 \right]} \newcommand{\definedas}{\overset{\small\text{def}}{=}} \newcommand{\definedalign}{\overset{\phantom{\text{defn}}}{=}} \newcommand{\eqeventual}{\overset{\text{eventually}}{=}} \newcommand{\Err}{\text{Err}} \newcommand{\expect}[1]{\mathbb{E}[#1]} \newcommand{\expectsq}[1]{\mathbb{E}^2[#1]} \newcommand{\fw}[1]{\texttt{#1}} \newcommand{\given}{\mid} \newcommand{\green}[1]{\color{green}{#1}} \newcommand{\heads}{\outcome{heads}} \newcommand{\iid}{\overset{\text{\small{iid}}}{\sim}} \newcommand{\lik}{\mathcal{L}} \newcommand{\loglik}{\ell} \DeclareMathOperator*{\maximize}{maximize} \DeclareMathOperator*{\minimize}{minimize} \newcommand{\mle}{\textsf{ML}} \newcommand{\nimplies}{\;\not\!\!\!\!\implies} \newcommand{\pedge}[2]{\require{enclose}\enclose{circle}{~{#1}~} \rightarrow \; \enclose{circle}{\kern.01em {#2}~\kern.01em}} \newcommand{\pnode}[1]{\require{enclose}\enclose{circle}{~{#1}~}} \newcommand{\orange}[1]{\color{orange}{#1}} \newcommand{\outcome}[1]{\textsf{#1}} \newcommand{\param}[1]{{\color{purple} #1}} \newcommand{\pgsamplespace}{\{\green{1},\green{2},\green{3},\purp{4},\purp{5},\purp{6}\}} \newcommand{\prob}[1]{P\left( #1 \right)} \newcommand{\purp}[1]{\color{purple}{#1}} \newcommand{\sign}{\text{Sign}} \newcommand{\spacecap}{\; \cap \;} \newcommand{\spacewedge}{\; \wedge \;} \newcommand{\tails}{\outcome{tails}} \newcommand{\Var}[1]{\text{Var}[#1]} \newcommand{\bigVar}[1]{\text{Var}\mkern-4mu \left[ #1 \right]} \]

Who Am I? Why Is Georgetown Having Me Teach This?

Prof. Jeff Introduction!

• Born in NW DC → high school in Rockville, MD
• University of Maryland: Computer Science, Math, Econ

The World Outside of DC

Studied abroad in Beijing (Peking University/北大) → internship with Huawei in Hong Kong (HKUST)

Stanford, MS in Computer Science

Research Economist, UC Berkeley

Columbia, PhD in Political Economy

Why Is Georgetown Having Me Teach This?

• Quanty things \(\leadsto\) PhD in Political Economy
• PhD exam major: Political Philosophy
• PhD exam minor: International Relations
• PhD exam paper: “How to Do Things with Translations”
• Game-changing Research Fellowships at… (Nothing was the same -drake):
• Santa Fe Institute: dedicated to the multidisciplinary study of complex systems: physical, computational, biological, social

Figure 1: Years spent questing in dungeons of academia

• Centre for the Study of the History of Political Thought, Queen Mary University of London (QMUL): New approaches to the history of political thought [Quentin Skinner, “Cambridge School”] have changed how we study ideas from the past and their relevance to contemporary politics. The focus of the Centre is to explore [ts].

Dissertation (NLP x History)

“Our Word is Our Weapon”: Text-Analyzing Wars of Ideas from the French Revolution to the First Intifada

(IR Part) Wars of Ideas I: Cold War

• Cold War arms shipments (SIPRI) vs. propaganda (Печать СССР): here, to 🇪🇹

(Middle East Part) Wars of Ideas II: First Intifada, 1987-1993

Research Nowadays

• Most cited paper: “Monopsony in Online Labor Markets” (Uses Double-Debiased ML for Causal Inference!)
• Most recent paper: “Operationalizing Freedom as Non-Domination in the Labor Market” (Cambridge U Press)
• Rarely cited paper but often-thought-about obsession: “How To Do Things With Translations”
• Related thing you can buy in a bookstore: [Editorial board for new translation of] Capital, Vol. 1 by Karl Marx (Princeton U Press)

But Now… Teaching!

