A thesis is a more ambitious undertaking than a project. Most thesis writers within Applied Mathematics spend two semesters on their thesis work, beginning in the fall of senior year. Students typically enroll in Applied Mathematics 91r or 99r (or Economics 985, if appropriate) during each semester of their senior year. AM 99r is graded on a satisfactory/unsatisfactory basis. Some concentrators will have completed their programs of study before beginning a thesis; in situations where this is necessary, students may take AM 91r for letter-graded credit, for inclusion in Breadth section (v) of the plan of study. In the spring semester, the thesis itself may serve as the substantial paper on which the letter grade is based. Econ 985 is also letter-graded, and may be included in the Breadth section of the plan of study in place of AM 91r.
Another, somewhat uncommon option, is that a project that meets the honors modeling requirement (either through Applied Mathematics 115 or 91r) can be extended to a thesis with about one semester of work. Obviously the more time that is spent on the thesis, the more substantial the outcome, but students are encouraged to write a thesis in whatever time they have. It is an invaluable academic experience.
The thesis should make substantive use of mathematical, statistical or computational modeling, though the level of sophistication will vary as appropriate to the particular problem context. It is expected that conscientious attention will be paid to the explanatory power of mathematical modeling of the phenomena under study, going beyond data analysis to work to elucidate questions of mechanism and causation rather than mere correlation. Models should be designed to yield both understanding and testable predictions. A thesis with a suitable modeling component will automatically satisfy the English honors modeling requirement; however a thesis won't satisfy modeling Breadth section (v) unless the student also takes AM 91r or Econ 985.
Economics 985 thesis seminars are reserved for students who are writing on an economics topic. These seminars are full courses for letter-graded credit which involve additional activities beyond preparation of a thesis. They are open to Applied Mathematics concentrators with suitable background and interests.
Students wishing to enroll in AM 99r or 91r should follow the application instructions on my.harvard.
The timeline below is for students graduating in May. For off-cycle students, a similar timeline applies, offset by one semester. The thesis due date for March 2024 graduates is Friday, November 17, 2023. The thesis deadline for May 2024 graduates is Monday, April 1 at 2:00PM. Late theses are not accepted.
Mid to late August:
Students often find a thesis supervisor by this time, and work with their supervisor to identify a thesis problem. Students may enroll in Econ 985 (strongly recommended when relevant), AM 91r, or AM 99r to block out space in their schedule for the thesis.
All fourth year concentrators are contacted by the Office of Academic Programs. Those planning to submit a senior thesis are requested to supply certain information. This is the first formal interaction with the concentration about the thesis.
A tentative thesis title approved by the thesis supervisor is required by the concentration.
The student should provide the name and contact information for a recommended second reader, together with assurance that this individual has agreed to serve. Thesis readers are expected to be teaching faculty members of the Faculty of Arts and Sciences or SEAS. Exceptions to this requirement must be first approved by the Directors, Associate Director, or Assistant Director of Undergraduate Studies. For AM/Economics students writing a thesis on a mathematical economics topic for the March thesis deadline, the second reader will be chosen by the Economics Department. For AM/Economics students writing for the November deadline, the student should recommend the second reader.
On the thesis due date:
Thesis due at 2pm. Late theses are not accepted. Electronic copies in PDF format should be delivered by the student to the two readers and to firstname.lastname@example.org (which will forward to the Directors of Undergraduate Studies, Associate and Assistant Director of Undergraduate Studies) on or before that date and time. An electronic copy should also be submitted via the SEAS online submission tool on or before that date. SEAS will keep this electronic copy as a non-circulating backup and will use it to print a physical copy of the thesis to be deposited in the Harvard University Archives. During this online submission process, the student will also have the option to make the electronic copy publicly available via DASH, Harvard’s open-access repository for scholarly work.
Contemporaneously, the two readers will receive a rating sheet to be returned to the Office of Academic Programs before the beginning of the Reading Period, together with their copy of the thesis and any remarks to be transmitted to the student.
