Since the retention of under-represented groups within computing remains an ongoing challenge for colleges and universities, the ACM Education Committee established a Retention Committee in November 2016 to "address the current issue of retention in 4-year, post-secondary CS education programs, specifically of the retention of women and URM students following CS1 and CS2 (where the pipeline is most leaky)." [1] This article focuses on curricular factors that may impact the retention of students, and is based on several conversations and observations within the Committee. Starting with a high-level overview, the discussion identifies common curricular issues, especially in courses at the introductory level, and then outlines some established approaches that can address these challenges.

Curricular Factors: An Overview

Although retention of students at each stage of an undergraduate computing program merits review, specific attention is paid at the introductory levels, where the pipeline sometimes seems particularly leaky. For example, in [4], cited for "Best Presentation" at ITiCSE 2016, Andrew Luxton-Reilly reported on "a study of 161 introductory programming courses across 15 countries," indicating a passing rate of only 67.7% and stating, "An analysis of historical data revealed no significant differences in pass rates since the 1980's, suggesting that a pass rate of around 67% is typical and has remained so for more than 30 years." [4] Of course, some schools have high passing rates, but others clearly continue to struggle.

Overall, the goal should be to change introductory courses, as needed, to accommodate backgrounds of all incoming students, rather than to try to change the students themselves.

Such circumstances led the ACM Retention Committee to focus primarily at the introductory level. In restricting attention to curricular matters, this article focuses upon creating an inviting and supportive environment for all beginning computing students. To clarify jargon, outreach sometimes is described as targeting women and under-represented minorities (URMs), and the acronym URMs is associated with "students of color." In this article, discussion related to URMs certainly includes "students of color," but also is intended to be broader. The goal is to include students of all gender identities, ethnic backgrounds, cultural perspectives, socio-economic statuses, and intellectual/career interests. Rather than trying to encapsulate one perspective or a set of initial assumptions regarding incoming students, introductory courses need to connect with students of all backgrounds, perspectives, and interests as they first walk through the door. Overall, the goal should be to change introductory courses, as needed, to accommodate backgrounds of all incoming students, rather than to try to change the students themselves.

In what follows, this article reviews common circumstances and curricular factors and then identifies several approaches known to address some of these difficulties.

Curricular Issues at the Introductory Level

The audience for CS1 and CS2 is quite varied, and introductory computing courses must meet incoming students where they are. Although challenges have arisen at the introductory level for years, the need to addresses non-homogeneous populations may have become more pronounced as programs reach out to an increasingly broad range of students. These difficulties are further exacerbated with record-setting demand both by those interested in computing majors and those in client disciplines who realize the importance of learning about computing. Addressing current needs requires many considerations.

  • Encouraging students to explore a computing course
    Incoming students may have many misconceptions regarding the nature of computing. Also, some schools report that some students, particularly women and under-represented minorities, may delay taking introductory computing until their third or fourth undergraduate year—too late to become majors once they decide they truly like the discipline.
  • Shaping and clarifying student expectations
    Incoming students may have encountered fictionalized media representations of computing that provide a distorted or inaccurate perspective on the discipline and the people in it. For example, in one course, a student cited the movie, "2001: A Space Odyssey" to document current capabilities of artificial intelligence. Further, true novices may feel overwhelmed by this new subject and intimidated by students who seem to know much more.
  • Adjusting assumptions for introductory courses to be consistent with student backgrounds
    Rather than trying to encapsulate one perspective or a set of initial assumptions regarding incoming students, introductory courses need to connect with students of all perspectives and interests as they arrive on campus; the goal should be to change introductory courses, as needed, to accommodate all incoming students, rather than to try to change the students themselves.
  • Organizing introductory curricula to address diverse student perspectives and interests
    Students in beginning courses likely have varying interests, backgrounds, motivations, and technical experiences. Students with little background may require considerable initial mentoring and may be intimidated by students with substantial experience. Curricula with different starting points for students with different backgrounds may have a substantial impact on long-term success.
  • Balancing the perceived difficulty of exercises
    Students with modest experience may find printing a table of quart/liter equivalents easy (if less than motivating) while novices may need substantial thought and learning to accomplish the task. Accommodating both true novices and more experienced students can be challenging, and some schools split CS1 into multiple sections based upon students' previous computing experience.
  • Addressing vast differences in student interests and backgrounds
    To accommodate varying interests, schools may offer different versions of introductory courses—all with common core content, but different applications and contexts. Placing students in the relevant sections can be challenging, as student backgrounds may be uncertain, and students may react poorly to placement in courses that emphasize elementary problem solving and other basics.
  • Providing support and guidance in beginning courses
    True novices may struggle as they begin computing courses. However, after a semester or two including substantial orientation and experimentation, these students can flourish and become among the best majors! In contrast, students with moderate computing background may become restless with a review of activities that they have been doing for years.
  • Articulating course progressions or pathways to middle and upper-level courses for students with varied backgrounds and interests, based upon what/when initial computing courses were taken
    Beginning courses ultimately lead into upper level courses, so curricular programs need to be both flexible and rigorous, allowing students to start where they are and steadily progress to where they need to be.
  • Managing high student demand with limited staffing and constrained resources
    With high student demand, staffing and/or facility constraints stretch resources—sometimes beyond a sustainable level. At some schools, these constraints create pressure to limit the number of students allowed to declare a major. Because schools have different timelines and requirements for majors, the impact of these limits can be immediately apparent or only become obvious at later stages [2].

Addressing Curricular and Pedagogical Issues: Some Approaches

Although introductory courses face many challenges in meeting the needs of a diverse student population, experiments and projects from many computing faculty have yielded a wide range of helpful best-practice documents, pedagogical approaches, curricular models, and programs. The following lists identify several approaches that have surfaced in initial conversations within the ACM Retention Committee.

