2018 ACM Education Council meeting
About a year ago I joined the ACM Education Council. The council is part the Association for Computing Machinery, the professional society that represents computer science across both academia and industry. The Education Council’s job is is to “promote ACM’s educational mission to… universities, community colleges, high schools, corporations, and the US government.” It’s not exactly clear what ACM’s educational mission is, but through the council’s medley of ad hoc groups, tasks, and other priorities, many interesting missions emerge.
I didn’t make last year’s in person meeting, but this year it was in Portland, so it was easy for me to take a quick flight south from Seattle to network with the ~30 members of the council. And of course eat some tasty local donuts:
Attendees represented many pockets of ACM and CS education, including the ACM CEO Vicki Hanson, ACM president Cheri Pancake, the director of the CS Teacher Association (CSTA), representatives of special interest groups that have education activities (SIGGRAPH, SIGCHI, SIGPLAN, and SIGCSE), the and representatives of industry. There were also council members (like myself) from the US, Europe, Canada, New Zealand, China, India, and Brazil, representing education-interested faculty across academia.
Updates on ACM
Vicki Hanson, the new ACM CEO, kicked off the meeting with an overview of ACM. The financial health of ACM is sound. There are about 100,000 members, half from the U.S., and the rest spread across Europe, India, China, and Canada. About one third of these are student members. There’s a new board trying to learn more about who the practitioners in this member community are. So far, the data shows that most have a curiosity about research and work for larger organizations. Most aren’t interested in certification. This has led to higher priorities around ACM learning webinars and more local meetups that better fit into practitioners’ work schedules. It’s also led to efforts like the ACM Future of Computing Academia, to draw in younger faculty to identify future priorities for ACM.
Vicki talked about some of the main activities of ACM councils around the world. For example, China has kicked off an annual conference, an editorial board for translating CACM articles into Chinese. ACM Europe has been doing policy work on inclusion, developing support for local chapters, and also doing an insert for CACM. ACM India has also been doing an annual conference and several smaller community building conferences. One of the biggest initiatives is trying to reach 300 million students across 1.6 million schools to teach computational thinking. One of the big challenges in India was developing teaching materials in the various languages and dialects in India.
ACM has also been trying to address ethics issues such conflict of interest rules and plagiarism and gaps in ACM’s topical coverage of CS around ethics. For example, it’s been prioritizing events like AI for Good, which has included people from all kinds of disciplines, not just computing. This conference is in its second year, and it’s developing annual reports. It’s also been developing policies against harassment in ACM activities, which includes training and enforcement guidelines, especially for conference leaders. And then, of course, there was a major revision to the ACM Code of Ethics. The interesting thing about the code is that it’s not framed as rules, but as issues on which professionals should reflect.
Vicki also talked about future directions for ACM. She’s talked to a lot of different stakeholders around the world, especially around new topics in computing. The big discovery of this listening tour was how to better structure how practitioners and researchers engage with ACM. For example, why aren’t new conferences seeking out ACM? It turns out (not surprisingly), that few people know how ACM works or what benefits it provides. The feedback also made it clear that societal issues, and interdisciplinary, are central to most people’s concerns, and that ACM needs to do more to guide and support the community. ACM’s new “Ask an Ethicist” service is a start, but most people in the room admitted that there was far more to do.
Mehran Sahami, outgoing co-chair of the ACM Education Board, summarized some of the ongoing initiatives by the council. The first big chunk of activities were computing curricula for CS, data science, and information systems, cybersecurity, and information technology. The big vision behind these curricular recommendations is to help shape the curriculum in the world. It turns out these these gargantuan documents are used broadly, especially outside the U.S., primarily to inform revisions to curriculum, to support changes to university leadership, and to inform the design of new programs. For example, the new data science curriculum is aiming to find some consensus around the role of computing in data science. There’s also a partnership with the Association for Information Systems to develop a curriculum for information systems programs. These are nuanced, massive, complex efforts to wrangle consensus out of diverse perspectives.
There’s another effort to run a survey for non-doctoral granting departments in computing to complement the Taulbee survey, which the Computing Research Association runs. This is a hugely important data source for discussing national trends in CS. This will be the first year the survey runs, so I’m excited to see the data and report.
Mehran also discussed the Learning@Scale conference, which isn’t under an ACM SIG, but is under the education board. It’s revenue positive, has 130–200 attendees, and it’s pretty vibrant. It’s also bridged to other communities, including ICLS and AI in Education. The council received some feedback that the co-location with other conferences was really valuable. There was some discussion about how ACM could foster and resource more of this cross-community activity.
