IIT Bombay app to transform evaluation of reading fluency among school students

Numerous reports show that the reading and literacy skills of a majority of school students in the country are not commensurate with the level of education

Mumbai: The Indian Institute of Technology (IIT) Bombay has developed a mobile application called TARA (Teacher’s Assistant for Reading Assessment) which is set to change the way reading skills of school students are assessed. The app – which uses advanced speech processing and machine learning to evaluate reading fluency metrics like accuracy, speed and expression – has been deployed in around 1,200 Kendriya Vidyalaya Sanghatan (KVS) schools across the country, to assess the English and Hindi reading skills of nearly 700,000 students from classes 3 to 8.

Researchers at IIT Bombay, led by Prof Preeti Rao from the department of electrical engineering, developed TARA (Arne Hückelheim / Wikimedia Commons)

Numerous studies and reports show that the reading and literacy skills of a majority of school students in the country are not commensurate with the level of education. The Annual Status of Education Report (ASER) 2022 revealed that more than 50% class 5 students in India could not read class 2-level texts. Similarly, the National Education Policy (NEP) 2020 estimated that 50 million students in India lacked foundational literacy and numeracy skills.

The Covid-19 pandemic exacerbated the gap, with 90% students losing at least one specific language skill such as reading comprehension or describing pictures, said a statement issued by IIT Bombay. While governments are focussed on improving literacy, monitoring reading skills and other learning outcomes is both labor- and time-intensive, requiring trained evaluators. TARA hopes to bridge this gap, the statement noted.

“TARA measures not only the number of words students read correctly per minute, but also the phrasing, intonation and stress in speech, providing a comprehensive picture of a student’s reading abilities,” said Prof Rao. Its reliability matches that of human evaluators, she said.

Dr Shailaja Menon, a reading pedagogy expert, highlighted the app’s potential, saying, “Organisations have long sought a digital tool for real-time data on learning levels. TARA fills this gap by providing performance insights for individual students and larger cohorts.”

The deployment of TARA in KVS schools started in October this year, and it has been rolled out in nearly 1,200 schools across the country in less than two months. The app development team is also involved in designing effective remedial instruction for students struggling with reading in collaboration with KVS authorities, said Rao.

“These interventions will be assessed in subsequent evaluation phases, paving the way for a cycle of testing and improvement throughout the academic year,” she said.

IIT Bombay is seeking partnerships to scale TARA’s impact and improve literacy across India’s educational system. But concerns are rife about whether it would be able to assess reading skills of students from different linguistic groups whose pronunciations may vary from the samples the app has been trained on.

Responding to the concerns, Rao said, “In TARA, phrasing (grouping of words), intonation and stress in speech are also measured to obtain a holistic score which is indicative of the precise stage of reading development.”