Skip to content

Research Hub Logo

Self-directed learning and project-based learning

Written By: Adam Boxer
9 min read


Most students are only in school for 14 years of their lives. In that time, they are carefully guided, led and occasionally cajoled through the greatest intellectual treasures our society has to offer. Of course, we are constantly looking for ways both to excite our students about this journey and for them to conquer challenging material efficiently and with ease. During school, they have a guide through this process, but once they leave the school gates they are on their own. Questions therefore abound as to how we enable our students to become self-directed and long-term learners. Project-based learning (PBL) is often suggested as a solution: a way to excite and engage your students, to help them learn effectively and to prepare them for life beyond school.

What is project-based learning?

PBL resists simple definition. Over the years, many programmes and approaches have laid claim to the title, but drilling down to the essential characteristics that unite these programmes has provided researchers with quite the challenge. In turn, evaluating whether or not PBL is appropriate or effective might be intractable: if a particular programme that defines itself as PBL is found to be ineffective, adherents of PBL can easily claim that the ineffective programme was never ‘true’ PBL.

Some commonalities do seem to emerge, including:

  • the solving of a central problem or question
  • that problem or question being grounded in the real world
  • the self-directed acquisition by students of new knowledge and skills
  • the project emulating the activities of domain experts in that subject (see, for example, Thomas, 2000).

Identifying the advantages of PBL is equally difficult, with adherents often making reference to concepts like real-life decisions or authentic learning as ends in their own rights. Proponents generally claim that there are learning gains both in the content but also in soft skills like communication, collaboration or knowing how to conduct research or engage experts (see for example, Boss (2011) or Vega (2012)).

One common justification for PBL is the benefits to student engagement (Hasni, 2016). In the movement away from teacher-led delivery of content, the hope is that students more fully engage with content that they find meaningful and relevant, and through exercising their own autonomy, gain some kind of ownership over the material, bringing with it feelings of empowerment and intrinsic motivation. Of course, engagement is rarely a strong justifier for classroom practice, as we know that students are often at their most engaged when they are dealing with material they are already familiar with, and it could be that other approaches aimed at building student mastery might be more effective in the long term (Boxer, 2019). In my experience from running and delivering projects, students can be engaged by these, but that engagement could be better described as ‘busyness’ – where students are clearly active, but not necessarily productive in terms of mastering a curriculum.

What does the evidence say?

The evidence tends to indicate issues surrounding implementation, establishing evaluative criteria, poorly-controlled studies and a lack of reasonable comparisons, with a recent review finding that: ‘Evaluations of its effectiveness have been hampered by the paucity of valid, reliable, and readily usable measures of the kinds of deeper learning and interpersonal and intrapersonal competencies that PBL aims to promote’ (Condliffe, 2017).

In a recent trial, the Education Endowment Foundation (EEF) investigated a PBL approach in eight English schools (Menzies et al., 2016). Their findings fail to support the intervention, noting that there was no impact on pupils’ literacy or engagement with school and learning. Perhaps more worryingly, the evaluation indicated that PBL may have had a negative impact on literacy for pupils eligible for free school meals. It is also worth pointing to the attrition rate of the programme, with half of schools dropping out part way through the trial. Teacher and researcher Greg Ashman speculates that this rate indicates that the true effect would actually be worse than that measured; if a trial was going well in a school and the teachers were seeing the benefit, why would that school drop out? (Ashman, 2016). Regardless of the truth of that speculation, it appears that there is limited robust evidence in favour of PBL.

One interesting source of evidence is teacher questionnaires, with teachers often reporting qualitatively positive outcomes like improvements in problem solving skills, collaborative skills and critical thinking (Thomas, 2000). In one study (Tretten and Zachariou, 1997), despite teachers being positive about PBL, they gave a very low rating to the statement ‘I believe they learn important knowledge/content’, indicating perhaps that those teachers recognised that this learning didn’t take place, but was less important than the other benefits of PBL. In his review, Thomas (2000) concludes that because such reports rarely match up to any objective outcome, they should be used with extreme caution when justifying PBL; how we feel things went is all too often clouded by numerous biases, and can lead us to the wrong conclusions.

The counter-claim

In 2006, cognitive scientists Kirschner, Sweller and Clarke published their now well-known article ‘Why minimal guidance during instruction does not work’ (Kirschner, Sweller and Clark, 2006), in which they applied basic cognitive load theory (CLT) to a number of instructional methods – the common denominator between the methods being that content was not taught directly by teachers but encountered in a student-led way. Harnessing decades of evidence from psychological sciences (as opposed to educational research), they argued that learning was defined as a change in long term memory (LTM), a change that could only occur if content had been properly processed in working memory (WM). The central contention for teachers is that the WM is highly restricted, and only a few items can be stored and processed at any given moment. If too much information is transmitted at once, the ‘load’ becomes too high and learning cannot occur.

There are many ‘tricks’ available to teachers to circumnavigate this constriction, one being the use of prior knowledge. If you are teaching a topic and students already know quite a bit about it, then that will help them in acquiring new information about that topic. Students with little knowledge of a particular topic are called ‘novices’, whereas students with more knowledge are called ‘experts’. As such, novices need material to be broken down into very small chunks – they have no store of information that can help alleviate the cognitive load, so if they are given too much at once, they cannot process or learn. This would seem like a fairly strong opposition to any approach – including PBL – in which material was not broken down by an experienced teacher but was instead presented to students wholesale through textbooks, newspaper articles, videos or the like. This theory accounts for the empirical evidence marshalled by the scientists showing that approaches with minimal teacher guidance or explicit teacher-led instruction are less effective at promoting learning.

