Newsela’ Approach to Leveling Nonfiction Texts

Similarities

• The core meaning, facts, and topical standards alignment 

• Voice and engaging details that make the text worth reading

• Domain‑specific terms essential to the topic – with supportive context or definitions, as needed

• Sensitivity to developmental maturity and representation

• Accuracy and journalistic integrity

Differences

• Our editorial team tailors complexity across multiple dimensions that research shows support comprehension and fluency. We combine human editorial judgment and research-based tools, including the Lexile framework by MetaMetrics, to make decisions.

Beyond Lexile level estimates, we consider the following quantitative and qualitative dimensions of text complexity and developmental appropriateness beyond Lexile level, from sentence length to background knowledge demands to vocabulary.

Text and Language Features

• Sentence Structure: Shorter, more direct sentences appear in less complex versions.

• Sentence Length: More varied, multi‑clause sentences appear in more complex versions.

• Cohesion and Transitions: Clear connectors (e.g., first, next, as a result) help readers follow ideas and chronology, especially in earlier versions.

• Organization: When a source text is complex, earlier versions may use a more linear, chronological structure. Later versions preserve or reintroduce more sophisticated structures.

• Vocabulary Load: We reduce the density of low‑frequency words in earlier versions. Topic‑critical terms are retained with either context or a full definition.

• Knowledge Supports: We add brief, plain‑language explanations of concepts and institutions when the topic assumes background knowledge many readers may not yet have.

• Opportunities for Inference: Lower complexity versions make relationships explicit. Higher complexity versions rely more on context clues and allow readers to do more inferencing.

• Variety and Clarity: We avoid repetitive sentence openings and imprecise wording, and we maintain a proper tone at every complexity level for easier reading.

Appropriateness and Engagement

• Maturity Bands: Text complexity is separate from topic maturity. A high‑school topic can be offered at a less complex reading version for older students who need it.

  

  In Newsela, teachers can search by reading level, maturity level, or both. 

  High school teachers

  • Teachers with striving readers may want to find content written at a simple language that will be respectful of their age and level of knowledge.

  • The maturity filter can help find high school-appropriate content at a low text level. 

  Elementary teachers

  • Teachers with advanced readers may want to respectfully meet students not only at their reading ability, but also at their age and stage.

  • The maturity filter can find some of the most complex text that would still be developmentally appropriate for their readers. 

• Interest and Tension: We preserve the original article’s purpose and narrative drive while balancing added context. 

  • For news type content, we adhere (as best as possible) to the inverted pyramid structure.

  • For other types of content, we use editorial judgement.  

We provide grade and Lexile level estimates for each text. The Lexile ranges listed are what Newsela staff use as a guide to map each level to a specific text complexity:

• Grade 2: 300-500

• Grade 3: 501-620

• Grade 4: 621-780

• Grade 5: 781-920

• Grade 6: 921-1010

• Grade 7: 1011-1109

• Grade 8: 1110-1180

• Grade 9-10: 1181-1260

• Grade 11-12: Above 1260

Unlike other Lexile mappings you may have seen, our grade-level bands are designed to be non-overlapping. Each article level corresponds to only one grade level to avoid confusion within our platform. You may also find that the Lexile numbers assigned to Newsela articles may be slightly different from the numbers you see on other platforms. 

This is a deliberate choice. When we assign a Newsela grade level, we take into account not just the Lexile score, but also other factors affecting text complexity, as described above. This approach ensures that we provide students with an appropriate challenge at each grade level.

Each version of an article we produce includes a Lexile measure to describe its relative text complexity. We also align with other text complexity metrics like Flesch-Kincaid.  

Each quiz is specifically written to correspond with the phrasing and complexity of the article's reading level. For instance, if an article is written at a fifth grade reading level, our quiz questions are written to reflect standards as they will be assessed on end-of-year fifth grade state assessments. 

Our quizzes correlate with performance on state assessments. Read the research.

What constitutes “appropriate complexity” varies by learning goal. Shanahan (2020) recommends teachers provide their students with texts across a range of difficulty levels. Our approach is aligned with this research and recommends providing students with texts across a range of difficulty levels: 

• Complex texts on or above the student’s grade level: Enable students to strengthen comprehension strategies and develop academic language and syntax

• Independent reading texts at the student’s achievement level: Grow knowledge and vocabulary by independently reading large volumes of text at their current reading level

• Simpler texts if needed for striving readers: To help students build knowledge and fluency by reducing word-reading demands
  
  Research indicates that comprehension is stronger when text complexity is well matched to the task and the reader, and that knowledge‑rich reading across topics boosts later comprehension.

For over a decade, a team of former journalists and educators have carefully selected, curated and leveled our nonfiction texts. They have all been extensively trained and follow clear editorial guidelines to gradually reduce the complexity of text, adding embedded supports to each version, such as increased context and simpler structure and vocabulary.

We regularly review our editorial guidelines and engage in internal training to make sure our staff are equipped with the latest knowledge about the science of reading. Our grade-level estimates today are highly correlated with other measures of reading level, such as Flesch-Kincaid Grade Level (r > .8).

For guidance on choosing a text version, go here.

Note: This is an explanation of how we level the nonfiction texts in the Newsela library. Separately, users who copy non-Newsela texts into the AI-powered text leveler will get grade-level estimates for each AI-leveled text. For more on that feature, go here. 

Learn more about Newsela's article leveling and how it is used

Level Control

Student Article Level

AI-Powered Text Leveler