OBE AT THE CROSSROADS
Why Computing Education Needs a Lean, Meta-Level Outcome Model, Not a Paperwork Empire and Accreditation Wrath
A case for stopping OBE where it adds value: at PEOs, PLOs, and CLOs. Let transcripts, industry feedback, and technology trends do the rest.
Adnan Abid (PhD)
Professor
Department of Data Science
Faculty of Computing & Information Technology
University of the Punjab, Lahore, Pakistan
Program Evaluator, National Computing Education Accreditation Council (NCEAC)
This assessment is written from inside the system it examines: from years of classroom practice in computing, from serving as a program evaluator for Pakistan’s own computing accreditation council, and from repeated, candid conversations with senior colleagues across departments who describe the same gap between what OBE paperwork claims and what actually happens in their classrooms.
The Promise We Were Sold
Outcome-Based Education arrived with a simple, attractive premise: define what a graduate should be able to do, teach and assess toward that, and continuously improve based on evidence. For engineering disciplines built around licensure, safety, and public accountability, that logic made sense. A bridge either holds or it doesn’t, and the graduate who designed it must be demonstrably competent.
Nearly three decades after ABET’s Engineering Criteria 2000 pushed the world toward outcomes-based accreditation, and roughly a decade and a half after Pakistan’s own computing and engineering programs adopted the model, it is time to ask a harder question honestly: did OBE actually deliver the result-oriented education it promised, or did it deliver the appearance of one?
The evidence, both international and local, suggests the latter, especially for engineering and engineering-adjacent disciplines where OBE was implemented most rigidly.
OBE Was Never Truly Result-Oriented for Engineering: Here Is the Evidence
The irony is sharp. A framework named “outcome-based” has been repeatedly criticized by its own scholarly community for optimizing form over substance.
• The assessment machinery displaced the learning it was meant to measure. Writing roughly a decade after ABET’s EC2000 reforms took hold, engineering-education researcher D. C. Woolston argued that outcomes-based assessment shifts attention to the results of instruction rather than to how learning itself happens, training students to chase the expected answer instead of developing the underlying judgment.
• The core terminology was never settled. Even years into ABET’s outcomes system, researchers documenting OBE implementation for the Washington Accord found that standard, universally agreed operational definitions for terms such as objectives, outcomes, and performance indicators still did not exist across the field, meaning programs worldwide were reporting “attainment” against a vocabulary they had not agreed on.
• Poor curriculum design quietly produced false outcomes. The same body of literature notes plainly that flawed curriculum and outcome mapping produces misleading attainment data. In other words, a program can look compliant and “outcome-attained” while the underlying learning gap remains invisible.
• Industry never fully endorsed the premise. A study from the Faculty of Engineering at Universiti Malaysia Sarawak opens with the same uncomfortable finding echoed across the literature: graduates were, and often still are, not acquiring the competencies employers actually need, which is the very gap OBE was supposed to close.
• Recent systematic reviews catalogue the same failure modes across countries. A 2024 systematic literature review covering ten empirical studies published between 2013 and 2024 found the two most common challenges, reported in 80 percent of studies each, were continuous assessment and feedback, and efficient time and resource management, alongside persistent problems in developing teaching resources and shifting educator mindsets.
Put together, this is not a case of a good model poorly executed in one country. It is a pattern, visible across ABET-anchored engineering education globally, of an outcomes framework that became very good at documenting outcomes and comparatively poor at producing them.
Ask the question a working engineer would actually ask: after nearly three decades of ABET-style outcomes accreditation, is there reliable evidence of measurably better instructors or measurably better engineers as a result? Neither the international literature reviewed above nor the everyday experience of hiring managers and senior faculty across Pakistan’s own engineering and computing departments points to such a shift. The framework produced better files. It has not been shown to produce better engineers.
Where OBE Failed Outright: The Cautionary Case
If Pakistan’s computing sector wants a preview of what happens when an outcomes model is pursued as policy dogma rather than a practical tool, South Africa’s experience is the starkest warning available in the global literature.
