Michael Ekstrand

Information for Prospective Students

A student walking between library shelves.
Photo by Redd F on Unsplash

I am often looking for Ph.D. students looking to study information access systems (recommender systems, search engines, AI for information access tasks, etc.) with a particular eye to evaluation and research methods, social impact, and addressing societally-important information challenges.

I am also looking for first-year IS or CS undergraduate students interested in working with me through the STAR Scholars program, and am interested in B.S. and M.S. students who wish to work on research with me for independent study credit. See Types of Opportunities later on this page for more information.

On this page, I have collected links to resources on my site and elsewhere that you may find useful in determining whether you are interested in working with me and navigating the process of application and admission. You may also want to see my FAQ or watch my latest research seminar video.

On This Page

See also my FAQ. For another perspective, I highly recommend reading and watching Casey Fiesler’s Ph.D. advice, and Shiri Dori-Hacohen’s FAQ also has useful info (some points specific to her lab, others more broadly applicable).

Types of Opportunities

There are several ways for students to participate in my research:

  • I regularly supervise Ph.D. students in our program. Most of the rest of this page is for prospective Ph.D. students.
  • The STAR Scholars program supports first-year undergraduate students to work on research in the summer. Most summers, I am interested in having STAR scholars work with my research group.
  • Research-based independent study courses for M.S. or B.S. students. I am usually interested in having a few students pursuing this route. It can be especially productive if you have first taken one of my courses, like Recommender Systems, and want to do an independent study to develop your ideas more fully into publishable research.
  • M.S. capstone projects (on occasion).
  • I sometimes hire coop students (either B.S. or M.S.) to work with me on my research. I do not have any current openings of this type, but will update this page when I do.

For prospective Ph.D. students, read the rest of this page and submit an application. For current Drexel students interested in the other types of research opportunities I have listed, email me to discuss your interest.

Am I A Good Fit?

If you:

  • want to make information systems (with or without AI) help — not hurt — their users, society, and the particular people they effect
  • are interested in studying the effects of real systems in real contexts
  • have computing background (roughly the equivalent of a computer science minor or more, or comparable experience; students who don’t know how to program will have difficulty getting up to speed in my group at this time); most of our research has significant computational and statistical elements
  • want to work on a mix of sponsor-provided and self-determined projects in a collaborative setting
  • believe a Ph.D. in Information Science will advance your professional and/or personal goals

then I would love to see an application from you!

Ongoing Research

My work is grounded in particular applications rather than technologies such as deep neural networks. With my students and collaborators, I work to understand how information systems interact with both individual and social human interests, such as the need for relevant information and the need for fair systems that don’t reproduce society’s historical and ongoing patterns of discrimination and oppression. This work includes both building new systems and developing experimental methods, metrics, etc. to measure system behavior and effects. If you’d like a taste of how this works out in practice, you can read a few of my papers, including:

SIGIRec22
2022

Amifa Raj and Michael D. Ekstrand. 2022. Fire Dragon and Unicorn Princess: Gender Stereotypes and Children’s Products in Search Engine Responses. In SIGIR eCom ’22. DOI 10.48550/arXiv.2206.13747. arXiv:2206.13747 [cs.IR]. Cited 10 times. Cited 5 times.

UMUAI21
2021

Michael D. Ekstrand and Daniel Kluver. 2021. Exploring Author Gender in Book Rating and Recommendation. User Modeling and User-Adapted Interaction 31(3) (February 4th, 2021), 377–420. DOI 10.1007/s11257-020-09284-2. arXiv:1808.07586v2. NSF PAR 10218853. Impact factor: 4.412. Cited 198 times (shared with RecSys18◊). Cited 106 times (shared with RecSys18◊).

AJIM20
2020

Michael D. Ekstrand, Katherine Landau Wright, and Maria Soledad Pera. 2020. Enhancing Classroom Instruction with Online News. Aslib Journal of Information Management 72(5) (November 17th, 2020; online June 14th, 2020), 725–744. DOI 10.1108/AJIM-11-2019-0309. Impact factor: 1.903. Cited 19 times. Cited 11 times.

You may also want to see a topical view of my research or my complete publication list; I also post my talks, which can sometimes be more approachable than the papers.

One of my ongoing projects is the POPROX research infrastructure project to create a personalized news recommendation platform that will enable academic researchers to study news recommendation and related problems with real users.

Working With Me

There are a few resources useful for understanding what’s involved in doing research under my supervision:

My primary goal in advising is to help each of my students figure out what they want to do, and support them on the steps to get there. I work on a variety of projects related to recommender systems, information retrieval, and social impact. Most of my work is connected somehow to the human impact of information access, but I occasionally have other projects. Specific details depend on student interests, available funds, and current collaborations.

Applying

A clipboard with a form saying "Application", next to a laptop.
Photo by Markus Winkler on Unsplash.

If you want to earn your Ph.D. under my supervision, apply to the Ph.D. in Information Science and mention me in your application. You do not need to contact me before applying. The Ph.D. in IS program only admits for fall quarter; there is no spring admission. Further details about application process and requirements are available here. Note that you do not need to obtain an M.S. first — you can apply directly to the Ph.D. program.

