Reading comprehension depends firstly on our ability to decode the words on the page, and then to make sense of each individual word and groups of words joined into sentences, and then to make connections between groups of sentences. This requires you being able to quickly and accurately retrieve information from your long-term memory and to simultaneously manipulate the information contained in the sentences in your working memory.
Long-term memory involves the storage of information over an extended period of time so that it can be accurately and rapidly retrieved on demand. In terms of reading, you need to be able to rapidly recognise a string of letters as meaningful and relate it to the pronunciation of a particular word. Once the word has been recognised it needs to be rapidly associated with the correct meaning of a word within the context of the passage and then you need to be able to link this information to knowledge of the topic held in your long-term memory to support you in understanding the words you are reading.
Working memory is the ability to hold and process newly presented information. Good working memory is required in order for you to make meaning of strings of words (i.e., sentences) and then hold this information in your working memory long enough to be able to integrate this information with proceeding sentences and to your own knowledge of the world while monitoring all this information to ensure it makes sense.
An example:
Mary pulled the bottle from her school bag. The water was refreshing.
What was in the bottle?
Did Mary drink the water?
To answer these questions correctly, you must first be able to read and understand the meaning of each of the words in the two sentences by retrieving this information from your long-term memory. Then, you needed to hold the information contained in these two sentences in your working memory long enough to be able established a connection between the sentences and to your own world knowledge. This allows you to infer, even though it is not explicitly stated, that Mary took the cap off the bottle and that in the bottle was water which Mary drank.
In an experimental study, Hamilton et al. (2016) investigated the effect of the working memory capacity of proficient adult readers on their ability to make inferences. The participants read four-sentence paragraphs. After reading the passage, participants were shown a ‘target’ word and had to quickly decide whether the word was a real word or a non-word. In some passages, the target word was related to an inference that related to the last sentence in the passage. In the above example, it could have been the word ‘drink’. In other passages, the target word was unrelated to the passage (e.g., throw). Participants were then asked a comprehension question about the passage.
To assess participants decoding skills, some sentences contained pseudo words which when pronounced sounded like real words which would make sense in the context of the sentence. For example: Greg was amaized (amazed) that his hoarshoo (horseshoe) landidd (landed) ryte (right) in frunt (front) of the mettul (metal) poste (post).
Participants also completed separate tasks designed to assess their working memory capacity and their decoding ability. On the basis of these two tasks, participants were categorised as being fast, medium or slow decoders.
The researchers found that all participants responded faster to the ‘target’ words when the target word was related to the last sentence in the passage (i.e., inference priming) compared to when the target word was unrelated to the passage and this was true in both the real word and non-word passages. However, response time was significantly faster in the real word passages suggesting less processing was required in this condition.
Importantly, Hamilton found that slow decoders who had poor working memory capacity were significantly slower to respond to the non-word passages than the real word passages. However, for slow decoders with high working memory capacity and fast and medium decoders the response rate didn’t differ significantly between the real word and non-word passages.
These results suggest that when readers have to devote substantial mental processes to the decoding process, they have less capacity to engage in high-level comprehension processes such as inference making.
The take-away message is that students with poor decoding skills will have difficulty integrating concepts and ideas that are presented in a text that they find challenging and so consequently will have poor comprehension of the text. This difficulty will be diminished to some extent if the student has a reasonable working memory. Given that many students with reading difficulties also have poor working memories, it is not surprising that these students have to work so hard to become proficient readers.
How can we help these students?
- Include memory retention strategies into literacy programs to assist in storing information in long-term memory. This is particularly important in terms of developing automaticity for decoding.
- Provide sufficient scaffolding to limit the load placed on working memory. This might include identifying and pre-teaching words the student doesn’t know.
- Provide the student with the required background knowledge to understand the text.
Reference
Hamilton, S., Freed, E., & Long, D. (2016). Word-Decoding Skill Interacts With Working Memory Capacity to Influence Inference Generation During Reading. Reading Research Quarterly, 51(4), 391–402. doi: 10.1002/rrq.148