The Office of Academic Programs will send readers' comments to the student in late May, after the degree meeting to decide honors recommendations.
The thesis is evaluated by two readers, whose roles are further delineated below. The first reader is the thesis adviser. The second and reader is recommended by the student and adviser, who should secure the agreement of the individual concerned to serve in this capacity. The reader must be approved by the Directors, Associate Director, or Assistant Director of Undergraduate Studies. The second reader is normally are teaching members of the Faculty of Arts and Sciences, but other faculty members or comparable professionals will usually be approved, after being apprised of the responsibilities they are assuming. For theses in mathematical economics, the choice of the second reader is made in cooperation with the Economics department. The student and thesis adviser will be notified of the designated second reader by mid-March.
The roles of the thesis adviser and of the outside reader are somewhat different. Ideally, the adviser is a collaborator and the outside reader is an informed critics. It is customary for the adviser's report to comment not only on the document itself but also on the background and context of the entire effort, elucidating the overall accomplishments of the student. The supervisor may choose to comment on a draft of the thesis before the final document is submitted, time permitting. The outside reader is being asked to evaluate the thesis actually produced, as a prospective scientific contribution — both as to content and presentation. The reader may choose to discuss their evaluation with the student, after the fact, should that prove to be mutually convenient.
The thesis should contain an informative abstract separate from the body of the thesis. At the degree meeting, the Committee on Undergraduate Studies in Applied Mathematics will review the thesis, the reports from the two readers and the student’s academic record. The readers (and student) are told to assume that the Committee consists of technical professionals who are not necessarily conversant with the subject matter of the thesis so their reports should reflect this audience.
The length of the thesis should be as long as it needs to be to make the arguments made, but no longer!
The most recent thesis examples across all of SEAS can be found on the Harvard DASH (Digital Access to Scholarship at Harvard) repository. Search the FAS Theses and Dissertations collection for "applied mathematics" to find dozens of examples.
Note: Additional samples of old theses can be found in McKay Library. Theses awarded Hoopes' Prizes can be found in Lamont Library.
Recent thesis titles
Theses submitted in 2021
|When the Tide Goes Out: Rescue Financing and its Impact on Capital Structure, Marketable Securities and Access to Capital Markets
|Investing in India: Does Home-Field Advantage Influence Success in Indian Venture Capital?
|The Uncertain Cost of Uncertainty: An Inquiry into Exchange Rate Volatility and Bilateral Trade Flows
|Agent-Based Modeling for Optimal Economic Policy with Exogenous Shocks
|Moral Values and Judicial Decision-Making
|Disadvantaged by Design: The Long-Term Effects of Redlining on Census Tract-Level Outcomes in American Cities
|Analyzing Changes in Intra-Chromosomal Interactions Following iPSC-to-PGCLC Differentiation
|Big Names, Bigger Barriers: Firm Reputation and its Role a a Barrier to Entry
|The Impact of Anti-Veiling Legislation on Employment Outcomes and Religiosity
|Diagnosing Prevailing Trends and Disparate Impacts of COVID-19 at the County-level
|Modeling Non-homologous Protein Sequences Using a Variation on Traditional kth Order Markov Models
|High Resolution, Greater Resolve: Downscaling Solar-Induced Fluorescence Using Land Cover and Soil Moisture Data in Madagascar
|Creation of a Geospatially Explicit Framework for Evaluating the Impact of Urban Form on the Growth and Performance of Cities: Theory and Evidence from Material Infrastructure and Resource Consumption in US Cities
|Quantifying trade-offs between economic recovery and lives saved during the COVID-19 pandemic
|Forecasting the Price of Fine Art at Auction through Machine Learning: Alternative Implementations of the Hedonic Model
|Communication and Evaluation of Risk and Uncertainty in Deep Learning: a Case Study in In-Vitro Fertilization
|The Electoral Consequences of Facebook Use in the United States
|Warm Glow from Voting vs. Direct Costs: Evidence from the 2020 Election, Black Lives Matter Protests, and Mail-in Balloting
