Remembering Prof. Howard Barrows: Notes on Problem-based Learning and the Schools of the Future

hbarrowsMe: “Howard, can I ask you a more general, philosophical question? Considering all your pioneering work in Problem-based Learning, how would you imagine the school of the future?”

Howard Barrows: “To start with, there would be no subjects. There would be no isolated classes for students such as geography, chemistry, accounting, history and so on.”

Me: “How can this work? What is the point of abandoning specialised subjects?”

Howard Barrows: “It is the designed learning outcomes that drive the learning process and in real-world problems, these outcomes are interconnected. As an example, you come from design. Supposed your students need to produce a TV program, then this is not only about holding a camera and putting it on air, but it is also about casting, accounting, clearing copyright, the ethics of journalism, buying airtime, mastering technology… a whole bunch of learning issues and they are all related. “

Me: “The school of the future would offer highly-integrated projects instead of traditional classes?”

Howard Barrows: “Yes. We would offer complex interdisciplinary projects that already include all the outcomes that traditional subjects intent to convey. Students develop knowledge and skills by achieving the designed learning outcomes. By presenting outcomes in the form of real-world problems, learning becomes more meaningful and relevant to students.”

Hallmarks of Constructivist Active Learning Pedagogy

These were not Howard Barrows exact words as this was not a recorded interview. But it is a truthful account of one of our last conversations during a teacher training workshop in Hamilton, Canada many years ago. Currently, almost two decades, later, Finland started abandoning traditional school subjects in favour of a curriculum reform under the title ‘Phenomenon-Based Learning’, which has much in common with Problem-based Learning (PBL) beyond sharing the same acronym.

(1) Problem-based Learning and Phenomenon-based Learning follow both a constructivist educational philosophy which refers to the idea that knowledge and the meaning of knowledge acquisition are actively created in the learner’s mind. The focus is on students’ critical evaluation of their learning, rather than passively internalising content. (2) Learning is contextual. A real-world problem or phenomenon such as climate change is contextual and requires considering different aspects and perspectives, such as e.g., the quality of people’s lives, mathematics, geography, meteorology, politics and policies, social psychology such as changing consumer behaviour and so on. This is different from most academic problems. Less complex and practical problems such as e.g., running a cafeteria would be another example. Each type of problem-solving or investigation into phenomena requires different kinds of competencies, each problem entails a distinct set of corresponding learning outcomes.

However, not all subjects can or should be replaced, such as e.g., languages, as they are by themselves meta-contextual in nature. Other subjects such as mathematics and music will e.g., still be taught in Finland which is adopting Phenomenon-based Learning. As Prof. Pasi Sahlberg of Harvard University comments “Finland’s National Curriculum Framework is a loose common standard that steers curriculum planning at the level of the municipalities and their schools. It leaves educators freedom to find the best ways to offer good teaching and learning to all children. Therefore, practices vary from school to school and are often customised to local needs and situations.”

(3) The combination of social immersion with autonomous motivation is the key to sustainable, self-directed learning within social contexts. Given the complex nature of most problems, problem-solving is ideally conducted within cooperating teams, rather than competing groups.

(4) The tutorial group structure of PBL avoids oversimplification that easily occurs on an individual level, but it also prevents ‘groupthink’ by encouraging open inquiry and critical, diverse thinking among group members. By discouraging individual power positions, social loafing but encouraging active participation, open inquiry, and consensus based on the better argument, tutorial groups resemble an approximation to what the social philosopher Jürgen Habermas called in his early philosophy ‘ideal speech situation’, which he later concluded as Discourse Ethics. Habermas’ doctrines almost read like a PBL tutorial guide, e.g., ‘Every subject with the competence to speak and act is allowed to take part in a discourse‘ (principle of social inclusion) or ‘Everyone is allowed to question any assertion at any time’ (principle of open inquiry).  In classroom practice, real world problems appear more meaningful to students as compared to decontextualized abstract tasks and thus support their intrinsic motivation.

pbl2Image above: Outline of the core PBL process for a tutorial group. It involves the stages of the group setting, problem identification, idea generation, the identification of learning issues, self-directed learning (research), research review, solutions development and the final self-assessment of learners of their individual and social roles. Educators such as Donald Woods have diversified the process structure for numerous faculties.

(5) In constructivist educational philosophy, reality is based on multiple representations allowing for multiple and multi-faceted solutions to emerge – unlike the single ‘model answer’ in traditional education. During the research and development phase, students become aware of different conceptual and practical approaches. Final solutions are the result of reiterative rounds of research and synthesis by the team, they are not spontaneous assumptions based on somebody’s pet ideas.