• Growth mindset: “I can’t do this” \(\leadsto\) “I can’t do this yet!”
• Think of anything you’re able to do… There was a point in life when you didn’t know how to do it! What happened? Your brain established and/or rearranged neural pathways as you struggled with it!
• How does this neural re-arrangement work? One “cheatcode” is spaced repetition:

Image source

Lil Wayne on Spaced Repetition

Figure 2: From The Carter (Documentary)

Maria Montessori on Your Final Project

Our teaching should be governed, not by a desire to make students learn things, but by the endeavor to keep burning within them that light which is called curiosity. (Montessori 1916)¹

• To this end: your final project is to explore some potential causal linkage from \(X\) to \(Y\) (more later!)
• The sole requirement is: sufficient curiosity to serve as fuel for your journey from associational world to causal world

1. “Couldn’t do classes, them lessons used to bore me, I was tryna school heads like Waldorf and Montessori” -Bar dropped by local rapper one day in freestyle that I never forgot

The Science \(\leadsto\) Social Science “Phase Transition”

Coarse-Graining Units of Observation

Field	Example Unit of Observation
Physics	Particle	[Particle = Fine-Graining of Molecules]
Chemistry	Molecule = \(\cup\)(Particles)	[Molecule = Coarse-Graining of Particles]
Biology	Cell = \(\cup\)(Molecules)
Neuro/Cog Sci	Brain = \(\cup\)(Cells)
Human Physiology	Body = Brain \(\cup\) Other Organs
↑ Science    🧐 something happens here… 🤔    ↓ Social Science
Anthropology	Human-Relational System (e.g., Kinship) = \(\cup\)(Brains) \(\times\) Natural Environment
Economics	Economy = Specific relational system of exchange w.r.t. scarce resources
Political Economy	Economy-Context = Economic Exchange \(\cap\) Relational Power
Sociology	Society = \(\cup\)(Relational Systems: Kinship, Friendship, Authority, Power, Violence)
History	Longue-Durée = Evolution of Societies over Time and Space

Homo sapiens/Homo arbitratus/Homo mischievous

• Latin sapiens denotes being “discerning” or “wise”

• But… technically nothing stops us from choosing to be “unwise” whenever we’d like… bc free will
• \(\Rightarrow\) For this class, humans are Homo arbitratus: arbitratus denotes choosing what to do, after we’ve sapiently “discerned”/thought about it

Laws of Physics vs. “Laws” of Social Science

• If we tell an inanimate object that we’ve discovered a law saying that it will accelerate towards Earth at \(9.8~\textrm{m}/\textrm{s}^2\)
  • …It will likely¹ still accelerate towards Earth at \(9.8~\textrm{m}/\textrm{s}^2\)

• If we tell a human we’ve discovered a law saying they will quack like a duck at 7:30pm EDT every Wednesday
  • …They can utilize their free will to violate this “law”

1. Obligatory quantum mechanics footnote: human agency [maybe] plays a role, in a quirky way, in physics at tiny subatomic scales… but once we coarse grain to atoms (and avoid speed of light), Newton’s Laws accurate to many decimals 🤯

Strangely-Relevant CS Topic: The Halting Problem

• Kurt Gödel \(\leftrightarrow\) Alan Turing: Entscheidungsproblem
• Theorem: It is not possible to write a computer program \(P(x)\) that detects whether or not a computer program \(x\) will eventually halt (as opposed to, e.g., looping forever)
• Proof: Assume \(P(x)\) is possible to write. Run it on mischievous_program.py. Infinite contradiction loop.

halting_problem_solver.py

def will_it_halt(program, input):
  # Put code you think will work here
  # Return True if program will halt, False otherwise

mischievous_program.py

def my_mischievous_function(input):
  if will_it_halt(my_mischevious_function, input):
    while True: pass # Infinite loop
  else:
    return # Halt

Input→ ↓Program	0	1	2	\(\cdots\)
0	Halt	Loop	Loop	\(\cdots\)
1	Loop	Loop	Halt	\(\cdots\)
2	Loop	Loop	Loop	\(\cdots\)
\(\vdots\)	\(\vdots\)	\(\vdots\)	\(\vdots\)	\(\mathbf{\ddots}\)
	Loop	Halt	Halt	\(\mathbf{\cdots}\)

The Takeaway: [Bayesian] Humility!

