Reading assessments enable schools to identify students who need additional support, track progress over time, and inform classroom practice. The need is clear: one in five children enters secondary school with reading or vocabulary ability two years behind expected levels, yet only half of those students are identified as needing additional support (caer.org.uk).

So, we know that accurate reading assessment matters - but what exactly should we be assessing, how often, and how can we translate our findings into more effective classroom practice?

Reading for meaning involves building a dynamic mental representation of a text that updates with every word read (Van Dijk, Kintsch & Van Dijk, 1983). This situation model is constructed from two interacting components:

• Text base: The explicit information on the page, held in working memory
• Background knowledge (schemata): What a student already knows, including vocabulary and prior experiences

These components do not operate separately. Background knowledge influences how explicit meaning is constructed, including selecting the correct meaning of ambiguous words (for example, “operation”) and resolving references such as pronouns (for example, identifying what “it” refers to in a sentence).

When comprehension fails

Comprehension can fail when:

• Decoding and word recognition are not accurate or not automatic, consuming working memory and limiting capacity for meaning-making
• Background knowledge and vocabulary are insufficient or not activated, weakening inferential processing

Research has examined what this looks like within a classroom setting (Smith, 2021). Low-skilled readers (those in the lowest quartile of the population) typically construct a less detailed situation model because the text base is less coherent, schemata are less well-formed, or both (Kintsch, 1998; McNamara, Ozuru & Floyd, 2011).

Without an effective text base (that is coherent with the content of the text), the reader has limited access to information that can be integrated with any related schemata (Kintsch, 1998). A less coherent text base results in a poorer understanding of the text, and could mean the reader is unable to recognise differences between characters or recall information from the text.

In contrast, problems associated with schemata tend to manifest as difficulties drawing inferences from the text.

"It is important to note that while decoding is a constrained task that typically requires instruction, linguistic comprehension is an unconstrained skill that continues to develop over the lifespan and is highly dependent on general knowledge.”

- Modelling the development of reading comprehension (Lervag & Melby-Lervag in ‘The Science of Reading’, 2022)

1) Universal screening

A universal screener is a standardised test given to a class, year group, or whole school to identify potential reading difficulties and students in need of additional support. Standardisation matters because reading comprehension is nuanced and requires consistent, comparable data.

2) Diagnostic assessment

Diagnostic assessments are given to select students who fall below a predetermined threshold on the universal screener. A common example is using a phonics screener to identify specific gaps in phonemic awareness or particular grapheme–phoneme correspondences. Diagnostic tools can also be used to measure the impact of targeted interventions over time.

3) Summative assessment

Summative assessments, conducted at the end of a term or school year, are used to evaluate progress and identify gaps that have emerged since the baseline measurement. As with universal screening, this can be taken by a class, year or whole school.

The importance of regular testing

Reading development is not linear. Difficulties often emerge over time, making one-off baseline testing (such as a single test taken at entry) insufficient (Stanovich, 1986). Regular assessment supports:

• Accurate diagnosis of specific issues rather than generic labels such as “struggling with reading”
• Identification of students whose needs emerge later
• Evaluation of progress and the impact of interventions

Common assessment cycles

At a minimum, reading ability should be assessed annually. The most common assessment intervals are:

• Once per year: Baseline testing at the start of the year, with additional tests used to measure the impact of specific interventions
• Twice per year: Baseline testing plus an end-of-year assessment to show progress and highlight emerging gaps
• Three times per year: September, January, and June/July testing to track progress and monitor interventions, sharing data with all teaching staff

“Reading comprehension ability, like word reading ability, exists along a continuum: there is no distinct threshold or break point at which we can separate good from poor comprehenders or good from poor word readers.”

- Children’s Reading Comprehension Difficulties, Kate Cain in ‘The Science of Reading’ (2022)

Reading comprehension exists on a continuum, so there is no natural cut-off separating ‘good’ from ‘poor’ comprehension (Cain, 2022). As a result, in practice, schools often use workable thresholds:

• Historically, in Ofsted reading deep dives - since the introduction of the EIF in 2019 - inspectors asked primary colleagues how to identify their lowest 20% of readers and what provision they have in place
• If you use a standardised test, a Standardised Age Score (SAS) below 85 is, by definition, “significantly below average” and schools often use this as a marker of where intervention is required
• Some settings consider a SAS of around 80 as the threshold for more intensive support, alongside softer approaches for students just above that range

This approach supports a tiered model: more substantial interventions for those most in need, and lighter-touch support for students who are below where they need to be but not to a critical extent.

Reading data often gets used to inform seating plans, intervention allocation, progress monitoring, and so on. The more valuable use is translating results into teaching and learning methods, especially when data indicates students are struggling with comprehension.

“When a teacher identifies that a pupil is struggling with aspects of literacy, the next step should be to accurately diagnose the specific issue(s) and then carefully plan how to support the pupil. Prompt identification of pupils’ specific literacy needs and the provision of appropriate support are critical to ensuring sustained progress.”

