Roughly 10% of primary school-aged kids have poor reading and oral language comprehension skills relative to age-appropriate word-reading skills (Cain, 2009). These kids – poor “comprehenders” – are often not picked up within classrooms because they can sound out (“decode”) words and read sentences just like good reading comprehenders (e.g. Nation et al., 2004).

We need to find these kids early, and then help them to develop their comprehension skills.

Broken links: what goes wrong with comprehension for kids who have no problems reading words?

To understand what we hear and read, we don’t memorise or rote-learn the words one-by-one. Instead, we retrieve the meaning of the words, then link the words and sentences together to make a “mental model” of what we’ve heard or read (e.g. Kintsch, 1998).  Recent research tells us the differences amongst children in language/listening comprehension skills are almost entirely explained by differences in:

(Lervåg et al., 2017).

We’ve talked about the important contributions vocabulary,  syntax, morphological awareness, and verbal working working memory make to reading and language comprehension.

But what do we mean by “inferencing skills”?

Inferencing skills: why they’re vital for good comprehension

The word “inference” comes from the Latin word “inferre”, meaning “bringing in”. Inferences are conclusions we reach (“bring in” ourselves) based on the evidence in front of us and our reasoning.

Inferences are vital to good reading and language comprehension because they help us to build mental models of what we read or hear. As readers or listeners, we must make inferences to:

  • expand on information we hear or read; and
  • connect information we hear and read to:
    • other information we hear and read; and
    • our knowledge about the world,

(Freed & Cain, 2017).

Inferences and how they help us to understand what we hear and read

Different researchers sort inferences into lots of different, overlapping, confusing and jargon-filled categories. In simple terms, there are two main types of inferences that help listeners and readers to understand language:

  • local inferences, also known as “gap-filling inferences”; and
  • global inferences, also known as “text-connection” or “bridging” inferences (e.g. Graesser et al., 1994; Cain & Oakhill, 1999).

Together, local and global inferences are sometimes referred to as “coherence inferences” because they both help readers and listeners to make unified, logical and consistent models of what they read and hear.

The meaning and differences between local and global inferences are best illustrated with a practical example.

Example

First, read this short story:

Today was Nicole and Rob’s wedding. All the cousins were getting dressed up for the ceremony in Rob’s bedroom. Once dressed, they were told to stay in the lounge room and to not go outside. While the kids watched cartoons, Uncle Matthew and Uncle Vince were decorating the marquee that had been set up in the front yard with lights and bunting. Grandma told them to put on their gumboots and raincoats, so they didn’t get wet.*

Now, answer these questions:

  1. Where were the cousins getting ready for?
  2. What was the weather like?

Local inferences fill in gaps within the story

Question 1 requires us to make a local inference. We make local inferences when we use information from the story to understand other information in the story.

For example:

  • to understand the story above, we need to use the information in the first sentence to understand the full meaning of the second sentence. We make a local inference that the cousins are getting dressed up because they are going to Nicole and Rob’s wedding ceremony, and not, for example, that they are going to a funeral or costume party, or simply playing “dress ups”;
  • in the third sentence, we infer that the word “they” refers to the cousins in the previous sentence;
  • in the fourth sentence, we infer that the word “kids” refers to the cousins referred to in the previous sentence; and
  • in the last sentence of the story, we infer that the word “they” refers to the uncles referred to in the previous sentence.

As you may have noticed, sometimes, local inferences are cued by pronouns (e.g. “he”, “she”, “they”, “it”), or synonyms (e.g. “cousins”, “kids”).

As a practical matter, this is one reason parents, speech pathologists and teachers should work hard with their kids on pronouns and vocabulary, especially with children with language disorders or children who are learning English as a second language.

Global inferences connect the story with the real world

Question 2 requires us to make a global inference: we need to take the information in the story and integrate it with what we know about the world (Cain and Oakhill, 1999). We need to figure out that it’s raining from the fact Grandma is telling the uncles to put on their gumboots and raincoats. This is based on our real-world knowledge that we wear gumboots and raincoats when it’s raining. Astute listeners/comprehenders may also link this knowledge with the information that the cousins are not allowed outside and reason that this is because it is raining and the adults don’t want the cousins to get wet or muddy.

As we read or hear the story, we may also make the global inferences that:

  • the cousins are kids (and not adults) from the fact they are being told what to do and are watching cartoons;
  • the story is happening at a house (and not a hotel) from the words “Rob’s bedroom”, “lounge room” and “front yard”);
  • the wedding ceremony is taking place today (not tomorrow) from the fact that the kids are getting dressed up; and
  • the wedding reception (and maybe even the wedding) will take place in the front yard, from the fact the uncles are decorating and the word “marquee”.

When do kids typically acquire inferencing skills; and who is most at risk for delays?

  • Even when they have difficulties comprehending what they hear or read, most children are aware of the need to make local and global inferences from an early age (e.g. Lynch et al., 2008).
  • The ability to generate local and global inferences improves significantly between the ages of 4 and 10 years (e.g. Currie & Cain, 2015).
  • Younger children and poor comprehenders make fewer inferences than older children and good comprehenders (e.g. Barnes et al., 1996; Cain et al., 2001).
  • Children with poor reading and language comprehension skills have problems making inferences (e.g. Caig & Oakhill, 1999; Bishop & Adams, 1992).
  • Interestingly, on average, children may be better at making global inferences than local inferences. This may be because:
    • children are more likely to remember information central to the overall meaning of the story, than they are to information that is peripheral;
    • some children with comprehension problems may use a strategy to focus on the “big picture” or “gist” of a story; and/or
    • there are usually fewer clues in the text for making local inferences than global inferences (e.g. Currie & Cain, 2015, Miller & Keenan, 2009; Freed & Cain, 2017).