Curricula Frameworks
Next, we dived deep into progress on new and revised curriculum frameworks. When I joined the council, I was really unclear on the purpose of these documents. I’ve learned over time that while many institutions of higher education have faculty with strong opinions about curriculum, there are many other institutions without the expertise to have an opinion about what to teach, and who to hire to teach it. My sense is that these curricular guidelines really serve the long tail of CS departments, giving them guidance and a proxy for expertise they might be lacking on their faculty.
The first curricular discussion was of the 2020 computing curricular revision. Alison Clear and Allen Parrish talked about the complexity of trying to synthesize the growing body of knowledge in CS, especially AI and Cybersecurity, which are affecting many aspects of CS. The bulk of the work has been in trying to define competencies that shape what is and isn’t taught. The team is now shifting into a phase of feedback from over a dozen academic communities. I asked about the task force’s plans for soliciting feedback from industry; they planned on relying on professional societies across the world and the social networks of the industry representatives on the task force itself.
Paul Leidig discussed an exploratory task force to investigate recommendations for information systems (IS) curricula that often appear in business schools and information schools. They’re new recommendations were to more strongly address the technical skills needed for IS graduates. The other issue was to focus on creating a more living document that is more responsive to change, and responsive to changes in the scope and focus of the broad diversity of degree programs related to information. Paul admitted that the motley collection of business schools and information schools is hard to organize. It wasn’t clear to me that these units are actually looking for guidance from ACM or how the document might be used.
Andrea Danyluk discussed the task force working on data science curriculum. She co-chairs a group of eight faculty and one industry member trying to offer recommendations on the content of data science degrees. The reason for creating this was to create a “stake in the ground” to prevent statistics from “claiming” all of data science. The goal of the curriculum framework is to articulate the importance of computing in data science and identify computing-related competencies for undergrad data science curriculum. It has a curious political motivate; I’m unsure whether the numerous programs that have been developed over the past decade would respond to the document or utilize it in their program planning.
Education Policy
Next, Jeff Forbes discussed the Education Policy Committee, which promotes public policies that ensure high quality computer science education at all levels, especially K-12. The mission involves reviewing issues, doing research on issues, and making recommendations.
For example, the committee has been working on pathways to help get K-12 CS into schools, giving resources to leaders who want to work on state and national policy. I’ve used these resources myself to help foster further efforts in Washington state.
Another priority is the Community College Transfer Pathways report, which is trying to address a particular concern of helping students navigate the myriad pathways to higher education, many of which result in failure (e.g., <15% completion rates). They’ve worked on case studies from five U.S. students and generated recommendations. For example, one insight is that transfer students really need relationships with faculty at the institutions they are transferring to. Without it, many don’t have confidence to transfer and succeed. Another example is that students get significant benefits from cohort models, involving groups of students that transfer together. These are all related to research on transfer student experiences, including some by my undergrad student Harrison Kwik, a transfer student himself. One surprising issue that was raised is that community college enrollment is actually down because the economy is good: most are taking jobs instead of pursuing further higher education.
Global education efforts
While ACM and it’s education council is heavily biased toward North America, it has an increasing number of activities in Europe and Asia.
The first was Informatics for All, a European initiative to broadly implement CS in K-12 education in Europe. (Note that in most of Europe, CS is referred to as Informatics). This includes ACM Europe, Informatics Europe, Council of European Professional Information Societies, and Bobby Schnabel, who represents the CS for All perspectives, and presented the initiative. It’s still a very nascent initiative, but it has very strong support from the EU. There’s a big workshop in 2019 that will hopefully involve industry. But it’s a big, big lift, even bigger than the U.S., because they’re entirely distinct countries, unlike the fifty states in the United States.
The next update was from ACM China, presented by Ming Zhang from Peking University. She talked about many challenges in China. The digital economy in China is growing steady, but not spiking dramatically as it has in the U.S. There are many new efforts to build more interdisciplinary work. However, there are significant funding issues for computing education research. The Chinese government, however, has started to prioritize engineering education, and the new China National Higher Education catalog now has more CS subjects, such as AI and big data. This should all lead to more growth in CS education in China, where curriculum is centralized. The ACM China SIGCSE chapter now has 469 members, and they’ve run a SIGCSE technical symposium, workshops on developing research capacity, and panel discussions at several other conferences. They also sent several members to the SIGCSE symposium in North America.