South Africa’s post-1994 school reform, Curriculum 2005, was built entirely on outcomes-based education. Within a few years it had to be reviewed and substantially rewritten, and OBE became, in the words of the academic literature examining it, shorthand for the country’s broader “education crisis.” Independent analyses found the model overambitious for classroom realities, especially where teachers lacked the training to operationalize a framework long on outcome language and short on content guidance.
Tellingly, one of the most cited academic papers on the subject was originally presented in 1997, before the policy even took effect, under the title “Why OBE Will Fail.” Its formal published version concluded that OBE, as designed, could not survive the philosophical, political, and practical contradictions built into its own implementation. That paper’s warnings were largely borne out.
The lesson is not that outcomes thinking is worthless. It is that an outcomes framework detached from institutional capacity, over-engineered at the operational level, and treated as an end in itself will collapse under its own documentation weight. This is exactly the trajectory computing education in Pakistan is now at risk of repeating at the university level.
Accreditation Accords: Signal, Not Substance
A parallel problem sits one level up, in how international accreditation accords are used to justify OBE’s most burdensome features.
This is not an abstract worry. In 2018, sitting in a National Computing Education Accreditation Council meeting as the council prepared to extend a Washington-Accord-style outcomes framework to computing programs, I raised exactly this concern: what specific problem would full OBE implementation solve for computing education in Pakistan? The justification offered at the time was that it would help the international recognition of Pakistani computing degrees and move the country toward Seoul Accord membership. Neither justification has aged well. Recognition of Pakistani computing degrees abroad was never, in practice, contingent on a domestic OBE paper trail; graduates were already being admitted to postgraduate programs and hired internationally on the strength of their transcripts, projects, and test scores. And Seoul Accord membership itself, examined honestly, changes very little: it is a mutual-recognition arrangement between accrediting bodies, not a credential that employers, graduate schools, or students themselves treat as decisive.
Accords such as the Seoul Accord were designed for an earlier era, when degree recognition was the primary gatekeeper of professional mobility and licensure. That era has largely passed for computing. Today’s software and IT industry hires on demonstrable ability: portfolios, open-source contributions, technical assessments, and project experience, not on whether a graduate’s transcript traces back to a signatory institution. Hiring in this sector is increasingly degree-agnostic, and accord membership does not measurably change outcomes for strong or weak students alike.
This matters because accords audit governance, documentation, and process alignment, not teaching quality, faculty depth, or graduate capability. Accord membership is therefore a useful hygiene factor for international recognition, but it should never have been allowed to dictate how granular, manual, and paperwork-intensive domestic OBE implementation becomes. When it does, institutions end up optimizing for auditability rather than learning, precisely the failure mode the South African case illustrates and the engineering-education literature documents.
Why Computing Cannot Be Governed Like a Licensed Profession
Engineering retains a plausible case for tight outcome governance: public safety, professional licensure, and multi-decade infrastructure consequences justify rigorous, traceable accountability. Computing has none of those constraints in the same form.
Three structural realities separate computing from regulated engineering disciplines:
• Computing is not a licensure-gated profession. No regulatory body certifies individual software engineers the way licensing boards certify structural or electrical engineers. Competence is validated continuously by the market, through code reviews, shipped products, and technical interviews, not by a one-time professional credential tied to accredited coursework.
• The knowledge base moves faster than any accreditation cycle. Cloud platforms, AI and machine learning tooling, data engineering practices, and security paradigms shift meaningfully within two to three years, often faster than a single accreditation review cycle. A CLO-PLO mapping frozen for compliance purposes is close to obsolete before it is ever audited.
• Employability signals already exist outside the classroom. GitHub activity, competitive programming records, internships, and certifications now carry more weight with employers than an institution’s accreditation status. Layering heavy manual PLO tracking on top of a domain that is already self-verifying through the market adds administrative cost without adding trustworthy signal.
