Updated on
July 2, 2026
A diagnostic guide for UK teachers on implementing reading comprehension interventions like inference training for learners who can decode words but struggle to understand the meaning of a text.
Many teachers are familiar with a particular type of reader. This learner can read aloud with perfect fluency and accuracy. However, when asked what the text was about, they can only offer a blank look or a vague summary. They have mastered decoding, the ability to translate printed words into speech, but they struggle with comprehension, the ultimate purpose of reading.
This gap between fluent reading and poor understanding is a common and complex challenge in classrooms from Key Stage 2 upwards. It highlights a important principle defined by the Simple View of Reading: comprehension is the product of decoding and language comprehension. If either part is weak, understanding will fail.
For learners whose decoding is secure, the difficulty almost always lies in the invisible, cognitive work of making meaning. They struggle to connect ideas, read between the lines, and build a coherent mental model of the text. Inference training is one of the most effective and well-evidenced ways to address this specific difficulty.
Inference training is a structured, evidence-based reading intervention designed to explicitly teach learners how to make inferences. An inference is a conclusion reached on the basis of evidence and reasoning; in reading, it’s the process of figuring out what the author means but hasn't stated directly.
This is not a whole-class strategy. It is a targeted intervention for a specific group of learners:
The core of the intervention is making the implicit act of inferring explicit. A useful analogy for learners is the iceberg. The words on the page are the tip of the iceberg, the visible part. The inferred meaning, the author's intent, and the connections to our own knowledge lie beneath the surface. The job of the reader is to look at the clues in the text and dive down to uncover that hidden meaning.
A common misconception is that comprehension strategies can be taught as generic, transferable skills that work on any text. The evidence suggests this is not the case. While strategy instruction is vital, its benefits depend heavily on the reader's existing knowledge.
Cognitive science shows that making an inference is not a single skill. It's an interaction between the clues in the text and the reader’s background knowledge, or schema, stored in their long-term memory. Without the relevant schema, a reader has nothing to connect the clues to. To infer a character is anxious, a learner must recognise clues like a 'racing heart' or 'sweaty palms'. They then connect these clues to their own knowledge of anxiety.
This explains why the benefits of strategy instruction can plateau quickly (Willingham, 2006). A few hours of explicit teaching can show a learner how to make an inference, but drilling the strategy for weeks on end with disconnected texts yields diminishing returns. The most effective approach, supported by recent meta-analyses, combines strategy instruction with the systematic building of background knowledge (Peng et al., 2023). Inference ability is central to processing discourse and is a cornerstone of all major models of reading comprehension (Elleman, 2017).
A typical inference training cycle runs for 8-10 weeks, with two or three short, focused sessions per week.
1. Identify the Right Learners Use a combination of standardised test data (if available) and classroom observation. Look for learners who have accurate and fluent decoding but score poorly on comprehension assessments. To test understanding, have learners read a short, age-appropriate paragraph. Ask both literal questions ('who' or 'what') and inferential questions ('why' or 'how').
2. Select Appropriate Texts Choose short, high-interest texts. The goal is to focus on comprehension, not stamina. Prepare the text by identifying key points where an important inference is required. Mark these with a sticky note as stopping points for discussion. Novels like Wonder by R.J. Palacio are often used because they are rich with opportunities for social and emotional inference.
3. Model with Think-Alouds The most powerful tool in the intervention is the teacher 'think-aloud'. The teacher reads a portion of the text and verbalises their own internal thought process, making the invisible act of inferring visible and audible. This modelling should be explicit and deliberate.
4. Explicitly Teach Inference Types Show learners the different kinds of thinking they are doing. This might include:
5. Use Graphic Organisers for Guided Practice During the guided practice portion of the session, use a simple graphic organiser to structure learners' thinking. This externalises the cognitive process and provides a scaffold for their thoughts.
6. Move to Independent Practice Gradually release responsibility. In later sessions, have learners identify their own stopping points in the text and fill out the graphic organiser with a partner before sharing with the group.