The goals of ACM India is to improve both CS education and CS education research. The strategy in India is to work with the government and provide services to students and faculty, and trying (and struggling) to influence CS curriculum. One big focus has been CSPathshala, trying to get CS into all Indian schools by 2030. The scale of this is astounding, trying to reach 1.6 million schools, 350 million students, and 44 education boards. They’ve succeeded in reaching 160 schools so far, building teacher workshops and partnerships to try to scale these efforts. My impression was that India is facing many of the same challenges as other countries, trying to change teacher practices, fighting for space in curricula. However, the lack of computers in India is one notable difference, which has led them to use many CS Unplugged activities. There are broad, open questions about how to motivate teachers, how to train them, how to assess students, and how to measure impact of the program.
Another issue in India has been the low quality of programming skills, mirroring evidence about the poor learning outcomes in the U.S., UK, and Australasia. They’ve been developing workshops and online content to help faculty better teach CS. They’re trying to give one way to teach it, but they face strong opinions amongst faculty, just as in other countries. This has also led to some interest in bootstrapping more research on CS education. It’s quite nascent, but workshops are drawing some interest.
Venkatesh also talked about the challenges of trying to improve the quality of CS curriculum in India. They’ve been using the ACM Curriculum Guidelines as a starting point, but are facing resistance, because many institutions prefer to mimic the curriculum of the respected IITs, rather than differentiate curricula. This suggests that the curriculum guidelines can be one of many resources to inform curriculum design, but are no means a turnkey solution.
Teaching in K-12, 2-year, and non-PhD granting institutions
ACM traditionally has focused on research, and so institutions that do not grant PhDs—four year colleges, two-year colleges, and K-12 institutions are relatively new to ACM. Jake Baskin, the new executive director of the CS Teachers Association (CSTA), presented on the latest activities of the CSTA. He talked about just how much support CS teachers need, and how little they get in these early days of K-12 CS education. They don’t have guidance, community, resources, or evidence. CSTA is the central organization that provides all of these things to CS teachers. Jake talked about the new era of CSTA, which is going to focus on bringing value to members through community building. This will involve new chapters, content to support chapter growth, staff and leadership development for chapters, and funding to help chapters implement programs.
Another activity related to non-research institutions was ACM survey of non-research institutions. This survey helps fill a big gap in U.S. data on what’s happening in 2 and 4-year institutions teaching CS, but not granting Ph.D.s. Response rate is still pretty low to this survey unfortunately. Results will appear in the 2018 issue of Inroads magazine. The data show that enrollments are up, tenure-track hires are up, teaching track hiring is up. But most of the growth was in adjunct faculty. There’s six years of data on this now, showing that number of degrees produced has more than doubled over the past six years. Yet the average faculty sizes in NDC institutions is still only 13, so the programs are small. Another trend shows that participation by women in CS is every so slightly increasing, but only by a few percent. There’s effectively no change in race and ethnicity in the aggregate, but the data hasn’t been disaggregated yet. None of the surveys gather data about disability; Stu argued that asking for it would reduce response rates because most departments don’t have data on it. I’m skeptical; until we started asking about data on gender, did they have that data?
Cara Tang from the Committee for Computing Education in Community (CCECC) gave an overview of the committee’s work over the last year. One was a bunch of research and presentations about supporting transfer students. Another effort was disseminating exemplars of cybersecurity certificates, helping to propagate stronger cybersecurity programs across the U.S. They’ve also been working on guidelines for supporting transfer students in IT programs and guidelines for data science. Many of these efforts focus on two-year curricular efforts, mirroring some of the four-year curriculum guidelines. Jane Prey made a fascinating political comment about how support for community colleges, both local and federal, is bolstered by the fact that every representative had a community college in their district. This suggests that they could and should be a major site for investing in high quality CS education nationally.
Diversity, equity, and inclusion
To wrap up the reports for the day, we had presentations from NCWIT and the retention task force. Beth Quinn of the National Center for Women in Information Technology (NCWIT) shared many exciting updates. Right now, they’re thinking a lot about communities of practice and strategic partnerships. Their goal really is to figure out how to put themselves out of business by changing the culture of CS departments to be more inclusive to women. One idea is to find ways to partner with other organizations to ensure some of NCWIT’s resources can exist even if NCWIT does not. One of the big areas of focus is EngageCSEdu, a collection of peer-reviewed materials for intro college and high school CS courses. Because the goal is to broaden participation in computing through inclusive pedagogy and engaging content, all materials accepted into the collection must use at least one of eleven research-based engagement practices. It’s an excellent example of applying research to practice.
The council also had some interesting discussions about retention. The committee on retention, led by Chris Stephenson, has written a massive 60 page article on evidence about retention in CS. The goal of the report is to really inform a lot departments about the evidence base and suggest best practices. The article, of course, has no silver bullets, but it is a great survey of important questions that departments should ask and recommendations for how to proceed.