These are not just theoretical claims. TestGorilla’s 2025 global hiring survey of over a thousand hiring decision-makers found that 85 percent of employers now practice some form of skills-based hiring, and that 53 percent have dropped degree requirements for at least some roles, up from just 30 percent the year before, a shift led disproportionately by technology employers hiring for exactly the roles computing graduates enter. An accreditation architecture built for a licensure-gated profession is being layered onto a labor market that is visibly moving in the opposite direction: away from credential-gating and toward direct, continuous skills verification.
Applying an engineering-grade, document-heavy OBE apparatus to this environment is a category error. It burdens computing faculty, who should be spending their time on curriculum currency and industry engagement, with a compliance architecture built for a fundamentally different risk profile.
The Computing Classroom, Up Close: Three Symptoms of a System Not Understood
Step down from accords and cross-country literature reviews to the level of an actual computing department in Pakistan, and the case against heavy OBE only gets stronger. Three patterns recur across institutions and are corroborated, independently, by colleagues teaching in different universities.
• A parallel ledger that adds little the transcript did not already show. Computing programs in Pakistan already maintain a full transcript, grades, credit hours, GPA, course-by-course performance, that is itself an outcome record, produced by the assessments a student actually sat for. OBE asks departments to additionally maintain a second, parallel accounting of the same performance, remapped onto PLOs and CLOs, often at a coarser and more subjective level of precision than the marks already on file. Two systems now claim to measure the same learning, and the newer, harder-to-maintain one is frequently the less precise of the two.
• An instrument most of its operators were never trained to play. In direct conversation with colleagues across several computing departments, and consistent with what Mahrishi and colleagues’ 2025 systematic review of the OBE literature found, keyword analysis of the field shows OBE research is rich in compliance vocabulary, outcomes, attainment, mapping, but strikingly thin on terms like active learning, student engagement, or industry collaboration. That absence mirrors the classroom: a large share of computing academicians can recite the definition of a CLO or a PLO, but far fewer can explain, unprompted, how a specific assessment item was designed to discriminate attainment of one outcome from another. Practicing OBE correctly, meaning writing assessments backward from outcome statements and using the resulting data to change teaching, remains rare. Most practice is compliance dressed as pedagogy.
• The midterm-to-CLO1 shortcut. The clearest evidence of the understanding gap is also the most common: in a large share of course files, every assessment up to the midterm is mapped to CLO1, and everything from the midterm onward is split between CLO2 and CLO3, almost regardless of what those specific questions actually test. This is not idiosyncratic; it is close to a default convention. Derouich’s 2025 analysis of CLO-PLO alignment describes exactly this failure mode in general terms: assessment mapping gravitates toward convenient, easy-to-grade shortcuts that chiefly capture lower-order recall and routine procedure, while higher-order outcomes such as analysis, synthesis, and authentic problem-solving go under-assessed or unassessed entirely. Once mapping becomes a scheduling convenience rather than a measurement decision, the entire CQI loop built on top of it, attainment percentages, action plans, and improvement narratives, ends up measuring the academic calendar, not the students.
None of this is a claim that individual instructors are careless. It is a claim about incentives: when the recordkeeping burden is high and the payoff for getting the mapping exactly right is invisible to anyone except an accreditation visit once every few years, convenience wins. A 2026 multi-level analysis by Shabbir and colleagues of CLO and PLO attainment across engineering courses found exactly this dynamic at scale: courses carrying a high teaching workload showed CLO attainment around 62 percent against 99 percent for courses with a light load, a 37-percentage-point gap driven not by weaker teaching but by administrative burden crowding out the very assessment-design work OBE depends on. This is the paradox worth sitting with: OBE’s contribution to curriculum design, forcing instructors to state upfront what a course is for and design lectures and assessments around it, is genuinely valuable and widely underappreciated. It is the recordkeeping layered on top of that design work that consumes the time, and it is exactly the layer most departments are failing to do honestly.
