Wide-spread cheating inflates students’ achievements in India, but technology-aided solutions can improve data accuracy in standardised testing
Read “Myths of official measurement: Auditing and improving administrative data in developing countries” by Abhijeet Singh here.
Misreported or manipulated administrative data is a key impediment to public service delivery in low- and middle-income countries. In India, cheating and misreporting on standardised education assessments which are relied upon to support the development of education policies, likewise threaten policy interventions designed to improve students’ low learning levels. In this VoxDevTalk, Abhijeet Singh discusses his recent work in a two-part study in India examining the extent of grade manipulation in standardised tests as well as the effectiveness of technological aids to reduce distortions in student testing.
His results show that in the state of Madhya Pradesh, maths and Hindi scores on the standardised tests are significantly higher than students’ actual learning levels. In a separate intervention in the state of Andhra Pradesh, he finds that utilising tablets to conduct tests that were independently graded greatly reduced incidences of cheating and produced test results that were more aligned to students’ actual learning levels. These results show that administrative data can suffer from large inaccuracies, with large implications as policymakers attempt to be more data driven in their approach to improving learning, and building state capacity in general, and that solutions to strengthen data accuracy can greatly improve data quality.