While basic reading patterns like Z- and F-patterns are common globally, the direction of the script and spelling complexity significantly alter how our eyes navigate text. Eye-tracking studies show how right-to-left script readers and East Asian language speakers process text. This demonstrates how culture and writing systems shape the reading experience.

Left-to-Right (LTR) Scripts

LTR reading direction relates to languages as diverse as English, Hindi, Tamil, Georgian, Armenian and Bengali. These cover radically different scripts.

LTR readers generally follow Z- and F-patterns for skimming and scanning and linear patterns for in-depth reading.

What can affect LTR scanning?

• Script density, such as in Tamil and Bengali, have more complex letterforms than English. This causes longer attention fixations and shorter eye jumps (saccades)

• Orthographic density, such as Hindi’s Devanagari script results in more eye regressions when processing conjunct consonants,

• Agglutination: Tamil uses long compound words and this subtly changes eye movements compared to alphabetic scripts.

Right-to-Left (RTL) Scripts

RTL readers, such as speakers of Arabic, Hebrew, Farsi and Urdu show patterns that mirror the Left-to-Right readers.

• The Z-pattern starts top-right, flowing diagonally to the bottom-right:

• The F-pattern also starts top-right and then skims down the right margin.

Key findings for RTL scripts:

• Readers experience more duration of fixation: Arabic, Urdu and Farsi readers show longer fixations due to the cursive nature of the script.

• Arabic letters connect within words (character connectivity), so readers rely on processing the whole word shape (gestalt processing) more than LTR readers.

  
  

Logographic Scripts

In languages like Japanese and other East Asian languages, eyes scan text top to bottom within column. Then, they move right to left to the next column. This is interesting because languages like modern Chinese and Japanese, horizontal, left-to-right patterns, especially in web content. However, readers still tend to exhibit vertical reading habits, hence the visual processing of text in columns, as in the following visualisation:

How does logographic density affect reading?

Chinese characters carry more meaning per symbol. This leads to shorter fixations and more frequent jumps. Meanwhile, Japanese readers may show mixed reading patterns, when switching between kanji (logographic), kana (syllabic) and Latin letters.

 Global Eye-Tracking Insights

• People spend more time looking at words in complex scripts. For example, readers take longer to process scripts like Tamil because the letters and words are more detailed.

• Eye jumps are shorter in logographic scripts. In Chinese, each character carries more meaning, so readers move their eyes in shorter steps across the text

• Reading direction: designers must mirror layouts for RTL scripts or respect column-based flows in vertical layouts.

Why This Matters

For UX designers, translators and global content creators, understanding these script-specific patterns can:

• Improve user engagement

• Enhance accessibility

• Boost comprehension.

This is especially critical in Clinical Outcome Assessments (COAs). These are key instruments that often contain complex instructions, scales, questions that patients need to navigate and understand clearly. The way patients interact with these documents is directly influenced by how their eyes scan and process text. As we have discovered, this is something that varies across scripts and cultures.

In LTR scripts, such as English and Hindi, placing important questions or instructions along the natural left-to-right reading path aligns well with typical eye-tracking patterns like the F-pattern. On the other hand, readers of RTL scripts, such as Hebrew, will expect the critical information to be mirrored: top-right started points and the right-hand margins for skimming.

If COAs are not adapted to fit the natural scanning habits of different reading directions, patients may:

• Miss key information, such as a rating scale explanation or the beginning of a shared-stem question

• Experience cognitive overload due to poor alignment with reading habits

• Take longer to complete assessments, increasing fatigue and risking lower data quality.

Moreover, for logographic script readers, such as Chinese, COAs formatted in dense horizontal blocks may hinder comprehension compared to layouts that respect columns that are aligned with their culturally bound visual scanning preferences.

Ignoring these script-driven eye-movement patterns could negatively impact how patients understand and respond to clinical assessments. Since COAs often inform clinical trial endpoints and regulatory decisions, small design inefficiencies could have wider consequences for both patient engagement and data reliability.

By designing COAs, as well as other patient-facing materials, with global reading behaviours in mind, developers can ensure that instructions, scales and questions are easier to process across diverse language communities. This will support better patient experience and more robust clinical data.

Thank you for reading.

Mark Gibson

Leeds, United Kingdom, March 2025