Effort Economics: Why the Deeper Practice Will Not Take Root
There is a simple labor-economics reason the pattern above persists, and it is worth naming plainly rather than moralizing about. Academic work in Pakistan, like knowledge work everywhere, runs on an implicit contract: people commit a certain number of units of time and energy to their teaching role, calibrated against pay, workload, research expectations, and the administrative duties already on their plate. Rigorous OBE, done the way its own literature prescribes, backward-designed assessments, item-level rubrics, honest CLO-PLO mapping, and a genuine CQI cycle, requires meaningfully more units of time and energy than conventional teaching, and it requires them indefinitely, every semester, for every course section. That additional demand was never resourced: no reduction in teaching load, no dedicated assessment-design support staff, and no meaningful weight in promotion criteria accompanied the mandate.
Get Adnanabid’s stories in your inbox
Join Medium for free to get updates from this writer.
Sociologists of organizations have a name for what happens next. Meyer and Rowan’s classic account of institutional ceremony describes how organizations facing a costly, hard-to-verify formal requirement often adopt its outward structure while decoupling day-to-day practice from it, satisfying the audit without absorbing the cost. That is a precise description of what several colleagues, independently and across different institutions, describe doing with OBE: course files, PLO matrices, and CQI reports are engineered after the fact to look complete and internally consistent for a visiting evaluation team, while actual teaching and grading proceed largely as they did before OBE arrived. Pakistani reviews of the country’s own OBE rollout reach a similar, more measured conclusion from the inside: Mirza’s 2022 review of OBE implementation in Pakistani institutions found that many departments have not devised a working Continuous Quality Improvement mechanism at all and continue running essentially traditional, syllabus-based courses underneath OBE paperwork; Haider, Siddiqui, and Ali’s 2023 evaluation of OBE in Pakistani engineering programs found close to 15 percent of faculty falling short of the KPI expectations the framework sets, and cautioned that it is still too early to say whether the reformed curricula are producing better graduates at all.
None of this is unique to Pakistan, and none of it is a character flaw peculiar to Pakistani academia. It is what happens whenever a compliance regime demands sustained extra effort without funding, staffing, or rewarding that effort, anywhere in the world. The honest response from a regulator is not to demand more rigorous enforcement of the current model; stricter audits of a system people cannot sustainably run will only produce better-engineered fiction. The honest response is to redesign the model so that what it asks for is something departments can actually deliver with the time and energy they realistically have, which is precisely the leaner division of labor proposed next.
The Way Forward: OBE Should Stop at the Meta Level
None of this is an argument to abandon outcomes thinking. It is an argument to be disciplined about where OBE adds real value and where it becomes theatre.
The evidence points to one clear boundary: OBE remains operationally relevant up to the point of defining the meta-level architecture: Program Educational Objectives (PEOs), Program Learning Outcomes (PLOs), and Course Learning Outcomes (CLOs). Beyond that point, its value degrades sharply the more manual and granular it becomes.
What should stay centralized and rigorous:
• A regulator-issued (HEC/NCEAC) standardized set of PEOs and PLOs for computing programs, so every institution starts from the same, currently relevant baseline.
• Predefined CLO-PLO mapping templates for core courses, removing duplicated design effort across hundreds of departments solving the identical problem independently.
• Periodic, centrally driven revision of this meta layer, not left to individual institutions to reinvent on their own schedule.
What should stop being manual and student-level:
• Discontinue parallel, student-wise PLO attainment registers maintained by hand alongside transcripts.
• Infer PLO and program-level attainment implicitly from transcript grades, course results, and assessment weightings already on record. The data already exists; it does not need to be re-collected in a second system.
• Treat Continuous Quality Improvement (CQI) as a data-driven feedback loop, not a reporting ritual, fed by four concrete streams:
Transcript and grade analytics. Aggregated, program-level trends in course and outcome performance.
