Skip to content ↓

Effective Teaching of Inference Strategies: Part Two

This blog post follows on from ‘Effective Teaching of Inference Strategies: Part One’, which presented the preconditions for inference. Once pupils have met these preconditions they are ready to learn how to make accurate inferences. However, research has continually shown that children must be explicitly taught the skills which they need for inference. This blog post will present some of the skills which pupils need to be taught and provide some practical suggestions on how to teach them.

Word Level Work: Vocabulary

Many researchers have highlighted the importance of vocabulary development as a foundation to reading skills such as inference. For poorer readers the attention has to be focused at a word level. Poorer readers often have difficulty using and understanding cohesive devices, such as pronouns and conjunctions, which are crucial in making accurate inferences. A study by McGee and Johnson (2003) provided one suggestion for what a teacher can do to develop pupils’ vocabulary. As part of their research, they trained pupils to explain what individual words and phrases contribute to a sentence. How to find these words was first modelled to the pupils and then pupils worked as ‘word detectives’ to choose key words and explain how these words contributed to the sentence. At the start of the activity pupils did choose ‘useless’ words but with practice they discovered which types of words carry most meaning. This activity allowed pupils to develop their understanding of which words are important to conveying meaning within a sentence.

Text Level Work with Narratives

When reading a narrative, readers rely on the conventional features of a story being present. Features such as setting, themes, plot and resolution guide the reader through the narrative. Familiarity with this structure is part of the background knowledge that a reader brings to the text which allows for accurate inferences to be made. However, research has found that only abler readers use the structural features of a narrative to build their understanding. Poorer readers do not realise the different ways a story can be linked and they are less likely to impose the conventional features of a story onto what they are reading, therefore they do not link up individual events and actions in a story. This research suggests that in order to develop inferences skills in poorer readers we need to develop their knowledge of story structure. This knowledge may be developed through writing stories and not reader per se, however, writing stories with a conventional structure would feed into comprehension skills.

Question Generating

The use of questions, generated by both teachers and pupils, is of huge importance to developing inference skills. Research into adult readers found that competent adult readers ask themselves questions throughout their reading, which aim to solve any gaps in knowledge or contradictions in the texts. Their comprehension is guided by why-questions, rather than what-happens-next or how, when and where-questions. Answers to these why-questions are what guide the inferences made from the text. Therefore, when teachers pose these why-questions to poorer readers they are prompting the sorts of inferences that occur naturally in good readers. However, this does not invalidate the use of other types of questions. Each type of question allows readers to form inferences from the text, but, according to Glasser (1994), the sorts of inferences generated by these other questions are less crucial to the overall understanding of the text. Pressley (2000) also found that why-questions were particularly beneficial to developing pupils understanding and recommended that teachers encourage pupils to ask themselves why-questions as they read a text. Pupils who are trained to ask themselves why-questions as they read are more likely to automatically relate what they know to ideas in the text. One suggestion from the research was that teachers can develop pupils’ ability to ask themselves questions by modelling this process aloud to them. Teachers may do this from an authentic text but could also modify texts to contain inconsistencies, difficult words and conflicts with prior knowledge. This would highlight how asking questions throughout reading can overcome these problems found in the text. Teachers can model how to answer these questions using strategies such as rereading, looking ahead in the text for clarification or using outside sources.

When it comes to questions posed by teachers Van den Broek et al (2002) conducted a study with interesting results. The results found that teachers must exercise caution when asking questions as questions can actually interfere with the inferencing process. Questions posed during reading or immediately after reading can compete with other ongoing processes for the limited cognitive resources which are heavily involved with all aspects of reading. This was especially relevant in younger readers, where the comprehension and recall of a text suffered most when questions were asked during and immediately after reading. It was suggested that questioning can be beneficial to younger readers but when the cognitive demands are reduced, such as when the story is read to them. This evidence produced by Van den Broek et al (2001) does seem to contradict the general research consensus and common classroom practice, however the differing views can be used alongside each other. Three teaching implications can be taken from the research which seem reasonable and allow for questioning from the teacher to develop inference skills. They are; not to interrupt pupils by asking questions during their reading, not to launch into questioning too soon after reading, allow time for the pupil to consolidate what they have read and, finally, pose inference questions to pupils based on texts that you have read aloud to them.

Activation of Prior Knowledge

Research has shown that knowing how and when to use general knowledge to understand a text is an important skill for making accurate inferences. Pupils need to be encouraged to relate information from a text to their own previous experiences, however it is difficult to explicitly teach this skill to pupils. Some research has suggested that there may not be much a teacher can do to influence how pupils use their previous knowledge to understand a text, however some research does put forward suggestions on to how to develop this skill. Although more than 20 years old, Langer’s (1981) model for eliciting and classifying prior knowledge is still relevant. The model follows three phases in a Pre-Reading Plan. They are; pupils generate initial associations, they discuss and clarify their collective knowledge and then they reassess the knowledge and classify what is now known as a result of the discussion. The model allows pupils to extend their own knowledge before they are introduced to the new knowledge, vocabulary and concepts in the text.

Aural Work and Listening Comprehension

Evidence from research found that problems in inference cause problems in comprehension and not the other way around, therefore poor inference skills are more likely to be the cause of problems in comprehension than a result of it. This finding suggests that inference is a separate skill and can be removed from reading and practised in other contexts. Therefore, developing inference skills can be done through other activities, such as speaking and listening activities. Poorer readers may be more willing to engage in inference activities that are not centred around reading, so inference skills can be taught across the curriculum and embedded into a variety of different contexts to engage children in using their inference skills.


Kispal, A. (2008). Effective Teaching of Inference Skills for Reading. National Foundation for Educational Research .