Insights from Learning as a Generative Activity Part 3: Learning by Summarising

You can find the first two posts in this series here: one, two.

In the last post, I explored the three different stages of cognitive processing that constitute ‘generative Learning’ and how each one is an essential part of the learning process.

Taken from ‘Learning as a Generative Activity’, the diagram and table below demonstrate how the three stages of processing fit into a model of cognition. The ‘SOI’ in Figure 1.1 refers to ‘Select, Organise and Integrate’.

SOI model of gen learning

SOI table

Each of the eight learning strategies within ‘Learning as a Generative Activity’ involve all three cognitive processes.

Boundary Conditions

One of the strengths of Fiorella and Mayer’s work is that, instead of treating their eight strategies as universally applicable and akin to educational silver bullets, each strategy is discussed in terms of the boundary conditions under which it is most effective. Cognitive Load Theory predicts that instructional choices should be made according to the level of expertise of the student: explicit instruction is almost certainly the correct approach when teaching novices; problem solving and more independent learning is better suited to those with more expertise. This distinction is also recognized in Learning as a Generative Activity. Other boundary conditions include the subject matter that is being learnt-some strategies lend themselves better to specific domains- and whether or not students require a period training or instruction before the strategy is deemed to be effective.

Strategy 1: Learning by Summarising

Creating a summary involves ‘restating the main ideas of a lesson in one’s own words’ and encourages learners to engage with all three cognitive processes. In a summary, students have to select the most important information, organize it into a coherent representation in their working memory, then finally integrate it with their prior knowledge by expressing the information using their own words. Additionally, the summary must be more concise and succinct than the original. Like ‘because, but so’ , writing a summary involves combining language skills with content knowledge, a combination that makes it a particularly powerful strategy for the classroom.

Although creating a summary can be used as a useful retrieval activity, here it refers to the process of generating summaries during learning when the student has full access to the materials that they are learning from.

Boundary Conditions of Learning by Summarising

Learning by summarizing is most effective when students are learning from text based materials and less so when the information being learnt contains complex spatial relations such as concepts in physics or chemistry.

Writing an effective summary is hard and when I taught IGCSE, students would always find it amongst the most difficult skills that they had to learn. As a result of this, summarisation is most effective when students receive training on how to do it.

According to this American Educator article, reading strategy programs-including teaching students how to summarise-that were relatively short (around six sessions) were no more or less effective than longer programs that included as many as 50 sessions. There are diminishing returns to teaching them; they are ‘a low-cost way to give developing readers a boost, but it should be a small part of a teacher’s job. Acquiring a broad vocabulary and a rich base of background knowledge will yield more substantial and longer-term benefits.’

Willingham draws an analogy between comprehension strategies and the need to check your workings in Maths. Firstly, checking your workings and comprehension strategies are tricks that are easy to learn and use, the only difficulty being the need to remember to consistently apply them. Secondly, while checking your work in Maths makes it more likely that you will get a problem right, it won’t tell you how to solve the problem. This applies equally to comprehension strategies like summarisation: strategies don’t achieve comprehension they encourage a student to apply a process. If you are summarizing, you still need to understand enough about the text-which will come from background knowledge-to summarise successfully.

Two types of Summarisation

Summaries can be split into two distinct categories: writer based summaries and reader based summaries. A ‘reader based summary’ is one produced for other people, for example completing a summary question in an IGCSE exam. These summaries require students to use accurate, precise and concise language and they often have to be written using specific sentence constructions.

A ‘writer based’ summary is one that students write for themselves as a study tool and, while both types of summary involve all three stages of cognitive processing required for generative learning, this second type is easier for students to learn as they don’t need to pay so much attention to the concision of their sentences; instead, they can devote their entire attention to making sense of the text that they are summarising.

How can you teach summarisation?

Students often find summaries difficult. Firstly, selecting the correct information from a text can be difficult if the text is lengthy, has low coherence or if students have low prior knowledge. Secondly, organizing the selected information in working memory can also be difficult, especially if the text is complex and lengthy because students will need to manipulate and organize disparate information from different text locations, perhaps spanning paragraphs, section and pages. Thirdly, integrating the information with relevant prior knowledge can be difficult if the content is complex or if there is the additional requirement-as in ‘reader based summaries’-that the student uses specific grammatical constructions and writes with accuracy  and concision.

Here are a few things that can help when teaching students how to summmarise:

Start with ‘writer based summaries’

Because this type does not require the same level of grammatical precision and technical proficiency as a ‘reader based summary’, they are much easier to master. They also provide a useful foundation if you then want to go on to teach students how to do reader based summaries.

Start with simple, short excerpts

Students should experience consistently high success rates and this can be achieved if they begin by using simpler, shorter texts.

Begin with ‘targeted summaries.’

Instead of asking student to summarise a text in a general sense, ask them to focus on something specific thereby increasing the chance that they will select the correct information. Instead of ‘Summarise the writer’s thoughts on nostalgia’, you could ask students to ‘summarise the writer’s thoughts on how nostalgia helps him to overcome homesickness.’ The more specific the task here, the more likely the student will be able to select the correct information. One of the key difficulties when producing summaries is deciding which information is important or not and using targeted summaries with more novice students can help overcome this.