Industry and alumni feedback. Structured input from employers and graduates on skill gaps, refreshed at least annually given how fast the sector moves.
Technological advancement signals. Tracked shifts in tooling, platforms, and in-demand skills (AI, cloud, data, cybersecurity) that should trigger curriculum review outside the normal accreditation cycle.
Environmental and contextual factors. Labor-market shifts, regional industry composition, and institutional resourcing that shape what “achievable outcomes” realistically look like for a given program.
This is not a weaker model. It is a leaner, faster, evidence-based one: precise where precision matters, in the definitions everyone must share, and loose where rigidity only produces paperwork, in how individual faculty and institutions operationalize teaching.
The Information-System Trap: Why a Heavy OBE Assessment Platform Is the Wrong Answer
A predictable response to years of OBE paperwork fatigue is to digitize it: build an information system that captures CLO attainment, PLO mapping, and question-level rubric scores for every student, every course, every semester, and presents it back as dashboards and compliance reports. Several such platforms are now being pitched to universities and regulators as the fix for OBE’s administrative burden.
This is solving the wrong problem, and doing so at real cost. The premise developed above is not that OBE’s record-keeping should be digitized, but that most of it should not be collected in the first place. A system that automates the capture of student-wise, question-wise, CLO-wise attainment data does not remove the burden identified throughout this article; it hardens it into permanent infrastructure.
Four objections apply directly:
• It digitizes a requirement that should be discontinued, not automated. If PLO and program-level attainment can be inferred implicitly from transcripts, grades, and assessment weightings already on record, as argued above, then a purpose-built platform for granular, per-student CLO tracking is solving a data-collection problem that a leaner model dissolves. Automating an unnecessary process does not make it necessary; it makes it harder to retire.
• It converts a policy choice into a vendor dependency. Once an institution has invested in licensing, data migration, faculty training, and workflow redesign around a specific platform, the sunk cost makes it politically and administratively difficult to later simplify the model, even if regulators agree the underlying requirement should shrink. The system’s continued existence becomes an argument for keeping the paperwork it was built to manage, regardless of whether the paperwork still serves a purpose.
• It creates exactly the data-governance exposure a lean model is designed to avoid. Centralizing granular, identifiable, question-level performance data on every student across an entire academic career raises real privacy and security obligations, obligations that a transcript-based, program-level inference model never incurs, because transcripts already exist, are already governed, and are not duplicated into a second, less-scrutinized repository.
• It rewards the wrong behavior once faculty start optimizing for it. The moment an information system exists to score CLO attainment, faculty incentives shift toward feeding it clean, defensible numbers rather than toward the teaching judgment the numbers were supposed to represent. This is precisely the reductionist failure mode the literature on outcomes-based assessment has been documenting for over a decade; a dashboard does not fix that failure mode, it industrializes it.
None of this means technology has no role. Aggregated, program-level analytics drawn directly from existing transcript and grading systems, the kind of lightweight reporting layer described in the CQI section above, is legitimate and useful. What is not justified is a standalone, student-level assessment information system built to formalize a compliance model this article argues should be discontinued. Institutions and regulators considering such platforms should ask a prior question before evaluating any vendor: does the underlying data-collection requirement need to exist at all? If the lean OBE model proposed here is accepted, the honest answer for most of what these systems are built to do is no.
The Bottom Line
OBE, as currently practiced in much of engineering and computing education, confused documentation of outcomes with delivery of outcomes, a distinction the international literature on engineering OBE has been raising for over a decade, and one South Africa’s Curriculum 2005 experience turned into a national cautionary tale. Accreditation accords compounded the problem by rewarding process compliance over classroom reality.
Computing does not carry the licensure stakes that justify engineering’s heavier model, and it moves far too fast for a document-heavy accreditation architecture to stay current. The responsible path forward is not to discard outcomes-based thinking, but to contain it: let regulators define the meta-level architecture (PEOs, PLOs, CLOs) once, centrally, and well, then let real data, transcripts, employer feedback, technology trends, and contextual realities, drive continuous improvement, instead of a second, redundant paper trail that exhausts the faculty it was meant to empower.