[T]his article reviews common circumstances and curricular factors and then identifies several approaches known to address some of these difficulties.

Pedagogical approaches: Generally, pedagogy that promotes active student engagement, often with collaboration, has been found helpful in keeping drop-out rates reasonably low and overall student performance high. Several successful and widely-used approaches, especially in CS1 and CS2, follow.

  • Exercises with pair programming promote learning and combat stereotypes that programming is a solitary enterprise. Since pair programming may be new to students, on-going discussion may be needed to explain roles and suggest constructive behaviors. When used, the person at the keyboard should change frequently to promote communication and shared responsibility, and partners should be changed often (perhaps weekly) to ensure one person does not become dependent upon another.
  • Peer Instruction (PI) in Computer Science, often including the use of clickers, has been shown to promote discussion and student engagement in large classes, as described in [6].
  • Process-oriented Guided-Inquiry Learning (POGIL) emphasizes "a learning cycle of exploration, concept invention and application - as the basis for many of the carefully designed materials that students use to guide them to construct new knowledge." [7] The POGIL organization provides extensive materials and training workshops with funding from the National Science Foundation, the Department of Education, the Hach Scientific Foundation, and the Toyota USA Foundation.
  • Team-based Learning (TBL) "is an evidence based collaborative learning teaching strategy designed around units of instruction," with lessons following a carefully-prescribed format. The Team-Based Learning Cooperative provides substantial training and resources, as described in [8].
  • Other lab-based approaches, such as the daily-lab formats for CS1 and CS2 described in [9], similarly utilize collaboration and active student engagement, within a relatively fluid format.

Curricular Models: Some curricular structures aim to address challenges encountered in drawing students from many disciplines into introductory computing.

  • Technically-oriented schools, such as Harvey Mudd College, may require all students to take computing. When computing is taken early, students gain understandings and tools they can use in many STEM disciplines. Students also experience the discipline of computing in their first (or second) year, giving them time to consider majoring if they are excited about what they find.
  • Other schools allow computing to count as one component of general education requirements. Such options can bring students into contact with computing and help them learn what computing is and is not. Unless general education requirements must be fulfilled early, however, timing may or may not allow interested students to continue onward as majors.
  • Some schools connect CS1 with image processing or robotics (Bryn Mawr, Georgia Tech, and Grinnell) or with music (UMass at Lowell). In some cases, faculty from art or music attend some CS1 class sessions to provide insights about application themes. These types of experiments may resonate with students in the arts or other populations and reach out to students outside STEM fields.

Best-practice documents: Several approaches provide suggestions and outline best practices for reaching out and supporting women and URMs in computing specifically and STEM fields more generally.

  • The National Center for Women and Information (NCWIT) maintains materials on "Promising Practices" that can help computing programs recruit and retain diverse student populations, see [5].
  • EngageCSEdu offers a repository of materials for CS1 and CS2, "including assignments, tutorials, labs, assessments, lecture notes, exercises and projects" with a special interest on recruiting women and URMs. With resources available at [3], EngageCSEdu is a collaborative effort of NCWIT and Google.
  • Several colleges and universities support faculty discussion groups to encourage effective teaching and learning. For example, the "Talking Teaching" Lunch Group at UC San Diego provides an ongoing forum for the discussion of best practices. Similarly, for several decades, Grinnell's Science Teaching and Learning Group has brought STEM faculty together for periodic discussions each semester. Several schools also run sessions for new faculty, including UC San Diego's New Faculty Workshop, Williams' Course Design Group, and many, many others.

Overall, the ACM Retention Committee has reviewed many curricular changes related to the retention of women and URMs in introductory computing, and it continues to collect and analyze data to provide further insights into programmatic issues. However, the Committee also provides encouragement and constructive hope, as it continues to identify resources and suggest possible directions.

• Acknowledgments

Many thanks to the members of the ACM Retention Committee for their support and help in the creation and editing of this article. Special thanks to the following members of the Publications Subcommittee: Colleen Lewis, Debra Richardson, and Alison Derbenwick Miller.


1. Association for Computing Machinery. Author's Letter of Appointment to the ACM Retention Committee, 2016 October 4.

2. CRA Enrollment Committee Institution Subgroup. Generation CS, CS Undergraduate Enrollments Surge Since 2006, (2017); Accessed 2017 September 21.

3. EngageCSEdu. Home Page: Foster diversity in your introductory computer science courses with quality content and engaging pedagogy; Accessed 2017 August 9.

4. Luxton-Reilly, A. Learning to Program is Easy, Proceedings of the 21st Annual Conference on Innovation and Technology in Computer Science Education, (ITiCSE) (2016), 284–289.

5. National Center for Women and Information Technology. Promising Practices; Accessed 2017 August 9.

6. Peer Instruction for Computer Science. Home Page: Peer Instruction for Computer Science; Accessed 2017 August 9.

7. Process Oriented Guided Inquiry Learning; Accessed 2017 August 9.

8. Team-Based Learning Collaborative. What is TBL? Overview; Accessed 2017 August 9.

9. Walker, H. M. Course Home Page: CSC 161: Imperative Problem Solving and Data Structures; Accessed 2017 August 9.


Henry M. Walker
Dept of Computer Science
Grinnell College
Grinnell, Iowa 50112 USA
[email protected]

Copyright held by author.

The Digital Library is published by the Association for Computing Machinery. Copyright © 2017 ACM, Inc.

Contents available in PDF
View Full Citation and Bibliometrics in the ACM DL.


There are no comments at this time.


To comment you must create or log in with your ACM account.