FAQs
Frequently asked questions for the Undergraduate program.
Courses
It depends on your background and interests. If in doubt, please discuss this with your academic advisors and the Statistics advisors. By far the most common places to start are Stat 100, 102, or 104 (which are introductions to applied statistics, each with its own flavor), or Stat 110 (which is an introduction to probability and is a necessary foundation for most of the upper-level Statistics courses).
View the Courses Tree Diagram
Statistics is guided by a few fundamental principles but has applications in a great variety of different fields. To best serve the needs of students with a wide range of backgrounds and interests, the department has several versions of its introductory course: Stat 100, 102, and 104. These courses introduce common core material in statistical inference at an introductory level, but with different areas of application emphasized. A student can take only one of these courses for credit.
Stat 100 is a general introduction to applied statistics and data science, exploring applications in a wide range of fields. Stat 102 emphasizes applications to the life sciences and medicine. Stat 104 emphasizes applications to economics and the social sciences, and is faster-paced than Stat 100.
Stat 108 and 109a are the two main examples. Stat 108 is an introduction to programming with a statistical focus using R, while Stat 109a is an introduction to data science using Python to perform copmutation (this is the same course as CS 109a). There are also various courses in other departments, such as Economics, that rely heavily on the ideas from Stat 100, 102, and 104.
Stat 111, 123, 170, and 171 have Stat 110 as their main or only prerequisite. There are also various courses in other departments, especially Applied Mathematics, Computer Science, and Economics, that rely heavily on the ideas from Stat 110.
They are complementary courses on mathematical finance. Stat 123 focuses on derivatives, with an emphasis on interest rate derivatives and issues about trading them in real markets. Stat 170 focuses on the pricing of assets, portfolio analysis, and derivative pricing. The mathematical level of Stat 170 is somewhat higher than that of Stat 123, e.g., the course delves into topics such as Brownian motion, stochastic differential equations, and Monte Carlo methods.
Students contemplating a career in investment management, sales and trading, investment banking, hedge funds, or policy-making should consider taking Stat 123 and Econ 1723. Students aiming for a research-oriented (“quant”) position in finance, considering graduate school in finance or economics, or who want further mathematical depth in the subject, should consider taking Stat 170.
Concentrations
http://statistics.fas.harvard.edu/pages/undergraduate-statistics-general-information
has links to information and resources about the concentration. For further information, please contact the Undergraduate Student Program Administrator, Lorna Blocksma (lblocksma@fas.harvard.edu), and/or the Directors of Undergraduate Study, Joe Blitzstein (blitz@fas.harvard.edu), Mike Parzen (mparzen@stat.harvard.edu), and Kevin Rader (rader@stat.harvard.edu).
The concentration requirements can be fulfilled via any of four tracks: a General track in core statistical principles and methods, a track in Data Science, a track in Quantitative Finance, and a track in Bioinformatics and Computational Biology (BCB). The diploma and transcript do not list the student’s track; all four tracks lead to the same degree: Bachelor of Arts in Statistics.
The general track provides students with a set of methods for making decisions in the face of risk and uncertainty, and for making sense of complex data arising from many applications. The data science track explores the interface of statistics, computer science, and application areas, emphasizing topics such as prediction, machine learning, and analysis of massive data sets. The quantitative finance track gives strong preparation for many careers in finance and actuarial work, and involves a mixture of Statistics and Economics coursework. The BCB track mixes together biology, statistics, and computation, giving models and tools for studying biological data such as gene and protein sequences.
Multivariable calculus and linear algebra, to at least the level of Math 19a and 19b, are required. The most common way for concentrators to fulfill the math requirement is to take Math 21a and 21b, but there are a variety of other fairly common choices. Students considering graduate school in Statistics are strongly recommended to take additional mathematics courses, especially real analysis (e.g., Math 112).
Summer school courses do not count for concentration credit unless specifically approved by one of the Co-Directors of Undergraduate Studies in Statistics.
As of December 2015, there were 201 Statistics concentrators: 67 seniors, 71 juniors, and 63 sophomores. The concentration has grown dramatically in recent years; by way of contrast, as of December 2005 there were only 8 Statistics concentrators.
This is a brief overview of thesis writing; for more information, please see our website here. Senior theses in Statistics cover a wide range of topics, across the spectrum from applied to theoretical. Typically, senior theses are expected to have one of the following three flavors:
1. Novel statistical theory or methodology, supported by extensive mathematical and/or simulation results, along with a clear account of how the research extends or relates to previous related work.
2. An analysis of a complex data set that advances understanding in a related field, such as public health, economics, government, or genetics. Such a thesis may rely entirely on existing methods, but should give useful results and insights into an interesting applied problem.
3. An analysis of a complex data set in which new methods or modifications of published methods are required. While the thesis does not necessarily contain an extensive mathematical study of the new methods, it should contain strong plausibility arguments or simulations supporting the use of the new methods.
A good thesis is clear, readable, and well-motivated, justifying the applicability of the methods used rather than, for example, mechanically running regressions without discussing the assumptions (and whether they are plausible), performing diagnostics, and checking whether the conclusions make sense.
Secondary Field
Yes.