• Social science, with “science” used in the same sense as for physics, may be a quixotic endeavor¹
• Instead, we’ll do social science, where we use data to…
• Infer tendencies: \(\mathcal{H}\) = «\(X\) tends to cause \(Y\)»
• With some degree of veracity: \(\Pr(\mathcal{H}) \approx 0.7\)
• Construct models that we can update with new evidence: Bayes’ rule! \(\Pr(\mathcal{H} \mid E) = \frac{\Pr(E \mid \mathcal{H} ) \Pr(\mathcal{H})}{\Pr(E)} \approx 0.8\)
• Pls notice “slippage” between aleatory probability within \(\mathcal{H}\) (“tends to”) vs. epistemic probability “outside of”, talking about \(\mathcal{H}\) (“I’m 70% confident about \(\mathcal{H}\)”)

1. At least, for the time being… BUT see Sperber (1996), which will come up later

Motivation I: Humble-Bayesian Social Science

The Logic of Violence in Civil War

Kalyvas (2006)

Matching Estimators

Image Source

Case Study: Military Inequality \(\leadsto\) Military Success

• Lyall (2020): “Treating certain ethnic groups as second-class citizens […] leads victimized soldiers to subvert military authorities once war begins. The higher an army’s inequality, the greater its rates of desertion, side-switching, and casualties”

Matching constructs pairs of belligerents that are similar across a wide range of traits thought to dictate battlefield performance but that vary in levels of prewar inequality. The more similar the belligerents, the better our estimate of inequality’s effects, as all other traits are shared and thus cannot explain observed differences in performance, helping assess how battlefield performance would have improved (declined) if the belligerent had a lower (higher) level of prewar inequality.

Since [non-matched] cases are dropped […] selected cases are more representative of average belligerents/wars than outliers with few or no matches, [providing] surer ground for testing generalizability of the book’s claims than focusing solely on canonical but unrepresentative usual suspects (Germany, the United States, Israel)

Does Inequality Cause Poor Military Performance?

Covariates	Sultanate of Morocco Spanish-Moroccan War, 1859-60	Khanate of Kokand War with Russia, 1864-65
\(X\): Military Inequality	Low (0.01)	Extreme (0.70)
\(\mathbf{Z}\): Matched Covariates:
Initial relative power	66%	66%
Total fielded force	55,000	50,000
Regime type	Absolutist Monarchy (−6)	Absolute Monarchy (−7)
Distance from capital	208km	265km
Standing army	Yes	Yes
Composite military	Yes	Yes
Initiator	No	No
Joiner	No	No
Democratic opponent	No	No
Great Power	No	No
Civil war	No	No
Combined arms	Yes	Yes
Doctrine	Offensive	Offensive
Superior weapons	No	No
Fortifications	Yes	Yes
Foreign advisors	Yes	Yes
Terrain	Semiarid coastal plain	Semiarid grassland plain
Topography	Rugged	Rugged
War duration	126 days	378 days
Recent war history w/opp	Yes	Yes
Facing colonizer	Yes	Yes
Identity dimension	Sunni Islam/Christian	Sunni Islam/Christian
New leader	Yes	Yes
Population	8–8.5 million	5–6 million
Ethnoling fractionalization (ELF)	High	High
Civ-mil relations	Ruler as commander	Ruler as commander
\(Y\): Battlefield Performance:
Loss-exchange ratio	0.43	0.02
Mass desertion	No	Yes
Mass defection	No	No
Fratricidal violence	No	Yes

Particularly Fun Non-“Standard” Examples

• DeDeo, French Revolution
• Grimmer, Mirrors for Sultans and Princes

Course Logistics

JupyterHub

https://guhub.io

Image source

References

Kalyvas, Stathis N. 2006. The Logic of Violence in Civil War. Cambridge University Press.

Lyall, Jason. 2020. Divided Armies: Inequality and Battlefield Performance in Modern War. Princeton University Press.

Montessori, Maria. 1916. Spontaneous Activity in Education: A Basic Guide to the Montessori Methods of Learning in the Classroom. Lulu Press.

Sperber, Dan. 1996. Explaining Culture: A Naturalistic Approach. Cambridge: Blackwell.