Caste, Purity, and Educational Outcomes: Evidence from a Nationally Representative Survey

• In Event: Access, participation, and inclusion in South Asia

Proposal

Introduction and Literature Review

A recent verdict of the Supreme Court of India permits sub-classification within Scheduled Castes (SC) and Scheduled Tribes (ST) for reserving seats in educational institutions and employment in government (Teltumde, 2024). However, existing scholarship shows that despite 70 years of reservations and other efforts, the literacy rates of SC and ST children (Raghavendra, 2020) are lower compared to the national average. Further, Borooah (2012) finds the learning outcomes of these children are lower compared to upper-caste children and argues that after controlling for parental, household, and school-related factors, the upper-caste children held a ‘structural advantage’ in some or all of the three competencies tested.

Other studies on the topic suggest a lack of human and physical capital in SC and ST households (Deshpande, 2011), school-level discrimination faced by SC children at the hands of their peers, and upper-caste teachers in social relations, and pedagogic processes and grading contributes to the gap in learning outcomes (Hanna & Linden, 2012; Nambissan, 2010). At the same time, village-level factors such as the low provision of public schools in SC and ST-dominant villages (Bailwal & Paul, 2021) and teachers’ differential treatment towards SC children in low and high-caste-dominated villages are possible factors that contribute to this gap (Bailwal & Paul, 2024).

Previous studies, however, fall short of providing direct evidence of the impact of village-level caste discrimination on the learning gaps between SC and ST, and upper-cast children. This discrimination is rooted in notions of ‘purity and pollution’ and manifests explicitly through the practice of untouchability (Dumont, 1966). Although untouchability has been banned in India since 1955, many untouchables still perform jobs that are regarded as “polluting.” Moreover, Thorat and Joshi (2015) find that 30% of households in rural India accept practicing untouchability. Hence, many Dalit scholars (Deb & Minj, 2024) have argued that the verdict permitting sub-classification within SC and ST is not based on sound empirical data and that the order could dilute the reservation provisions introduced initially to overcome the injustice of caste-based segregation that persists.

To address the gap in the literature, this paper examines if there is a learning gap in reading and math between SC, ST, and upper-caste children. Next, it asks if the practice of untouchability increases the performance gap between children belonging to SC and ST and upper-caste communities.

Data and Methods

This study utilizes the Indian Human Development Survey (IHDS-2012), a nationally representative multi-topic panel survey of households conducted in 1,503 villages, 971 urban neighborhoods, and approximately 40,000 households across India. While the researchers ran the survey in 2005 and 2012, only in the latter round were primary household respondents specifically asked whether their household practiced untouchability. Hence, we use only the 2012 round of the study for our analysis. Further, the study focuses on the rural areas subsample since a higher percentage of respondents (30%) than the urban respondents (20%) reported practicing untouchability, which could be an indicator of the true prevalence of untouchability in rural areas or suggests a more conservative mindset of the rural population as respondents felt more comfortable with sharing the information (Thorat & Joshi, 2015).

In the individual surveys, every respondent was asked whether they were currently in school and the years of education they had completed. Children aged 8-11 were given short reading, writing, and math tests. For the present analysis, we excluded individuals who were not given the tests. Children whose test scores were not available were also dropped. Finally, entries that did not have values for variables selected were dropped to make the models comparable and a total of 5,973 entries were analyzed.

The main dependents are reading and math scores, which are evaluated in separate models, while the independent variable is the children's caste/group. Our moderator variables are taken from the Yes or No responses to questions in the survey regarding the household practice of untouchability and were coded as 1 or 0 respectively. Since we are interested in the effect of village-level caste discrimination, we construct a ‘village-level practice untouchability’ variable defined as the fraction of households in a particular household’s village other than the household itself that practices untouchability, which is used as the moderator.

In this study, we first use state-fixed effects with standard errors clustered at the village level to examine how much of the learning gap can be explained by the ‘caste/group’ and ‘village-level practice of untouchability with the appropriate controls introduced in a step-wise manner. We control for student-level factors such as the child’s age, gender, whether they attend a private school or get private tuition, and household-level factors such as the practice of untouchability, log annual income, assets, and whether a primary-school-educated adult resides in the household. In the next specification, we interact with the ‘village-level practice of untouchability’ and the ‘caste/group’ of the child to examine the effect of untouchability on the learning gap.

Findings

Results reveal three important findings. First, the learning outcome of children belonging to SC and ST communities is lower than upper-caste children in both reading and math. Second, the ‘village-level practice of untouchability’ has led to higher reading and math performance of upper-caste children and lower reading and math performance of SC and ST children. Finally, it increased the learning gap between SC, ST, and upper-caste children.

This study uses cross-sectional data and findings are correlational. Despite the limitations, the study has important implications. Considering that caste discrimination continues to affect the educational outcomes of SC and ST students in addition to the other economic and educational disadvantages, doing away with the reservation for the economically better-off among SC and ST students will reduce the effectiveness of the reservation policy in achieving social justice. Further, even if sub-categorization among SC and ST communities is carried out, there should be some form of reservation for the castes and tribes that sit at the top of the pyramid as they are not immune from the persistent negative effects of caste discrimination.

Author

• Prayas Sutar, The University of Wisconsin-Madison