When we talked about grades, a hallmark of traditional education, Howard Barrows put forward a rhetorical question “If a pilot is taking 250 passengers from Frankfurt to New York, and you ask him how he feels, what would you think if he answers ‘Like a C minus’?” If an industry representative would ask me which of my students is able to do the job – wouldn’t it be ideal if I can honestly answer ‘All of them, all of my students can perform the job well’? Besides, what does a grade tell us? For example, if a student has obtained a ‘B’ in web-design, is this because she could handle the backend well, or the graphical user interface, or the integration into a database or any combination thereof? Does a grade tell us how a student was performing as a team member, a problem-solver or as a researcher? Traditional assessment is disappointingly un-informing when we take the requirements of a 21st-century workplace as a criterion. As compared to a mere summative assessment, assessment rubrics and formative assessment procedures have undoubtedly narrowed the gap to a fairer and more efficient evaluation, but social skills and research skills are rarely part of academic appraisals.

What is a ‘Problem’ and what is its function?

My colleagues and Howard also had intense discussions about the term of a ‘problem’. Doesn’t ‘problem’ sound too negative? We rarely perceive problems as something we would wish for. Problems can be of high and low complexity and they come in many forms, shapes, and sizes. Some problems might not be considered problems per se, but challenges. A product that works perfectly well without problems may just need an update to keep up with the times, or people might want to discuss openly competing models of desired social futures. More than often, we pose general questions even when we do not face an immediate functional problem, e.g., how do we conceptualise social fairness? What makes a happy childhood? What do we wish for the future of our communities? What is our political utopia?

At the end, the terminus ‘PBL’ was simply too well established in academic discourse so that alternate proposals, such as e.g., ‘Challenge-based Learning’ never gained popularity. For now, it needs to suffice that we talk about ‘problems’ in an extended sense meaning that issues can be of quantitative as well as qualitative nature. In a Kantian understanding, problems and system design issues can represent instrumental as well as non-instrumental (pure- intrinsic) types of rationality. In each case, it is the problem or issue at hand that drives the learning process, not the lecturer or teacher. From a research perspective, theoretical frameworks can complement PBL (by e.g., integrating Critical Theory, Social Conflict Theory, minority group perspectives, Social and Ecological Sustainability, Prospect Theory etc.), depending on the decided research approach.

In the light of terminology, ‘Phenomenon-based Learning’ as in the Finish example represents a limited signifier as well. Phenomenology typically refers to the particular structure of consciousness from a subjective, first-person view. It is a good term in a sense that it emphasises the active construction of new knowledge by a learner, but it neglects one of the key principles of constructivism which is the construction of knowledge with others in social context. It is in intersubjectivity, the reciprocal exchange of perspectives, where learning takes place. Related psychologists in support of constructivist learning are Albert Bandura (Social Cognitive Theory) and Lev Vygotsky. Vygotsky contributed an interesting perspective to learning by differentiating areas where a student can learn unaided, where a student can learn with assistance and a ‘Zone of Proximal Development’ (ZPD) where learning takes place in terms of a coherent social scaffolding for knowledge creation.

Unlike traditional education, PBL questions its own assertions by promoting metacognitive skills in students’ reasoning and epistemology. A probing tutor asks for the grounds and justifications of group members’ reasoning and asks how group members truly know what they assume to know. In academia worldwide, PBL has established itself as a sound pedagogy to address the deep social, economic, ecological and technological challenges of the 21st century.

Extraordinarily great pedagogy requires extraordinary effort: What is the price to pay?

Since much of what has been claimed about constructivist learning may sound too good to be true, the question arises if I had ever experienced any drawbacks. From my experience, there are critical institutional and individual obstacles to overcome. Constructivist pedagogy requires substantial training of staff and it demands a likewise substantial commitment by the education provider to promote student-centered pedagogy. Policy implementation should be conducted across the institution based on a voluntary buy-in. It requires support by consultants and needs a clear normative endorsement by management. In summary, the implementation of constructivist learning pedagogies is a huge collaborative effort.

On teacher level, resistance to new modes of curriculum delivery is rather common. Teachers that have been delivering classes by rote learning their entire careers often feel reluctant to let go of their cosy classroom authority. Many traditional teachers are not ready to give up their role as classroom authority and to embrace the more meaningful role as a facilitator of students’ learning. They feel that PBL is depriving them of their privileged status. To them, the classroom is their private kingdom. On the other hand, each school or college has some teachers that are motivated to try out learner-centered pedagogy. There are opportunities for supporting and recognising such teachers, to offer them the opportunity to become role-models for others, to organise’best practices’ sharing sessions and to publicise student feedback to persuade those that are resistant to change. On a university level, empirical studies might convince more scientifically oriented staff.