|A Hierarchical Modeling Approach to Understanding Tropospheric Ozone
|All for One, Pre-K for All: Assessing the Impact of Universal Pre-Kindergarten on Single Mothers in New York City
|A Story of Human Capital: Why the Paycheck Protection Program Had Huge Geographic Disparities
|Inferring Seasonal Cycles of Solar Induced Fluorescence for Different Vegetation Types, a Case Study in Madagascar
|The 2014 Hydrocodone Rescheduling: A Double-Edged Sword in the Fight Against Opioid Abuse
|Algorithmic Fairness in Medical Decision-Making
|Restricted Boltzmann Machines and Sequence Homology Search
|A Theoretical Three-Period Model of Early Bank Credit Line Drawdowns
|Identification and Characterization of Post-Merger Pharmaceutical Innovation
|Yuanjun (Sarah) Zeng
|In Pursuit of Alpha: Cross-Sectional Dispersion, Market Correlation, and Lagging Hedge Fund Returns
|Optimal Taxation with Imperfect Competition
|The Affordable Care Act’s Two-Legged Stool: The Effects of Eliminating the Federal Individual Mandate on Health Insurance Coverage
|Robust Rent Division
Theses submitted in 2020
|Deep Detection: Statistical Modeling and Deep Learning to Predict and Address Malaria and Anemia in Madagascar
|Migration, Skills-Biased Technical Change, and Human Capital Accumulation: Evidence from the Great Migration
|The Effect of Monetary Policy on Output According to BHANK
|Electoral Incentives, Race, and Judicial Decision Making - Evidence from one million criminal cases
|The Relative Accuracy of LMSR and CDA Prediction Markets
|Do Football Transfers Actually Improve Performance? Examining the effect of football transfers on player and club performance in Europe, the United States, and China
|Michael L. Chen
|Identifying Antibiotic Resistance in Mycobacterium tuberculosis with Machine Learning: a quick and accurate alternative to conventional diagnostics
|Ruoxi (Michelle) Chen
|Incorporating Mechanics into the Cellular Potts Model
|The Role of Disorder in Facilitating Conduction in Ionic Lattices: Insights from Monte Carlo simulations and mean-field modeling
|Electoral Composition and the Midterm Loss Cycle
|An Analysis of the Economic Forces Driving Partial Exits in Private Equity Transactions
|Expenditures, Savings and Temptation under Prohibition. Evidence from Bihar
|Riding with Charlie: Public transportation policy and its impact on businesses and road safety in Massachusetts
|Location, Location, Experience Creation: The market dynamics and financial impacts of experiential retail
|Faithful Saliency Maps: Explaining neural networks by augmenting "Competition for Pixels"
|Does More Capital Make Banks Safer? Market-based evidence form the 2011 European Banking Authority Capital Exercise
|Quantifying Uncertainty in Deep Learning
|Bi-Level Mutli-Agent Reinforcement Learning for Intervening in Intertemporal Social Dilemmas
|Characterizing the Dynamics of Movement in Drosophila melanogaster and Humans with Autistic Spectrum Disorder
|The Ex-Day Behavior of Preferred and Common Stocks
|An Analysis of Formation Disruption in Soccer
|Strict Algorithmic Scrutiny. A formulation of algorithmic procedural fairness through affirmative action jurisprudence
|Studies in Early Modern Social Networks, 1400-1750
|The Search for Dipper Stars in the Orion A Molecular Cloud Complex: Analysis of variable YSOs from NGTS photometric time-series data
|The Effect of Priming Activities on Algorithm-in-the-Loop Decision Making
|Not So Great Expectations: An analysis of expectation determinants and the predictability of expectation errors
|The Effects of Liquidity Regulation on Bank Demand for Reserves and the Federal Reserve’s Balance Sheet Policy
|Media Bias on Television and its Determinants
|Dictionary Learning for Image Style Transfer
|Constraining Adversity: Linear optimization methods for regional climate risk mitigation using solar geoengineering
|The Promise of a Cure: Measuring the welfare gain to Hemophilia A patients from gene therapy
|Moral Values and Corporate Social Responsibility
|The Effect of SoftBank Vision Fund on Venture Capital Cycles
|Find the Informed: How public pension funds choose asset managers
|Counting TB Cases in Kids: mechanistic modeling to quantify the underreporting of pediatric tuberculosis
|Lei (Cathy) Yin
|An Outlier in Zipf's World? A case study of China's city size and urban growth
|The Impact of Ride-Hailing Technology on Consumer Wages: Theory and evidence for a difference-in-differences analysis
|Housing Markets and Home-Sharing Services: An analysis on Airbnb in Boston
Theses submitted in 2019
|Does Integration of Bricks with Clicks Affect Online-Offline Price Dispersion?