A framework this lean would do something the current one has struggled to do: make OBE actually about outcomes again.
For NCEAC and HEC specifically, the ask here is narrow and actionable: commission an honest, anonymous audit of how CLO-PLO mapping is actually practiced across a representative sample of computing departments before the next accreditation cycle, rather than assuming compliance equals comprehension. If even a fraction of departments show the midterm-to-CLO1 pattern described above, and colleagues across the sector suggest it will be a large fraction, that is evidence for redesigning the requirement, not for auditing harder against a requirement most departments have already learned to perform rather than practice.
References
1. Woolston, D. C. (2008, October). Outcomes-based assessment in engineering education: A critique of its foundations and practice. In 2008 38th Annual Frontiers in Education Conference (pp. S4G-1). IEEE.
2. Kasilingam, G., Nithiyananthan, K., & Mani, P. R. (2017). Implementation and assessment of outcome based education in engineering education. International Journal of Pure and Applied Mathematics, 117(17), 217–228.
3. Rajaee, N., Junaidi, E., Taib, S. N. L., Salleh, S. F., & Munot, M. A. (2013). Issues and challenges in implementing outcome based education in engineering education. International Journal for Innovation Education and Research, 1(4), 1–9.
4. Mistamiruddin, E., & Nasri, N. (2024). Challenges in integrating outcome-based education (OBE) in higher education institutions: A systematic literature review. International Journal of Academic Research in Progressive Education and Development. https://doi. org/10.6007/ijarped/v13-i4/23378.
5. Spreen, C. A., & Vally, S. (2010). Outcomes-based education and its (dis) contents: Learner-centred pedagogy and the education crisis in South Africa. Southern African Review of Education with Education with Production, 16(1), 39–58.
6. Clercq, F. D. (2020). The persistence of South African educational inequalities: The need for understanding and relying on analytical frameworks. Education as Change, 24(1), 1–22.
7. Jansen, J. D. (1998). Curriculum reform in South Africa: A critical analysis of outcomes‐based education. Cambridge journal of education, 28(3), 321–331.
8. Mahrishi, M., Ramakrishna, S., Hosseini, S., & Abbas, A. (2025). A systematic literature review of the global trends of outcome-based education (OBE) in higher education with an SDG perspective related to engineering education. Discover Sustainability, 6(1), 620.
9. Meyer, J. W., & Rowan, B. (1977). Institutionalized organizations: Formal structure as myth and ceremony. American Journal of Sociology, 83(2), 340–363.
10. Mirza, A. (2022). An effective and efficient implementation of OBE framework within constrained Pakistani environment to attain desired learning outcomes. Sir Syed University Research Journal of Engineering & Technology. https://doi.org/10.33317/ssurj.390
11. Haider, F., Siddiqui, A. A., & Ali, S. M. (2023). Adaptation of outcome-based education system in Pakistan for engineering disciplines and its critical evaluation. Proceedings of the Pakistan Academy of Sciences: Physical and Computational Sciences, 60(1), 1–7. https://doi.org/10.53560/PPASA(60-1)773
12. Derouich, M. (2025). Ensuring outcome-based curriculum coherence through systematic CLO–PLO alignment and feedback loops. Discover Education, 4, Article 486. https://doi.org/10.1007/s44217-025-00915-7
13. Shabbir, A., Rehman, T., Salman, M., Rizvi, S., & Saaludin, N. (2026). Unmasking the impact of teaching workload on learning outcome attainment in outcome-based engineering education: A multi-level analysis of CLO and PLO achievement. Education Sciences, 16(9), 1473. https://doi.org/10.3390/educsci16091473
14. TestGorilla. (2025). The state of skills-based hiring 2025 report. TestGorilla.