Model and practice each stage of the process initially in isolation before combining the stages into a routine.

Writing a summary is what Engelmann would call a ‘cognitive routine’: it is a process that involves ‘a series of steps that lead to a solution’ where the learner is ‘logically required to process a series of concepts, details or discriminations to arrive at the appropriate solution.’ By initially modelling and practicing each stage of the process in isolation before combining them into a complete process, errors can be more precisely diagnosed and corrected, ensuring that students experience high levels of success from the outset. Asking students to attempt the entire summary writing process from the beginning will almost certainly result in poor output. Not only that, it will be difficult to ascertain exactly why performance has been poor because writing a summary can involve a number of individual steps, each of which can prove difficult for students. These steps include:

  1. Delete trivial information
  2. Delete redundant information
  3. Substitute superordinate terms for lists
  4. Substitute superordinate terms for series of events
  5. Select a topic sentence
  6. Invent a topic sentence (if the text doesn’t include one)

Although some of these are more suited to ‘reader based summaries’ (numbers 5 and 6), it would be useful to model and practice each step in isolation before combining steps and then finally asking students to practice the entire process. Teachers should use example problem pairs and follow a sequence that gradually fades out support to ensure that a consistently high success rate is achieved by all students.

Prevent common misconceptions and errors

Students often include information that they have directly copied from the text into their summary. While they may have selected the correct information and expressed it in a logical order, failing to use their own words may prevent them from integrating it with their prior knowledge. When students are challenged about this, they will often reply that they understand what they have copied or that the author has expressed it in the best way possible using words that have no viable synonyms. Although these replies are often justified, if we want the summary writing to be an effective learning process, students need to transform the words from the text into their own language, thereby integrating the information with their prior knowledge. There are a number of things that we can do to ensure that they do not merely copy:

a) Rewrite short sections of copied text

When teaching students how to summarise for the first time, the teacher can present a sentence or two of copied text and demonstrate how to transform it into their own words under a visualizer. Following the I-We-You continuum, they can then ask students to assist them with the next example before asking the class to attempt a few of their own.

b) Oral questions about the written summary

Many pieces of vocabulary have no suitable or precise synonyms and it is often difficult or impossible to express them in your own words: the more abstract and technical the text, the more likely that a summary will contain words that have been directly taken from it. If this is the case, then the teacher can mitigate the potential lack of thought involved in copying technical terms by asking students what they mean.

c) Annotating vocabulary and paraphrasing definitions before summarising

When reading a text in class, a teacher can provide elaboration and annotation, stopping to explain difficult vocabulary, ask questions and provide additional examples to ensure that students understand what they are reading. Pre-teaching some of the trickier content makes it more likely that students will be able to successfully summarise it.

We often provide definitions for students, choosing words that are hopefully not only new to them but also have high utility and can be used across texts, contexts and domains. Here is an example from an article on nostalgia:

vocab table nostalgia

In order to help students integrate these words with their prior knowledge, the teacher can ask students to read the definition before offering their own paraphrased version of it. This can then lead into a series of questions designed to help students make connections:


Teacher: What does alienated mean?

Student: It means being left out, being alone

Teacher: According to the article, how can nostalgia mitigate feelings of alienation?

Student: It can help you feel optimistic and connected to others.

Teacher: We learnt a similar word when we looked at Exposure that also means being excluded?

Student: Ostracised

Teacher: Why did the soldiers feel ostracised in Exposure?

d) Elaboration

Although writing a summary is about concision (especially when writing ‘reader based summaries’), if it is to be used as a generative learning activity, there is potentially less requirement to be as succinct as possible as the main aim is successful engagement with all three stages of cognitive processing. To prevent students from copying important technical terminology without thought, the teacher could ask them to initially include noun appositives to ensure that they integrate the new vocabulary with their prior knowledge.


Original Text: Is it healthy to dwell in the past? Up until about 15 years ago most psychologists would have suggested probably not. The habit of living in memory rather than the present, of comparing how things once were with how things are now, was for several centuries thought at best a trait to avoid and at worst a root cause of depressive illness. Nostalgia was the soldiers’ malady – a state of mind that made life in the here and now a debilitating process of yearning for that which had been lost: rose-tinted peace, happiness, loved ones. It had been considered a psychological disorder ever since the term was coined by a 17th-century Swiss army physician who attributed the fragile mental and physical health of some troops to their longing to return home.

Targeted summary question 1: Summarise what people used to think about nostalgia.

Summary Example 1: Nostalgia used to be considered a mental illness

Summary Example 2: Nostalgia, a feeling of affection for the past, used to be considered as a mental illness.

While clearly less concise, the second example demonstrates that all three stages of cognitive processing have been engaged with by the student, making it more of a generative activity.

If you want to read more about summarisation, Timothy Shannahan has written two blogs, the first of which is available here.

Additionally, this article gives a helpful overview of how to teach students to summarise.

Next post: Learning as a Generative Activity part 4-Concept Mapping