7. Measure Impact Avoid relying solely on reading age scores. The best measure of success is improved verbal reasoning and articulation during the sessions. Can learners now explain how they know something? Can they justify their inference by pointing to a specific clue in the text and connecting it to their own knowledge? This qualitative data is far more valuable than a standardised score.
Mrs. Adler is working with four Year 7 learners who are all competent decoders but struggle to follow plots and understand character motivations. They are reading Wonder.
Teacher: (Reads aloud) "'Via was furious with Auggie, but she knew he couldn't help it.' Right, let's pause here. The author uses the pronoun 'she'. If we don't figure out who 'she' is, the rest of the chapter won't make sense. I think 'she' must be Via. Let me test that by putting the name back in: 'Via was furious with Auggie, but Via knew he couldn't help it.' Yes, that makes sense. I've used the clues to make an inference."
Later in the chapter, she stops at another point.
Teacher: (Reads aloud) "'Auggie's mother sighed and held his hand tightly.' Okay, let's use our organiser. In the first column, 'Text Clue,' what should we write?" Learner 1: "She sighed and held his hand tightly." Teacher: "Perfect. Now, the next column: 'My Background Knowledge.' What do we know from our own lives about why a parent might sigh and hold their child's hand?" Learner 2: "Maybe because they're sad? Or worried?" Teacher: "Good thinking. So, for our final column, 'My Inference,' what can we infer about Auggie's mother?" Learner 1: "She's feeling worried about him." Teacher: "Excellent. And how do you know? You've connected the text clue, the sigh, with your own knowledge about what a sigh can mean. That's it. That's making an inference."
The graphic organisers used in inference training are a perfect application of Structural Learning's 'Map It' tool. The goal of 'Map It' is to make thinking visible by externalising the connections between ideas. An inference map helps learners structure their reasoning and provides a repeatable process.
A simple and effective inference map could have three columns:
By using this structure, learners are explicitly guided to combine information from the text with their own schema to generate a logical inference. The process becomes systematic, not guesswork.
Inference training is one of the most well-researched reading comprehension interventions. A major meta-analysis found that inference instruction has a reliable and positive impact on both literal and inferential comprehension for a wide range of readers (Elleman, 2017). Intervention research from the UK has also shown that programmes with a strong focus on inference can double the normal rate of reading progress for struggling comprehenders (Lavan & Talcott, 2020).
However, it's important to be aware of the limitations:
| Text Clue (What the author says) | My Background Knowledge (What I already know) | My Inference (What I can figure out) |
|---|---|---|
| "The boy ran from the room, slamming the door behind him." | People often slam doors when they are angry or upset. | The boy is feeling angry. |
| "Her eyes were glued to the television screen." | 'Glued' means stuck fast. When you are very interested in something, you can't look away. | She is completely focused on the television programme. |
Before starting an inference training group, consider the following:
Screening and Selection:
Implementation:
1. How long does it take to see an impact? You should see learners begin to use the language of inference and apply the strategies during sessions within a few weeks. A full 8-10 week cycle is typically needed to see a more stable improvement.
2. Can this be used with a whole class? The core components, like teacher think-alouds and using graphic organisers, are excellent universal teaching strategies. However, inference training as a targeted intervention is most effective in small groups where interaction and feedback can be tailored to individual needs.
3. What's the difference between this and guided reading? While there are overlaps, inference training is more specific and systematic. It has a singular focus on teaching the cognitive strategies of inference, whereas a guided reading session might cover a broader range of reading skills.
4. Is this suitable for EAL learners? Yes, with adaptations. For English Language Learners, inference difficulties can be compounded by vocabulary gaps or unfamiliar cultural contexts. The explicit, structured nature of inference training can be very supportive. Pre-teaching key vocabulary and carefully selecting texts with familiar contexts is important (Chen, 2023).
You do not need a formal programme to begin applying the principles of this intervention. Next lesson, take a single paragraph from a text your class is reading. Before reading, tell students you will show them how to read between the lines. Use a think-aloud to explain an inference, using text clues and background knowledge. Make one piece of your invisible thinking visible.