K-12 in Oregon
Jim Hook and Joanna Goode visited from Portland State the University of Oregon, respectively, to give a talk about their efforts to build K-12 computer science in Oregon public schools. Joanna discussed the equity issues first: only 436 students took the AP CS exam in 2017, and they were 65% white, 25% asian, and only 17% were girls. Only two test takers were black boys. And so their efforts primary goals were equity, in addition to scaling access in general. They were funded by a 3-year NSF CS for All grant last year and have spent the last year trying to find a way forward in their research-practice partnership. They explicitly rejected a jobs framing, instead focusing on framing it as for everyone, regardless of job outcomes. They also focused on the Exploring Computer Science program, and bringing it to both in-service professional development and pre-service teacher preparation. The model they’ve built is a 2-year professional development, with a summer institute, four quarterly workshops, and a teacher professional community.
One of the interesting discussions was on state policy. In Oregon, there is no state policy on anything around CS, so the team is starting from zero. That means they’re really building community from nothing. There is no memo, there is no leadership motivating teachers to learn to teach CS. The upside, however, is that they have time to try to envision good policy that best supports equity in the state. Another interesting barrier was the $45,000 cost of becoming a teacher, and the complexities around creating policies that don’t increase that cost. A third is the massive number of unfunded mandates in teacher education, with CS being yet another.
Another interesting dimension was the mundane problem of supporting travel, both by the team and the large number of teachers engaged in professional development. The team’s strategy was primarily targeted investments in collocated meetings, combined with lots of synchronous online meetings on Zoom. It sounded functional, but not necessarily ideal, but ultimately optimized for being respectful of teachers’ time and money.
I was quite impressed by Joanna and Jim’s team. They’re really tackling every detail of building K-12 CS in Oregon, and doing so with a very pure equity mindset, despite many immense barriers to progress. They’re inspiring me to continue my efforts in Washington state with my own colleagues and community.
The education council’s next priorities
We spent the end of Wednesday brainstorming priorities for the council and Thursday morning workshopping some of the most popular ones, including:
- How to build a more sustainable data pipeline to monitor changes in CS engagement and retention across all of higher education, not just research institutions.
- More rigorous policy and guidance on incorporating ethics and social responsibility in curricula and amongst faculty.
- Guidance on how higher education should be responding to the soon to be rapidly shifting diversity of incoming college students, both racially, ethnically, and with respect to prior knowledge.
- How the council should support teaching capacity in higher education, while maintaining equity.
- How to better disseminate the good work of the council.
- Developing resources to improve teaching practices in higher education.
We broke out to discuss many of these topics and others. I unfortunately had to leave for the airport to make the last day of my summer high school class (and then my flight was delayed by 2-hours, making me want to race back to the meeting to join in the fun), but I’m sure many critical discussions ensued, shaping the council’s work for the coming year.
Reflections on the council
As a newcomer to the council, and a newcomer to community-level academic efforts in general (outside of a research, of course), I left with a few observations and thoughts.
First, organizing large groups of independent, motivated people with strong opinions is incredibly difficult, and the ACM Education Council isn’t immune to this. I think it mirrors the many challenges of local community organizing, and shares many of the same weaknesses: the lack of a single clear purpose and organizational hierarchy limits how much a community can do. I think the council, and ACM in general, would benefit from creating some best practices for organizing groups and running meetings, such as how to facilitate large group conversations, how to support remote participation, and how to facilitate brainstorming and leadership development. I know as newcomer I’m supposed to sit back and learn, but I really think many of the norms around decision making are broken and ignoring many useful best practices grounded in evidence from social psychology and decision sciences.
Second, I think the Education Council’s mission is too vague. If the council’s job is to promote ACM’s education mission, what is ACM’s education mission? Is it to ensure an educated professional class of software engineers? Is it to ensure a trained and resourced community of CS educators? Is it to disseminate research findings? Or is it’s purpose to choose from opportunities like these? As a new member of the council, it wasn’t clear at all what the purpose of the council is, what the purpose of the annual meeting is, or what strategic objectives ACM wants to achieve with the council.
All that said, the council does some fantastic, important work despite the lack of clarity in it’s purpose. It’s getting good data, doing thankless work on identifying competencies and learning objectives, and wrangling a broad community of experts to identify gaps in CS education construed broadly. With a clearer purpose, some improved organizing, and revised norms, I think the council could play a crucial role in improving the knowledge, abilities, and beliefs of both CS teachers and software engineers across the globe.