How do we typically assess inferencing skills?

In our clinic, we typically administer valid and reliable standardised, norm-referenced oral language and reading assessments, e.g. the:

  • Clinical Evaluation of Language Functioning (4th and 5th Editions) (Semel et al., 2006; Wiig et al., 2013) for language, including the Understanding Paragraphs Sub-test; and
  • York Assessment of Reading for Comprehension (Snowling et al., 2009) for reading comprehension, including two reading passages.

Both these tests include questions to probe local and global inferencing skills. The comprehension questions are asked after the child has heard or read the entire story, article or other text.

One criticism of this “whole story” approach is that it doesn’t necessarily assess a child’s potential to make inferences as the story is read or heard. Some terrific new research by Dr Jenny Freed and Professor Kate Cain has suggested that we should think about taking a more “dynamic” approach to assessing kids’ inferencing skills, similar to what we do in other areas of speech language pathology and in teaching (2017, see citation below).

This would involve us (for example) breaking down (or “segmenting”) stories into parts and asking specific questions at the end of each segment during the story – rather than after it – to reduce memory demands and to activate comprehension strategies as the story is read. This is a great idea, and I’m looking forward to the development of standardised, norm-referenced assessments based on dynamic assessment principles.

How can we help kids to improve their inferencing skills?

There’s evidence supporting a range of adult-guided techniques to improve children’s inferencing and comprehension skills (e.g. Kang et al., 2015). These include:

  • Shared (“dialogic”) reading: This is a useful approach for toddlers, preschoolers and young school-aged children. Shared reading may develop kids’ use and comprehension of complex language structures (e.g. Cameron-Faulkner & Noble, 2013) and to make “causal” elaborations (e.g. Makdissi & Bosclair, 2006). We use shared reading in our Hanen therapy for late talkers, PreLit Program for preschoolers, and in language therapy, generally, including with school-aged kids with language disorders. It is not difficult to do with a bit of practice. (You can read more about how to do shared reading, and the evidence supporting it here). There’s some very promising new research going on right now about improving children’s inferencing and comprehension skills with shared reading interventions, including by Australia’s own Emily Dawes. It can also be used as a vocabulary intervention. Fair to say, I’m watching this space keenly!
  • “Think-alouds”: This requires children to explain, after listening to each sentence of a story, what they have understood so far. Like Freed & Cain’s suggested approach for assessments, think-alouds explicitly segment the processing of a story. Research shows that think-alouds help both good and poor comprehenders to create inferences (e.g. Laing & Khami, 2002). However, the approach can be cognitively demanding, especially for kids with significant developmental language disorders.
  • Segmented story-telling with questions: This is similar to thinking-aloud, but with specific questions asked at the end of each segment (rather than open-ended “What do you think is happening?” questions). It can also be seen as a “more structured” version of shared book reading. Based on Freed & Cain’s 2017 study, it appears there is a clear and strong benefit of breaking stories up and asking specific questions as kids read or listen to a story for younger, typically developing kids up to about Year 3, perhaps because it encourages kids to monitor their comprehension as they go.
  • The “why” technique: This is also called “elaborative interrogation”. It sounds a little foreboding, but it’s actually pretty simple to implement. You ask, or teach kids to ask themselves, a “Why?” question at the end of each sentence (e.g. Dunlosky et al., 2013). I use it a lot with school-aged children with developmental language delays. You can read more about it, with examples, in section B.3 of my article on evidence-based study skills.

Clinical bottom line

Reading and oral language comprehension is critical for school, work and life success. Comprehension skills for both reading and listening to language depend on a range of sub-skills including vocabulary, syntax, morphological awareness, working memory and inferencing. Some children need help to develop their local and global inferencing skills.

Several evidence-based techniques can help kids to develop inferencing skills, including shared reading, think-alouds, segmented story-telling and elaborative interrogation techniques. Exciting research is happening on new inferencing assessments and interventions, including in Australia. We need to keep a close eye on these developments to make sure that kids with reading and language comprehension difficulties can access the best, evidence-based strategies to improve their skills.

*Example story loosely adapted from the example in Appendix A of Freed & Cain, 2017.

Related articles:

Principal source: Freed, J. & Cain, K. (2017). Assessing school-aged children’s inference-making: the effect of story test format in listening comprehension. International Journal of Communication Disorders, 52(1), 95-105.

Image: http://tinyurl.com/ydbv7xbm

Banter Speech & Language Banter Speech & Language
Banter Speech & Language is an independent firm of speech pathologists for adults and children. We help clients in our local area, including Concord, Concord West, North Strathfield, Rhodes, Strathfield and all other suburbs of Sydney’s Inner West.

Banter Speech & Language is owned and managed by David Kinnane, a Hanen- and LSVT LOUD-certified speech-language pathologist with post-graduate training in the PreLit early literacy preparation program by MultiLit, the Spalding Method for literacy, the Lidcombe and Camperdown Programs for stuttering, and Voicecraft for voice disorders. David is also a Certified PESL Instructor for accent modification.

David holds a Master of Speech Language Pathology from the University of Sydney, where he was a Dean’s Scholar. David is a Practising Member of Speech Pathology Australia and a Certified Practising Speech Pathologist (CPSP).

Posted by David Kinnane

Speech-Language Pathologist. Lawyer. Father. Reader. Writer. Speaker.

Leave a reply

Your email address will not be published. Required fields are marked *

Time limit is exhausted. Please reload CAPTCHA.