Funding decisions are made along with admission, and full-time students receive a funding offer with their admission. You do not need to apply separately for funding.

After reviewing applications that are relevant to my work and meet the program’s requirements for potential admission, I will usually schedule interviews with the most promising candidates to make final decisions.

The Statement of Purpose

Statements of purpose are a weird document, and unfortunately there is a lot of very bad advice out there about how to write them. For good advice, I recommend reading Vijay Chidambaram’s Twitter thread.

I also have a few specific suggestions:

  • Focus on the future. The SOP is a statement of purpose, not history. The emphasis should be on what you hope to achieve in and through your graduate career. It should answer a few questions:

    • What do you hope to accomplish in your Ph.D.? It’s fine to change research areas later, or not be entirely clear on your desired research area, but the committee should be able to see “if we admit this student, what might they do?”

    • How will obtaining a Ph.D. advance your life or career goals? Remember that a Ph.D. is a research degree — why do you want such a degree?

    • What prepares you to succeed in the Ph.D.? Examples of existing research or coursework, particular skills or interests, etc. can be evidence here.

    • Why do you want to study with this advisor at this university?

    Your own background can be useful as evidence or context for some of these points, and what you will contribute to the program, but the primary focus of the essay should be on your purpose in the program.

  • Be specific. What particular domains or applications interest you? Are there problems you might be interested in solving? This is particularly important for working with me on research — my work is highly applied and connected to particular applications. While we do make use of a lot of different machine learning, data science, etc. techniques, the focus is usually on the problems with these tools as a means to an end. If there is a particular technology or field that fascinates you, such as deep learning or natural language processing, why? What problems do you see it useful for solving?

  • Be consistent. To the extent that you state specific research goals, be consistent. It’s fine to not entirely know what you want to do; however, if you do state a specific goal, such as wireless network security, and then list potential advisors who only work on something completely different like compiler optimizations for machine learning, it looks disconnected.

    I am also in the Information Science department and advise students in that Ph.D., which is distinct from the Computer Science department. Your SOP and application materials should be consistent with the program to which you are applying — there are many computer scientists (like myself!) in information science, so a CS background is good and useful, but your SOP should talk about why you want to join the IS program.

There are also a couple of problems I see that I would recommend avoiding:

  • Don’t plagiarize. Just don’t. Ever. This includes taking SOP examples or templates and plugging in your target institution and research keywords, even if those SOPs are published for the purpose of being examples in books about getting in to grad school. Mediocre text you wrote yourself is better than good text you copied.

  • Don’t flatter. Say why you want to pursue a Ph.D., what you think you might want to do, what qualifies you for the work, and why you want to go this institution. Say specific things about how the program will fit your goals; general statements about rankings and reputations are not helpful. Such statements backfire in two ways: if they are true, the people reading your application don’t need you to tell them and are in a better position to judge impact and prestige. If they are not true, they demonstrate a lack of critical thinking that is a red flag.

    Many example SOPs in the books I have seen about how to get into graduate school are full of flattery. I consider them to be bad examples.

Contacting Faculty

You do not need to contact me in advance when applying to work with me for your Ph.D. — no permission is required to list me as a prospective advisor. I also do not take Ph.D. applications by e-mail; all applications need to go through the application system. If you do want to e-mail me (or other faculty) in advance of your application, I recommend Casey Fiesler’s video on contacting prospective advisors.

I don’t mind e-mails in advance; these e-mails should be personalized and they should be specific: what information are you looking to convey, or obtain? They should also mention your application status: have you already applied, are you planning to apply, are you deciding whether to apply? And they should not include trackers such as MailTrack.

E-mailing me in advance will have minimal, if any, impact on your application; it might help me identify which applications to look at, but if I have any openings I will review all applications that express interest in working with me and meet the program’s requirements for potential admission.

I am happy to answer specific questions about our program, about working with my group, or about research and graduate school in general. It’s also fine send a brief e-mail just letting me know you applied; I’ll file it away and make sure I look at your application. I do recommend checking my FAQ before e-mailing me.

Some things you might want to consider e-mailing about:

  • Specific thoughts or questions about one of my papers.
  • Questions about how the application process works, or a question that isn’t already answered on my website about how I work with students.
  • What projects I am recruiting students to work on (although for Fall 2024, this information is already on this page).
  • A specific idea of your own you want to work on, to see if it is something that would fit with my research portfolio.
  • Skills you should work on to be prepared for research with my group (but see my skills list first).

Potential Poor Fits

No one is the perfect advisor for everyone. A few reasons why I might not be a good fit to be your Ph.D. advisor:

  • Your primary interest is in advancing specific AI designs or models. My research agenda treats particular models as tools to be applied to particular human ends, not as the goal in themselves. Strong interest and background in AI can be good preparation for working with me, but if you want the main outcome of your research to be improved deep learning models, then there are probably others who would be better.
  • Your goal is to maximize adoption of a particular technology. My approach looks for whether a technology is useful, and does not assume that it is.
  • You specifically want a degree in computer science. I am in the information science department, and am not currently taking CS Ph.D. students. If you are admitted to Drexel’s Ph.D. in CS program, I may be able to be on your committee if you want, but cannot be your major advisor.