Another issue to resolve is the topic of prior knowledge. How much of prior (mostly conceptional) knowledge should be taught to students as a professional ‘body of knowledge’before they can embark on PBL? Especially students from disadvantaged socio-economic backgrounds deserve to start on the same level as their better-off peers. Related academic discussions contributed new ideas to the qualifying issue of prior knowledge such as e.g., initial lectures on common underlying concepts, ‘lectures on demand’ (by external experts, not necessarily lecturers) or ‘flipped classrooms‘, among many others. During policy implementation, it is critical to carefully balance the dissemination of prior knowledge with classroom learning expectations and to explain to all academic staff how both aspects complement each other fairly.

As a memorable experience, I remember my very first PBL class at Temasek Polytechnic in Singapore. Long after the class ended, I had to chase students literally out of the classroom (they ran overtime and the room was booked for other students) as they were still passionately immersed in discussions. I asked myself when was the last time that I saw students truly thrive this way. International studies have confirmed PBL as a superior and more efficient pedagogy as compared to traditional teaching, but practitioners and consultants are in rare supply.

Howard Barrows died in March 2011. To me, he was one of the most influential mentors, reformers, and innovators in contemporary educational philosophy. The list of groundbreaking innovations that PBL brought to the most diverse fields of education is incredible. It ranges from establishing students’ personal responsibility for their learning to learning in structured tutorial groups, employing simulated clients (today we also use computer simulations) and basing curricula on real-world multidisciplinary issues. Most importantly, Howard Barrows’ concept of PBL contributed to empowering learners of all ages by taking an active role in knowledge construction. PBL brought back meaning to students’ learning by introducing process-based solutions development for real-world applications. To conclude with a perspective of Albert Bandura, people are not merely onlookers of their behaviour, but proactive subjects capable of creating desired futures.



Barrows, H. S. (1971). Simulated patients (programmed patients): The development and use of a new technique in medical education. Springfield, Ill: Thomas.

Barrows, H. S., & Tamblyn, R. M. (1980). Problem-based learning: An approach to medical education. New York: Springer Pub. Co.

Barrows, H. S. (1992). The tutorial process. Springfield, Ill: Southern Illinois University School of Medicine.

Barrows, H. S. (1996), Problem-based learning in medicine and beyond: A brief overview. New Directions for Teaching and Learning, 1996: 3–12. doi:10.1002/tl.37219966804

Barrows, H. S., & Wee, K. N. L. (2007). Principles & practice of aPBL. Singapore: Pearson Prentice Hall.


Why it is Time to Retire Bloom’s Taxonomy


Picture above: Exam among Chinese Students (Source: Tomo News)

“You cannot teach today the same way you did yesterday to prepare students for tomorrow. ” John Dewey

1. Historical Credit and Positioning

Bloom’s Taxonomy of Learning has reigned as one of the most influential pedagogical concepts for the design of school curricula until today. Formulated by Benjamin Bloom and colleagues in the mid-50s (Bloom et al., 1956), the taxonomy attempted to break away from behaviorist theories as well as learning via remembering (rote learning) by promoting higher-order thinking skills, such as analyzing, synthesizing and evaluating concepts. Taking a more holistic approach, the taxonomy includes the cognitive- (knowledge-based), affective- (emotive-based) and psychomotor (action-based) domain which explains its intuitive appeal to many teachers. We do not only learn with our heads but also by our actions and emotional experiences that reinforce cognitive processes and give them meaning.

In 2001, Anderson & Krathwohl (2001) published a revised edition of Bloom’s Taxonomy, suggesting that in the cognitive domain, creation appears as a higher-order process as compared to evaluation (ISU, 2017).

2. Limitation of Bloom’s Taxonomy

The most commonly voiced out critique to the taxonomy is that thinking does not operate within hierarchies, but that cognition and affect are neurologically and phenomenologically distributed processes that can assume a plethora of possible configurations. Additional reasons that cast doubt on the usefulness of Bloom’s taxonomy as a pedagogical concept shall be outlined in the following.

2.1 Lack of Scientific Validity

Currently, Bloom’s Taxonomy is more than 60 years old and it had been developed before extensive empirical research into cognition, metacognition and motivation were conducted. As such, the taxonomy’s main categories (Knowledge, Comprehension, Application, Analysis, Synthesis, and Evaluation) are not supported by empirical research on learning, be it as a category or as a category within a hierarchy ranging from lower to higher-order thinking skills.