|Predicting Mood in College Students: Developing a Predictive Model from Multivariate Time Series
|An Agent-Based Numerical Approach to Lenski’s Long Term Experiment
|The Price Isn't Right: The Effect of Passive Ownership on Stock Price Informativeness
|Surveying Harvard Dining: An Architectural and Mathematical Assessment of
University Dining Services and Spaces
|Information and Market Efficiency: Evidence from the Major League Baseball Betting Market
|Investigating the Causal Effects of Climate Change on Lyme Disease Incidence
|Causal Inference in Matching Markets
|A Game Theoretic Approach to the Israeli-Palestinian Conflict
|The Closed-End Fund Puzzle: New Evidence from Business Development Companies
|Applied Linear Algebra and Big Data Course Book
|The 'Taxi Driver' Process and Passive Mobile Sensing on Networks
|From a Hashtag to a Movement: Modeling Conversation Around #MeToo on Twitter
|License to Exclude: The Concentrated Costs of Occupational Licensing
|Geographic Clustering for Neighborhood Boundaries:
A Spatial Analysis of Chicago using Public Data
|Don't Hate the Players, Hate the Game: Designing a Provably Trustworthy
Stock Market in the Age of High-Frequency Trading
|Constrained Bayesian Neural Networks
|Toward a Khipu Transcription “Insistence”: A Corpus-Based Study of the Textos Andinos
|Fairness in Machine Learning
Methods for Correcting Black-Box Algorithms
|A Fork's Impact: The Reach of Mission-Driven Fine Dining
|A Quantitative Analysis of Reproductive Rights and Right to Life
Advocacy Organization Mission Statements
|Learning Strategies for Bidding in Online Advertisement Auctions with Noisy Feedback
|Administrative Assistance and the Labor Integration of Migrant Roma
Informal Settlement Residents: Results from the MOUS Program in Grenoble
|Genome-Wide Analysis of NET-seq Data to Understand RNA Polymerase II Behaviour
Around Transcription Factor Binding Sites
|Effects of Job Displacement on Prescription Opiate Demand:
Evidence from the Medical Expenditure Panel Survey
|Determinants of Volatility: Calling it Quits on the VIX
|The Bayesian Synthetic Control: A Probabilistic Framework for Counterfactual Estimation
in the Social Sciences
The Effect of Hospital Market Consolidation on Health Care Quality and Wages
|Subsidized Contraception and Teen Fertility: The Effects of Medicaid Family
Planning Program Eligibility Expansions on the Teen Birth Rate
and 12th Grade Dropout Rate
|One Step Back, Two Steps Forward
A Machine Learning-Powered Options Trading Strategy
|Tip Nudging: A 3-Part Experimental Analysis of Influencing Tips
through Technological Default Menus
|Real-Time Scheduling for a Variable-Route Bus System
Theses submitted in 2018
|Designing Prediction Markets with Informational Substitutes
|An Electronic Health Record Interface for Clinical Use in Inpatient Settings
|Labour-market Impacts of Industrial Robots: Evidence from US occupations
|HIV Dynamics: Predicting Treatment Efficacy and Rebound
|The Impact of Medicaid Expansion on the Quality of Care in the United States
|Gaming the Blockchain: How Subversive Mining Strategies Can be Profitable
|Differentially Private Inference over Exponential Random Graph Models
|A Comparison of Sketch-Based Cardinality Estimation Algorithms
|Weighting the Dice: A demographic based look at slot machine profits.