As I have grown more experienced and more in touch with the evidence base, I have found that my students learn more and enjoy my lessons more when they are fully guided; when they have the security of clear routines and habits, and when I carefully sequence and explain material in a way that allows them to understand some of the most challenging ideas our curriculum has to offer. All too often when presented with my students’ errors or misconceptions, I can trace them back to my own activities: did I give them too much information at once? Did I leave any material to chance? Did I assume they had a certain piece of background prerequisite knowledge which in reality they didn’t?

Students as experts?

As mentioned above, one of the arguments in favour of PBL is that it allows students to emulate domain experts, like professional scientists or historians, for example. However, the logic presented above would argue that there is a fundamental, categorical, difference between the novices we teach and the experts they may one day become. Professional scientists do indeed tackle problems like how bird migratory patterns have changed, but the reason they can do so is because of the vast knowledge of the topic they bring to bear. Their knowledge of global weather patterns, seasonal change, nesting grounds and distribution of appropriate food allow them to bring solutions to the problem that are rich, sophisticated, and accurate. Could we reasonably expect a 13-year-old to be able to do the same?

Professor Paul Kirschner argues that students in school are novices, and it is not their role to emulate experts, but to be initiated into a domain: they are not doing science or history or philosophy, they are learning those subjects so that they might one day be able to do them (Kirschner, 2009). This is a philosophical argument about the purpose of education. It is perfectly conceivable that one could respond to Kirschner by accepting that students will learn little of substance about migratory patterns, but the experience of acting as a miniature scientist is empowering in its own right.

The expertise reversal effect

We noted earlier that novices learn best through teacher-led instruction, but as time goes on and they progress on the path to becoming expert, the equation starts to flip. An expert in a subject does not need a teacher to break down new material in that subject for them; they can rapidly assimilate that material from primary sources in a way that is self-directed. Researchers term this the expertise reversal effect (Kalyuga, 2009) and it accounts for a number of important findings in instructional design. Of course, for the practising classroom teacher this means that PBL approaches are more likely to work on topics that students are already familiar with. One could argue that such an approach removes one of the central planks of PBL (the self-directed acquisition by students of new knowledge and skills), however, it is at least conceivable that a teacher could design a task that relates heavily to material students are already familiar with, but then pushes them significantly further into a real-world context of inquiry.

Into the classroom?

Based on my experience in the classroom and my engagement with the research, it is fairly clear that PBL should not be used as a basis for the teaching of new material. The risks of patchy and misconception-filled learning are too high for novice learners. However, it does seem reasonable to allow students, once in a while, to branch out beyond the traditional confines of the curriculum and the teacher-led classroom and encounter the boundaries of the knowledge they have already acquired through expert teacher instruction. We just need to be realistic about the impact of such activities, and the time they take from the necessary work which we are called on to perform: taking ambitious, challenging, complex and beautiful content and breaking it down in a way that our students can understand.

Key questions:

  1. How can I account for my students’ progression from novice to expert?
  2. In a lesson that I felt went well, what is that feeling based on? How could I get better evidence?
  3. When is it appropriate to treat students as beginners/novices, and in what circumstances should I allow them to emulate the behaviour of experts?



Ashman G (2016) Project based learning: further analysis. In: Filling the Pail. Available at: (accessed 31 May 2020).

Boss S (2011) Twenty ideas for engaging projects. In: Edutopia. Available at: (accessed 31 May 2020).

Boxer A (2019) What is the best way to motivate students in your subject? Impact 5: 10-11.

Condliffe B (2017) Project-based learning: A literature review. Working Paper. MDRC.

Hasni A, Bousadra F, Belletête V, et al. (2016) Trends in research on project-based science and technology teaching and learning at K-12 levels: A systematic review. Studies in Science Education 52(2): 199-231.

Kalyuga S (2009) The expertise reversal effect. In: Managing Cognitive Load in Adaptive Multimedia Learning. London: IGI Global, pp. 58-80.

Kirschner PA (2009) Epistemology or pedagogy, that is the question. In: S Tobias and TM Duffy (eds). Constructivist instruction: Success or failure? Abingdon: Routledge, pp. 144-157.

Kirschner PA, Sweller J and Clark RE (2006) Why minimal guidance during instruction does not work: An analysis of the failure of constructivist, discovery, problem-based, experiential, and inquiry-based teaching. Educational Psychologist 41(2): 75-86.

Menzies V, Hewitt C, Kokotsaki D et al. (2016). Project based Learning: Evaluation report and executive summary. London: Education Endowment Foundation.

Thomas JW (2000) A review of research on project-based learning. Available at: (accessed 15 June 2020).

Tretten R and Zachariou P (1997) Learning about project-based learning: Assessment of project-based learning in Tinkertech schools. San Rafael, CA: The Autodesk Foundation.

Vega V (2012) Project-based learning: Research review. In: Edutopia. Available at: (accessed 31 May 2020).


Share your experiences with educators globally by joining the discussion below. When do you think it appropriate to treat students as beginners/novices, and in what circumstances might you allow them to emulate the behaviour of experts?

    4 1 vote
    Please Rate this content
    Notify of
    Inline Feedbacks
    View all comments

    Other content you may be interested in