The only terms of Bloom’s taxonomy that are validated by research are factual-conceptual knowledge (described in modern pedagogy as ‘prior knowledge’) as well as procedural- and metacognitive knowledge. In psychology metacognition is further differentiated into metacognitive knowledge, metacognitive regulation and metacognitive experiences (Efklides, 2006; Schraw & Moshman, 1995; Schraw et al., 2006) and it can appear in the form of individual metacognition, the reflective thinking related to mental content ‘due to me’, or social metacognition, the reflective thinking related to thinking about mental content ‘due to others’ (Briñol & DeMarree, 2012; Kim et al., 2013).

The question remains if lower- and higher-order thinking skills exist as such. A closer look questions this assumption. Some examples: The hierarchy of Bloom’s cognitive domains is broken in the case of (a) a problem-solving scenario where a solid comprehension of basic facts may outweigh an evaluation that is based on biased perceptions (besides, where does comprehension stop and where does evaluation start since both processes work reciprocally) or (b), where a concept is tested for the robustness of its causal and conditional relations (analysis) in order to obtain approval (final evaluation). In such case, analysis and evaluation are interdependent and one cannot be confirmed without acknowledging changes in the other.

If e.g., a situation is evaluated as problematic then this stimulates analysis on how to deal with it, entailing a subsequent evaluation of potential solutions. Even if a final solution is decided this leads to a retrospective analysis and check on the efficacy of the applied solution – and so on and so forth. There is, strictly speaking,  neither a clear-cut hierarchy nor sequence of cognitive processes since we are dealing with interactive, mutually dependent processes: no analysis without prior evaluation, no evaluation without prior analysis.

Creation, to comment on the revised taxonomy (Anderson et al., 2001), does also not necessarily constitute a higher order domain when underlying data analysis and conclusions of a project are faulty. Without proper research, creations remain guesses and assumptions. More than often, people try to promote their pet ideas and care little about thorough procedural solutions development. How do we determine the value of creation? In research, deductive and inductive reasoning are interdependent: we cannot blindly analyze things without an initial sense of intuitive comprehension (such as notions of purpose) and we cannot comprehend things without some sort of prior evaluation. Inductive-empirical and deductive-theoretical inferences relate reciprocally. The more tightly deductive and inductive inferences relate, the smaller the margin for error in research and development. For this reason, assuming a static hierarchy of domains like in Bloom’s Taxonomy is not helpful.

2.2 Lack of an Epistemological Base

Knowledge creation and relating thinking skills do not exist as a priori phenomena, but they are evoked and engaged by people. Knowledge is a foremost social construct while learning is facilitated by social processes (Bandura 2001, 2006). In this light, Bloom’s Taxonomy does not take into consideration the social relation of persons in the creation of knowledge. This includes crucial aspects such as the motivation to acquire knowledge, reiterative and diverse cycles of research, dynamics of open inquiry, the validation of related arguments or the ongoing refinement of concepts within teams. Bloom’s Taxonomy tells us nothing about the role that learners play in knowledge acquisition and creation, including a learner’s intellectual values, the psychological effects of learning experiences, individual differences in cognitive processing, or the communicative processes involved in research and development. Bloom’s Taxonomy does not explain how people collaboratively create, manage and modify knowledge.

Epistemological questions ask things like ‘How do we know that we know?’ or ‘How do we make sure that our knowledge is valid, reliable and relevant?’ The answers to such complex, but critical questions cannot be concluded by attributing general categories (e.g., ‘to analyze’, ‘to synthesize’), but via open deliberation among multiple learners.  Assessment cannot be based on ticking boxes of which cognitive domains have been covered by a student, but by assessing the quality of underlying reasoning.

2.3. Practical Disadvantages and Methodological Flaws

Other potential disadvantages of applying the taxonomy in curricula are (a) the lowering of expectations for higher-level deliberation and reasoning among students by ascribing complex, interrelated processes to simple domain identifiers (b) creating a false notion of ‘higher order’ versus ‘lower order’ outcomes. The taxonomy misleads educators to apply these perceived categories in separation, hampering a natural flow of logical reasoning such as in group discussions and (c) the identification of cognitive processes within an individual learner makes little sense. Instead, a student project can be structured according to logical stages, such as problem identification, problem reframing, identification of learning issues, self-directed research, research review, solutions development, solutions presentation and team/ self- review.

