Reading has long been an important theme, especially given the negative trends in PIRLS and PISA. In my talks about this topic, I often mention the Simple View of Reading. This model posits that reading comprehension is essentially the result of the multiplication of two components: the ability to decode on the one hand, and (language) comprehension on the other. But how does this model relate to other cognitive skills, such as working memory, processing speed, or general intelligence?

A 2025 study by Hajovsky and colleagues sheds more light on this. To do so, the researchers took a rather large-scale approach. They collected 49,959 correlations from 599 correlation matrices derived from 137 different test batteries. This means the data comprise nearly 200,000 individual observations. Do note that this does not necessarily mean an equal number of unique participants. To handle this large dataset, the researchers used a meta-structural equation modelling approach. They subsequently attempted to develop an overarching model of the relationships between cognitive skills and reading.

What did they find?

First, the researchers distinguished various aspects of reading. For instance, they examined phonological decoding, word recognition, word decoding speed, fluent reading, and reading comprehension.  They also included eight broad cognitive skills, such as working memory, auditory processing, processing speed, reasoning, and prior or acquired knowledge.

One important finding is that these reading skills are strongly interrelated, which shouldn’t surprise anyone. For example, phonological decoding appears to be an important predictor of word recognition (β = 0.55). Word recognition, in turn, is related to decoding speed and also has a significant direct relationship with reading comprehension (β = 0.31).

Also not really surprising is the role of what the researchers call comprehension-knowledge. They describe it as a broad cognitive skill that includes vocabulary, general knowledge, and acquired language skills, among others. What do the findings show? Of all the cognitive skills examined, this one exhibits the strongest correlation with reading comprehension (r = 0.56). Furthermore, auditory processing (r = 0.46) and fluid reasoning, or the ability to solve novel problems (r = 0.43), also seem associated with reading comprehension.

It became slightly more surprising when the researchers combined these skills into a single model. Many of these relationships turn out to be indirect or overlap with other skills. For instance, auditory processing is particularly important for phonological decoding. And for the speed at which words are decoded, general processing speed plays a greater role. For reading comprehension, acquired knowledge, including vocabulary and language-related knowledge, remained particularly relevant. In this way, the model ultimately explains 54 per cent of the differences in reading comprehension. For word recognition, that figure is even 59 per cent.

The Simple View of Reading holds up remarkably well.

What does this tell us about the Simple View of Reading? As noted, according to that model, reading comprehension develops by combining two things: decoding and language comprehension. To understand texts, you should not only be able to read but also understand what you read. The findings seem remarkably consistent with Gough and Tunmer’s idea, later adapted by Hoover and Gough. First, several types of decoding processes are associated with reading comprehension. Second, other factors, such as acquired knowledge, also play a role.

Still, one needs to mention one very important thing about the way the research is conducted. The researchers did not measure language comprehension separately, which is part of the original Simple View. Moreover, while acquired knowledge includes some language factors, the researchers’ concept is broader than this. Thirdly, this Simple View implies testing the multiplication of two variables. Here, that is not being done.

This means the study does not prove the Simple View, but it works especially well within it. Even when you build a complex model with eight cognitive abilities and different elements of the reading process, decoding and language-related knowledge remain central.

But there are also limitations

The study’s scale is impressive, but that does not mean we can simply generalise the results. I must confess that I already had doubts when I saw the journal’s publisher. After all, it does not have such a good reputation.

Furthermore, the heterogeneity between the investigated correlations is very high. The average I² is as high as 87.8%. In concrete terms, this means the strength of the observed associations differs considerably across data sources. The researchers did not systematically investigate the extent to which, for example, age can explain these differences.

Furthermore, the tests are in English. This is not unimportant, as English spelling is considerably less transparent than Dutch. Therefore, it is not self-evident that all correlations would be equally strong in Dutch.

And then there is the classic: yes, this is correlation research. The model does contain arrows between different skills, but that does not yet demonstrate causal relationships. For instance, auditory processing and phonological decoding are highly correlated, but that does not automatically imply that training auditory processing will improve reading.

This also applies to working memory or processing speed; a cognitive process may be essential for reading, but training it doesn’t necessarily improve reading.

What can we take away from this?

For me, this research primarily confirms that we cannot view technical reading and reading comprehension separately, but we also shouldn’t confuse them.

Anyone who wants to learn to read well must be able to decode words fluently and accurately. But anyone who wants to understand texts also needs language comprehension, vocabulary, and knowledge. Although other cognitive skills play a role, there seems little reason to put forward any one of them as the sole key to good reading.

The study certainly offers no miracle solution for teaching reading. Nevertheless, it offers an intriguing confirmation of an idea that has long been at the heart of reading studies: reading comprehension requires technical reading ability, plus sufficient language and knowledge.