Designing for retention – the ICEBERG model

Last year I worked with some colleagues from the Learning Design team here at the OU on a project focusing on designing for student retention. We of course, have many different aspects in mind when designing a course, but my pitch for this project was that it was worth devoting some time to specifically focusing on how design can influence whether students stay on a course or drop out. When thinking about retention there are, I would suggest, four categories of factors that can impact upon whether a student stays with a course:

1) Design – are there elements in the way that the course is constructed that make it more or less likely that a student will persist?
2) Delivery – when the course is delivered, what support and interventions can influence retention?
3) Personal factors – these can range from whether the student has taken on too much study, changes job, has a shift in personal circumstances, etc
4) Contextual factors – broader context within which the student and course operate within, for instance whether student fees are introduced, if there are increased requirements for this qualification, etc.

The first two factors are the ones most directly under the control of the higher education institution. And in this project, it was just the first one we were focusing on in this particular project (others within the OU are looking at Delivery aspects). So the question was “What can we do during the design of a course to increase the number of students who finish the course?” This is a particular issue for open entry courses, as MOOCs are now discovering.

We conducted research by interviewing course chairs in the Ou where retention had been noteworthy (in either direction), and undertaking literature review across a range of topics, including motivations for learning, MOOC retention data, and analysing our own student’s reasons for withdrawal data. Some of the findings of this might seem fairly obvious, and were part of what most curse teams did anyway, but it’s worth gathering them together and elucidating each clearly I think. My colleague Jitse van Ameijde did some excellent work in gathering all this together into a model. This has the acronym ICEBERG, for seven design elements that can influence retention:

  • Integrated – A well-integrated curriculum so it appears as a coherent whole where all the parts work together in a meaningful and cohesive way. This means that there is constructive alignment between learning outcomes, assessments, activities and support materials which all contribute effectively to driving students to pass the module. I like to think of it as the course having a clear narrative and identity.
  • Collaborative – group work is often stressful for students and difficult to successfully negotiate but there is also good evidence that students tend to persist with a course when they form social bonds with other learners. It also aids understanding of concepts, so courses need to create opportunities for collaboration, which can take different forms, while avoiding some of the frustrations these activities can create.
  • Engaging – An engaging curriculum draws students in and keeps them interested and enthusiastic about their learning journey. This can include varying the types of activities students do, so it’s not one long slog, but also deliberately trying to make the course engaging, for example in the first week providing an exercise that helps them see the relevance or excitement in this subject.
  • Balanced – this is mainly with reference to the workload. Our research showed that excessive workload can correlate with increased student withdrawal, but worse was wildly fluctuating workload. Students like to be able to plan and if what is required varies from one week to the next, this undermines their ability to do so (and often their confidence as a result)
  • Economical – too often the solution when designing a course is to give students more. If they were having difficulty with a concept or an activity we provide more explanation. In order to meet the needs of different learners and perspectives, we give more content than is needed. This can lead to a sense of being overwhelmed and so being economical with what is required and how key information is conveyed is useful for distance learners negotiating their pathways.
  • Reflective – reflection allows students to pull concepts together, and also to understand their own development. It’s important to provide space for this and structured reflective activities, and not just assume it happens. It can also be through the use of informal assessment, including quizzes, to help learners reflect on their own learning and any areas they need to focus on.
  • Gradual – one sure way to lose students is to dump them into complexity. It’s a bit like those “learn to draw” books that go circle, circle with triangle, and then full running horse with flowing mane. Nothing makes you think “this isn’t for me” than a very sudden increase in difficult. A well designed course then has gradually exposes students to increasingly complex and challenging materials, tasks and skills development.

We have tips for how to achieve this, but I’ll save that for another time. While the model is especially relevant to distance ed, open courses, I think it’s applicable for any good course design, whether it’s face to face or online. I should stress this is only one aspect, there are related ways of viewing design to achieve different aspects, so it’s not the only consideration in course design. But it should be one consideration, so we recommend that one meeting is given over where Designing for Retention is the sole focus, rather than being subsumed in other design aspects.


  1. Fascinating stuff. Student retention is a very important issue at my institution (and presumably many others). Do you and your colleagues plan a publication based on your work in this initiative?

  2. Thanks Martin. Great post. I’m doing a course redesign at the moment and these points are important online or offline.

  3. Does this connect in any way to the Iceberg model that is the systems thinking tool, or does it just share the same name? I can see how the behavioural aspects of that model could fit into this one — but is that a whole separate area that isn’t a focus in your model?

    Asking for a friend 😉

Leave a Reply