Bloom listed specific ‘action verbs’ that he claims are identifiers for the main cognitive domains, but it is easy to demonstrate that such simple correlations using ‘action verbs’ are misleading. For example, if we take Bloom’s domain of ‘evaluation’ in isolation and only look at action verbs, a student may e.g., ‘compare’ facts without involving analysis, ‘describe’ a phenomenon without explaining its underlying causality and context, or ‘justify’ an argument without giving valid reasons to why is should be believed. It is the power of interconnected, reflected and articulated reasons that drive cognition, not the mere presence of verbs.

3. Conclusion

Educators are looking for evidence-based strategies to enhance their students’ learning. Since Bloom’s Taxonomy is neither based on scientific findings nor offers an epistemological base that explains how knowledge is specifically created and modified within a socio-cultural context, it provides little reason to why it should be employed in educational settings.

The advantages of a constructivist approach, by contrast, are obvious: what matters is not the categorical identification of cognitive processes for the sake of ticking boxes. What matters is to determine how cognitive constructs have been assembled by the learner, which reasons and motivations went into the formulation of mental content and how knowledge-creation ties into larger meaningful frameworks such as cultural identity, human relationships, consensus finding, policy making, or the advancement of local and global communities.

In closing, Bloom’s Taxonomy, despite its historical merits, should be retired as an educational philosophy on the following grounds:

  • The taxonomy is not empirically validated
  • The taxonomy focusses on abstract cognitive domains rather than on learners. The taxonomy is not learner-centered and does not answer questions regarding a learner’s autonomy, competence and social relatedness (Deci & Ryan, 2012), all critical to learning.
  • Real-life contexts and their relevance for knowledge creation are not part of Bloom’s taxonomy
  • The taxonomy does not take into consideration the meaning that knowledge creates for a learner or a community of cooperating learners
  • The role of prior knowledge is not operationalized from an epistemological perspective
  • Motivation, the key component to learning, is not part of the concept
  • Individual differences in learning styles and attitudes remain unaccounted for
  • The taxonomy provides no sensible, specific criteria for assessment, such as evaluating students in their role as team members, researchers, and problem-solvers
  • Thinking processes are not based on strict sequences or hierarchies. Depending on the kind of problem at hand and its complexity, learners structure affective, cognitive and metacognitive processes accordingly
  • The obsession with individual cognitive skills and processes is often exercised at the expense of personal development, social skills, communication skills and the development of cooperative behavior

In all fairness, we have to consider that Bloom lived in a time and culture that celebrated uncompromised individualism. Bloom still shared the assumption of solipsistic learners whose learning can be objectively measured by a clear-cut hierarchical taxonomy. Empirically validated theories of social learning and studies investigating intrinsic versus extrinsic motivation or cognitive construction had not yet appeared on the horizon when Bloom worked on his taxonomy.



Anderson, L. W., & Krathwohl, D. R. (2001). A taxonomy for learning, teaching, and assessing: A revision of Bloom’s taxonomy of educational objectives. New York: Longman.

Bloom, B. S.; Engelhart, M. D.; Furst, E. J.; Hill, W. H.; Krathwohl, D. R. (1956). Taxonomy of educational objectives: The classification of educational goals. Handbook I: Cognitive domain. New York: David McKay Company.

Bandura, A. (2001). Social Cognitive Theory: An Agentic Perspective. Annual Review Of Psychology, 52(1), 1

Bandura, A. (2006). Toward a Psychology of Human Agency. Perspectives on Psychological Science, (2). 164.

Briñol, P., & DeMarree, K. G. (2012). Social metacognition. New York, NY: Psychology Press.

Deci, E. L., & Ryan, R. M. (2012). Motivation, personality, and development within embedded social contexts: An overview of self-determination theory. In R. M. Ryan (Ed.), Oxford handbook of human motivation (pp. 85-107). Oxford, UK: Oxford University Press. doi: 10.1093/oxfordhb/9780195399820.001.0001

Efklides, A. (2006). Metacognition and affect: What can metacognitive experiences tell us about the learning process? Educational Research Review, 13-14. doi:10.1016/j.edurev.2005.11.001

Iowa State University (2017). Revised Bloom’s Taxonomy. Retrieved from:

Kim, Y. R., Park, M. S., Moore, T. J., & Varma, S. (2013). Multiple levels of metacognition and their elicitation through complex problem-solving tasks. The Journal Of Mathematical Behavior, 32(3), 377-396. doi:10.1016/j.jmathb.2013.04.002

Schraw, G., & Moshman, D. (1995). Metacognitive Theories. Educational Psychology Review, (4). 351.

Schraw, G., Crippen, K. J., & Hartley, K. (2006). Promoting Self-Regulation in Science Education: Metacognition as Part of a Broader Perspective on Learning. Research In Science Education, 36(1-2), 111-139.