- EEF, Improving KS2 literacy

1) Pre-teaching vocabulary improves comprehension

Vocabulary instruction has moderate positive effects on reading comprehension when it focuses on words in the passage. Effects are greater for students with reading difficulties (Elleman et al., 2009). This supports pre-teaching key vocabulary and background knowledge before students encounter complex texts.

2) Depth and fluency matter more than breadth

Vocabulary depth and fluency* are more strongly related to comprehension processes, such as inference-making, than vocabulary breadth (Cain & Oakhill, 2014). As such, teaching fewer words in greater depth aligns is more likely to improve comprehension than attempting to cover large word lists.

*Fluency here includes automaticity in word recognition, reducing cognitive load and freeing working memory for meaning-making.

3) Knowledge must be activated

Background knowledge alone does not predict performance – it is the activation of that knowledge that supports inferential processing (Yeari et al., 2017). Therefore, classroom routines that activate prior knowledge before reading reduce comprehension barriers.

4) Informational texts require additional structure

Different text genres have specific structures; for example, expository (informational) texts are structured very differently to narrative texts. Knowledge of these structural nuances is important for reading comprehension – yet poor comprehenders may lack such knowledge. Indeed, young readers find expository texts harder to process and comprehend compared to narrative texts (Currie et al., 2021).

By teaching students the ‘cue words’ for various structural elements, we can enable students to recognise them. For example, the five most common structures found in expository texts are:

• Cause and effect
• Definition-example
• Problem-solution
• Proposition support
• Sequential listing

Teaching cue words such as “because”, “consequently”, “for example” and “finally” help signpost the above and therefore reduce processing demands.

5) Combining content knowledge and reading instruction yields stronger outcomes

Merging content knowledge and reading instruction builds the background knowledge that strengthens inferential processing, improving subject outcomes and reading development overall. In a meta-analysis, students who were taught this way were found to retain more vocabulary and understand content better than children who learned science or social studies separately from reading instruction. These students also did better on standardised tests of reading comprehension.

This integrated approach not only made them better readers of the content they were learning, it made them better readers overall (Hwang, Cabell & Joyner, 2022).

A good reading test allows you to confidently:

• Identify students needing additional support
• Monitor the progress and impact of interventions
• Provide all teachers with reliable, actionable data on reading ability
• Understand exactly which subskills readers are struggling with
• Benchmark your students to better understand their progress
• Save time on admin so you can spend more time teaching
• Save money on testing so you can invest in improvement

A standardised reading test supports benchmarking and progress tracking over time. During the development of Bedrock’s Reading Test, items were standardised on a sample size of 11,283 students. This sample was nationally representative for both gender and FSM percentages.

The reliability of the test is measured on a scale of 0 to 1. Values above 0.80 are considered good. Bedrock’s Reading Test is measured as having a reliability of 0.94 which is high.

It is critical that the data you glean from your reading assessment is actionable, thus translating into reading progress for every student of every ability. Output data to look out for includes:

• Reading age
• Standardised Age Score (SAS)
• National Percentile Ranking (NPR)
• Group Ranking
• Subskill analysis

The utility of SAS and NPR

SAS and NPR are a helpful way of understanding an individual student’s reading comprehension ability relative to other students of the same age.

Percentile ranks and standardised scores that stay the same show us that the student is making expected progress. If they go down, then progress is slower than expected; if they go up, more progress has been made than expected.

The importance of assessing subskills

Bedrock's Learning and Assessment framework was developed in collaboration with Professors Jessie Rickets and Kate Nation, two of the country’s leading experts on the science of reading.

Our Reading Test breaks down the complexity of reading comprehension into seven subskills. Subskill analysis supports interpretation at cohort and individual level and helps identify implications across subjects, including areas such as inference demands in humanities and science. The subskills analysed are as follows:

• Retrieval (explicit): The ability to understand information explicitly presented in the text
• Retrieval (implicit): The ability to understand information not explicitly presented in the text via inference
• Meaning: The ability to understand the specific meanings of words and phrases in context
• Evidencing ideas: The ability to understand how authorial intent is communicated by language choice
• Summary: The ability to understand the overall meaning of a paragraph or text
• Predictions: The ability to use understanding of what has happened in a text to predict what will happen next
• Inter- and intra-textuality: The ability to use understanding of overall text structure and argumentation to make comparisons within or across texts (this is a core reading skill underpinning the integrated reading-into-writing cognitive process)

Effective reading assessment requires:

• A clear model of what reading for meaning involves (text base plus background knowledge)
• Regular assessment cycles rather than one-off baselines
• Standardised screening supported by diagnostic follow-up where needed
• Practical thresholds that enable tiered support
• Classroom applications of data focused on vocabulary depth, knowledge activation, and explicit teaching of informational text structures

This approach not only improves the identification of students needing support, but critically, increases the likelihood that reading data changes classroom practice for the better rather than remaining a standalone data point.