|Automation-Driven Job Displacement and Subsequent Employment Outcomes: Individual-Level Evidence
|Markov-based model of cervicovaginal bacterial dynamics predicts community equilibrium states in young South African women
|Money Illusion and Attitudes Toward Trade
|Tweeting and Dialing: Quantifying Influence in Online and Offline Protest
|What is the role of social value on complex decision-making tasks?
|Pointing Isn't Always Rude: Using Pointer Networks to Improve Word Prediction in a Language Model
|Mumps at Harvard: Modeling the Spread of Disease on College Campuses
|Adaptation in Corn Production Sensitivity to Extreme Weather Events: the Effect of Weather Calculation Procedures
|Testing for Racial Discrimination in Mortgage Lending via Mortgage Default Equations
|Defuse the News: Predicting Information Veracity and Bias in Social Networks via Content-Blind Learning
|Fertilizer Supply Chain in Ethiopia
|Wrong Turns and Right Tails: Identifying Fraudulent Detours in New York City Taxi Rides
|I'm Not Sexist, I Just Prefer My Companies to Be: An Experiment in Discrimination by Customers
|Estimating the Causal Effect of Chloroquine Treatment on Mortality in Malaria Patients Using Marginal Structural Models
|Urbanization and Social Networks: Evidence from a Development Context
|The Role of Merit-Based Scholarships on Economic Mobility in the United States
|Smarter Prescriptions: Personalized Radiation Treatment Using Self-biomarkers
|Does Network Influence in Global Value Chains Lead to Innovation?
Senior Thesis Submission Information for A.B. Programs
Senior A.B. theses are submitted to SEAS and made accessible via the Harvard University Archives and optionally via DASH (Digital Access to Scholarship at Harvard), Harvard's open-access repository for scholarly work.
In addition to submitting to the department and thesis advisors & readers, each SEAS senior thesis writer will use an online submission system to submit an electronic copy of their senior thesis to SEAS; this electronic copy will be kept at SEAS as a non-circulating backup. Please note that the thesis won't be published until close to or after the degree date. During this submission process, the student will also have the option to make the electronic copy publicly available via DASH. Basic document information (e.g., author name, thesis title, degree date, abstract) will also be collected via the submission system; this document information will be available in HOLLIS, the Harvard Library catalog, and DASH (though the thesis itself will be available in DASH only if the student opts to allow this). Students can also make code or data for senior thesis work available. They can do this by posting the data to the Harvard Dataverse or including the code as a supplementary file in the DASH repository when submitting their thesis in the SEAS online submission system.
Whether or not a student opts to make the thesis available through DASH, SEAS will provide an electronic record copy of the thesis to the Harvard University Archives. The Archives may make this record copy of the thesis accessible to researchers in the Archives reading room via a secure workstation or by providing a paper copy for use only in the reading room. Per University policy, for a period of five years after the acceptance of a thesis, the Archives will require an author’s written permission before permitting researchers to create or request a copy of any thesis in whole or in part. Students who wish to place additional restrictions on the record copy in the Archives must contact the Archives directly, independent of the online submission system.
Students interested in commercializing ideas in their theses may wish to consult Dr. Fawwaz Habbal, Senior Lecturer on Applied Physics, about patent protection. See Harvard's policy for information about ownership of software written as part of academic work.