Playbook
Why Arabic-Speaking Students May Lean on Text-to-Speech More Than Their Spanish-Speaking Peers
Key Ideas
- On the 2017 NAEP Grade 8 digital math assessment, Arabic-speaking English Learners activated Text-to-Speech (TTS) roughly 2.5 times as frequently as Spanish-speaking peers.
- Arabic-speaking students used TTS consistently across all difficulty levels, relying on it as a routine reading scaffold rather than a last resort.
- TTS usage was associated with spending more time per item, but not with higher or lower math scores for either group. Its value may lie in providing equitable text access rather than boosting test performance.
Imagine you’re an eighth grader taking a digital math test. The problem in front of you is dense with words, and English isn’t your first language. You have a text-to-speech (TTS) button that can read the item aloud. Do you use it? And does the answer depend on which language you grew up speaking at home?
In education research, we tend to lump all “English learners” (ELs) together into a single category. But EL is an umbrella term covering students who speak dozens of home languages, with very different relationships to English. A Spanish-speaking student can lean on a shared alphabet and hundreds of cognates. An Arabic-speaking student is working across a completely different script, written in the other direction, with almost no vocabulary overlap. In my recent study, I ask a simple question that’s rarely been asked: do ELs from different language backgrounds actually use accessibility tools differently?
A Striking Usage Gap
Arabic-speaking students activated TTS roughly 2.5 times as often as Spanish-speaking students, even after I accounted for their English proficiency, accommodations, gender, and economic background. Without extended time, 40% of Arabic-speaking students used TTS at least once, compared to 28% of Spanish-speaking students. Arabic speakers activated it far more often on average: 24 times across the 15-item test, versus nine times for Spanish-speaking students.
The gap showed up most clearly on the items with the most words. As word count went up, Arabic-speaking students activated TTS more often, and that increase was steeper than the increase seen among Spanish-speaking students. And while Spanish-speaking students tended to save TTS for the hardest questions, Arabic-speaking students used it steadily across easy and hard items alike, more like a routine reading aid than a last resort.
Why Would This Be?
This lines up with what we know about how far apart Arabic and English are as languages. Arabic uses a different script, often written without vowels, and builds words through patterns quite unlike English. Spanish shares our alphabet and a deep well of cognates. The further a student’s home language sits from English structurally, the more that student may need support accessing the words on the page, regardless of their proficiency rating.
I make these statements with caution: my Arabic-speaking sample was small, about 50 students, so this finding is suggestive rather than conclusive. But the pattern is what a linguistic-distance explanation would predict.
The Twist: More Usage, Same Scores
Using TTS more wasn’t linked to higher or lower scores, and that held equally for both groups. Once I accounted for English proficiency, students who activated TTS more often didn’t outperform or underperform students who used it less. And that null relationship didn’t differ by home language either, so TTS wasn’t quietly helping one group more than the other. What the tool use did move was time—each activation added a few seconds per item, but Arabic-speaking students also spent about 14 seconds longer per item even after accounting for TTS use, suggesting the extra processing time isn’t just attributed to the tool.
TTS doesn’t look like a shortcut, or like a crutch that holds students back. It’s a tool students can use to process English text at their own pace, helping them access the item, but not necessarily helping them get it right by itself.
What This Means for Practice
Treating ELs as one group hides significant differences in how students actually engage with the tools we build for them. If Arabic-speaking students are relying on TTS as a default reading support, that’s worth knowing when we design assessments, train teachers, and decide which accommodations to offer whom.
As more assessments go digital, we’re gaining process data that can tell us not just how ELs perform, but how they get there. Our job now is to use it—to disaggregate, to ask which students are using what tools and why, and to build accessibility supports that actually match the diversity of the students we’re trying to serve.
- Topics: Evidence-Based EdTech
- Learning Environments: K-12
- Audiences: Educators , School and System Leaders
- Offerings: EdTech Strategy and Standards , Research and Development
- Programs: Learning Sciences Research