Where Do Children Look when Watching Videos with Same-Language Subtitles?

Can same-language subtitles help children learn to read? Maybe, but only if children look at them.

An innovative solution to a global problem

Learning to read is challenging for many children. Despite substantial resources invested in teaching literacy, approximately 57% of children in low- and middle-income countries and about 6% of children in the European Union fail to learn to read.

These statistics are alarming. If children can’t read, they can’t use reading to learn in secondary school. Poor reading is also associated with a host of negative outcomes in adulthood such as lower income and poorer health. 

This is a global problem that is difficult to solve. Surveys conducted by the National Literacy Trust show that only around one-third of children enjoy reading in their free time, and just one in five children report reading daily outside of school.  

One innovative solution has been proposed by the public campaigns Turn on the Subtitles in the UK and CaptionsON in the USA. These campaigns suggest that same-language subtitles should be switched on by default for children’s programming so that “television time” becomes “reading time”.

A weak evidence base

This proposal has been likened to sneaking vegetables into a child’s dinner: they don’t notice but it does them a “world of good”. Yet, the evidence base behind the proposal is weak. There is no evidence that children even look at same-language subtitles, much less read them, or whether they need to have already become fluent readers before they do so. 

There are a few published studies that report that watching videos with same-language subtitles improves children’s vocabulary and reading skills. However, these studies have serious methodological limitations: small sample sizes (and hence low power), lack of control groups, and reliance on retrospective self-reported viewing habits.

We therefore started with the basic question of whether children look at same-language subtitles and whether this depends on their reading skills.

The experiment

We asked 180 primary-school children in Years 1-6 to watch short movies with and without English subtitles. We recorded their eye movements using a high-precision camera that monitors where the eyes look on the screen at 1000 times per second. The presentation of subtitles followed the BBC Subtitle Guidelines. We also measured each child’s reading fluency using a standard test.

The eye-tracking set-up (EyeLink Portable Duo system).

Our results show for the first time that children do look at same-language subtitles. Our results also show that when children look at subtitles they are reading them. We know this because we observe some of the hallmarks of reading: namely, that the eyes spend more time looking at longer, less common words. We found no evidence that reading subtitles improves or impairs understanding of the movies. 

These findings may sound like they support the subtitles campaigns described earlier.  However, importantly, our results also revealed a strong relationship between engagement with subtitles and reading proficiency. Children at the end of Year 1 mostly ignored the subtitles. Even when they looked at the subtitles, they skipped most of the words in them. It wasn’t until children reached the end of Year 2 or Year 3 that they showed reasonable engagement with the subtitles. By this time, children were already able to read at about 1 word per second. 

Our findings are illustrated in the heatmaps below. These show the distribution of eye-fixations aggregated across all participants in Year 1 and Year 6 while they watched the whole video Taking Flight (5 minutes and 5 seconds).  The top row shows eye fixations when subtitles were present, while the bottom row shows eye fixations when there were no subtitles. 

It’s also possible to see the difference between pupils in Year 1 and Year 6 in the videos linked below. These videos show eye-fixations for these pupils as they watched a short 32-second segment of this same video with subtitles. It’s important to remember when watching these videos that the data are taken only for this short segment, and that frame-by-frame shots do not represent the overall pattern of results.

The key conclusions

Our study suggests that children need to be reading at around 1 word per second before they start paying attention to same-language subtitles. Even for children who have learned their phonics to a high standard, subtitles appear and disappear too quickly for beginner readers to follow. These results allow us to be fairly certain that same-language subtitles will not help children in the initial years of learning to read. 

It’s possible that same-language subtitles may provide valuable reading practice for older children who have already acquired sufficient reading fluency to engage with them. However, further research is needed to test this causal hypothesis. It’s also important to remember that the language used on television is less sophisticated than the language used in books. Thus, even if subtitles do offer a degree of additional reading practice, they cannot replace books in terms of their benefits to the developing reader. 

Some final words 

It’s surprising how quickly the ‘Turn on the Subtitles’ campaign gained widespread, unquestioning support, given the limited evidence base. It’s seductive to think that the hard work of learning to read could be (partially) replaced by watching subtitled television. Yet, our study suggests the reverse: learning to read probably allows children to engage with subtitles. 

In the absence of further evidence, we would advise families to turn on the subtitles if they find them useful. However, we would argue based on the available evidence that teaching children to read requires us to focus on providing three essential building blocks: strong oral language experiences, high-quality reading instruction, and supervised book reading at home.

For more information about the experiment, please see the paper:

Lopukhina, A., van Heuven, W. J. B., Crowley, R., & Rastle, K. (2025). Where Do Children Look When Watching Videos With Same-Language Subtitles? Psychological Science. https://doi.org/10.1177/09567976251325789

We are grateful to the Nuffield Foundation (FR-23450) for supporting this work and to the advisory board of the project for their feedback and support.