Data Science Program Accreditation

28 June 2022

From the desk of Catherine Cramer

(read more about Catherine Cramer)

On February 23, 2022 ADSA hosted the last of the 2021 Virtual Meeting sessions, with an installment on the important and complicated issue of Data Science Program Accreditation [video link]. As all of us in the academic data science community are well aware, data science programs have been proliferating, and with their emergence comes a need for some programs to be accredited. The process of accrediting data science programs is now in development, and was the subject of this panel discussion. 

What is accreditation? From the ABET website

Accreditation is a periodic review process to determine if educational programs meet defined standards of quality. In the U.S., accreditation is voluntary, decentralized and carried out by many non-governmental, non-profit organizations. The process typically involves an external quality review by a team of professional experts from academia and industry. Additional volunteer experts review team findings and make the accreditation determinations while ensuring consistency of the decisions. 

University gates with students walking toward the camera

Bryan Y.W. Shin, CC BY-SA 3.0 via Wikimedia Commons

Some additional background (and spelling out of acronyms) might be helpful. As of April 21, 2021, the American Statistical Association (ASA) became a full member of the Computing Sciences Accreditation Board (CSAB), joining the world’s two largest professional and technical societies for computing - the Association for Computing Machinery (ACM) and the Institute of Electrical and Electronics Engineers (IEEE) Computer Society (IEEE-CS). As the lead ABET (Accreditation Board for Engineering and Technology) member society for computing, CSAB is responsible for developing accreditation criteria for a number of programs, including Data Science (DS). Programs in data science and data analytics, which have previously not fallen under ABET’s accreditation umbrella, will soon be eligible for ABET accreditation. There are two distinct ABET groups working on data science program accreditation criteria - the Applied and Natural Sciences Accreditation Commission (ANSAC) and Computing Accreditation Commission (CAC). (Confused? Hopefully you will be less so after reading this.) In this session Donna LaLonde (ASA Director of Strategic Initiatives and Outreach), Dave Hunter (Penn State Professor of Statistics) and Ron Wasserstein (ASA Executive Director) provide an update on this process and the drafting of criteria, and answer many questions.

Donna LaLonde started off with general descriptions of the various associations. The ASA is the world’s largest community of statisticians and is the second-oldest continuously operating professional association in the US, founded in Boston in 1839. As data science as a field has grown, so has its representation within ASA. There are now data science sections and interest groups within ASA as well as education programs, a data challenge and data fest. ABET accredits programs in engineering, applied and natural sciences, and is the final decision-maker on accreditation. There are over 4000 accredited programs, mostly in the US, and ABET contains four commissions for accreditation standards, as well as 35 member societies that provide expertise on criteria.  One of those member societies is CSAB, lead ABET member society of accreditation of degree programs in CS, cybersecurity, data science, information systems, information security, and software engineering and includes members from the ACM, ASA, and IEEE-CS. (Dave Hunter pointed out that ASA joined CSAB because data science accreditation criteria are being established.)

Dave Hunter discussed the current status of DS criteria. ABET and member society CSAB formed a joint task force of professional society experts to investigate data science accreditation. The task force found that a majority of employers and existing academic programs focus on the computing aspects of data science. Criteria for such programs were developed, presented at several public forums and revised in response to public feedback. On July 26, 2020, the ABET Computing Area Delegation approved the release of Program Criteria for Data Science and similarly named computing programs for public review and comment through June 15, 2022. This gives the Computing Accreditation Commission an opportunity to get constituent feedback while ensuring the proposed changes, revisions and criteria are as clear, effective and inclusive as possible. Pilot accreditation will took place in the 2021-22 accreditation cycle. The task force continues to investigate programs that have other academic disciplinary emphases beyond computing. Engagement with the public on non-computing focused criteria will be part of this process. The goal is to make accreditation available for all qualified programs, regardless of a specific program’s data science focus.

Dave pointed out that there are both general and program criteria, and right now two separate commissions are in the process of approving DS criteria, with a goal to bring them in line with each other. (The actual curriculum is the same between ANSAC and CAC, and Dave made sure to point out that the curriculum criteria had “a lot of influence from statisticians”, saying further that “ANSAC joined the party a little late but stats have an important role”.) ANSAC - Applied and Natural Sciences Accreditation Commission – looks at data science under natural sciences, not computing, while CAC – Computing Accreditation Commission – got a one-year head start and has already advanced a set of DS criteria that will go into effect in 2023-24; pilot programs are being run and program evaluators are observing the pilot courses. These courses include a DS-specific student outcome: that graduates of the program will have an ability to apply DS theory, techniques and tools throughout the DS lifecycle and employ the resulting knowledge to satisfy stakeholders’ needs. 

The panel then answered many questions from attendees. Among the answers:

Colorful question marks

  • DS criteria only apply to 4-year undergraduate courses – for now. They are working toward the accreditation of Associates and Masters programs. 

  • To become a program evaluator you need to first take ABET training, then you are assigned as an observer. “It’s a lengthy process but you will learn what satisfies the criteria.” 

  • Look at the general criteria to see what is the best match for your program.  If there’s another program on your campus that is already accredited by CAC then you can get some expertise from them.


  • Can CS, statistics, domain knowledge, ethical training, soft skills, diversity training, etc. all be reasonably covered in an undergraduate degree program? Pilot programs will give us some answers. “A solid DS program could do all of these things and it’s good to be aspirational in setting the bar, but time will tell if the bar is too high.” 

  • If your program meets the criteria then it does not need to be called a “DS Major”. This is why the word “program” is used. “It needs to be a 4-year undergraduate program, but if it’s an option within a different major or a minor that meets the criteria then it’s a program that can be accredited.” 

  • A DS program within an engineering school would still need to be accredited as a DS program.

  • The accreditation criteria are a baseline, and programs are free to make any additions that they feel will make their program stronger.

  • Accreditation is good for 5 years then you have to reapply, ensuring continuous improvement.

  • In terms of where to start when developing a data science program: look around at what other schools are doing and find the approaches that most closely fit your school. No one can replicate one particular program at another school – each school has its own eccentricities — so there isn’t one generic approach.
Group of people looking at charts on paper

Dave concluded by saying, “There have been many decades of discussion about DS in the stats community and we’re trying to bring that expertise into the process. It’s better to be part of this ABET process. There might be a discrepancy at some point between what ABET is saying and what the stats community is saying but these differences will likely be ironed out over the years.” 

He pointed out that the community has to get to consensus on questions such as, what does self-study look like? There is a rapid growth of programs happening, with so much that still remains to be worked out, and be worked out in a community that is itself rapidly evolving – bringing the “building the plane in mid-flight” image to mind. “We have to figure it out as we go along,” said Dave. “It will be a several year-long process, and we can't let perfect be the enemy of good.” 

The panel reiterated that if you have interest in becoming a program evaluator to contact Scott Murray at rsmurray1@gmail.com and that you are encouraged to provide comments on the first set of criteria, and these comments will be read and taken seriously!