A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in...

57
A RIGHT WORTH FIGHTING FOR? UNDERSTANDING DECLINES IN STUDENT ACHIEVEMENT FOLLOWING INDIA’S RIGHT TO EDUCATION ACT Erika V. Keaveney Master of Arts Paper International Comparative Education Graduate School of Education Stanford University July 2014

Transcript of A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in...

Page 1: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

A RIGHT WORTH FIGHTING FOR? UNDERSTANDING

DECLINES IN STUDENT ACHIEVEMENT FOLLOWING

INDIA’S RIGHT TO EDUCATION ACT

Erika V. Keaveney

Master of Arts Paper

International Comparative Education

Graduate School of Education

Stanford University

July 2014

Page 2: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

Graduate School of Education

Stanford University

INTERNATIONAL COMPARATIVE EDUCATION

A Right Worth Fighting For? Understanding Declines in Student

Achievement Following India’s Right to Education Act

Erika V. Keaveney

July 2014

A Master of Arts Paper in partial fulfillment

of the requirements for the degree of Master of Arts

Approvals:

ICE/IEAPA Master’s Program Director: ____________________________________

Christine Min Wotipka, Ph.D., date

Advisor: ____________________________________

Martin Carnoy, Ph.D., date

Page 3: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

Acknowledgements

I owe a great deal of gratitude to the Stanford Graduate School of Education ICE faculty,

whose support made this project possible. In particular, I wish to thank my advisor, Dr. Martin

Carnoy, for his guidance, mentorship, and inspiration over the past year. Many of the ideas and

concepts explored in this paper were first introduced to me by him, and his research and

philosophies have indelibly shaped my views on education policy. Special thanks are also due to

my relentlessly thoughtful Program Director, Dr. Christine Min Wotipka, whose excellent

guidance transformed me from an aspiring researcher into an actual one. Finally, I am much

indebted to Sen Zhou for her clarity, ingenuity, good humor, and patient hand-holding as I

moved through many rough drafts and incarnations of my research question(s).

I would also like to give special thanks to my ICE/IEAPA cohort with whom I shared

much joy and rumination as we moved through this transformative process together (not to

mention our shared despondency as our respective country teams were eliminated from the 2014

World Cup). I feel very fortunate to have acquired 13 new lifelong friends.

This paper was made possible by the generous tuition fellowship from the David and

Susan Douglass family (awarded by the Stanford Graduate School of Education). I would also

like to acknowledge the extra financial support from the ICE MA Fund and the GSE Travel

Grant, which allowed me purchase a licensed copy of Stata as well as present the proposal of my

thesis at the 2014 conference of the Comparative and International Education Society in Toronto.

This project would not have been possible without the efforts of Pratham, which allowed

me to use the unpublished raw data from its Annual Status of Education Report (ASER).

Finally, I owe a huge amount of gratitude to my husband Courtland Creekmore who has

supported me emotionally and spiritually through all of my professional and academic pursuits

Page 4: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

since the beginning of our relationship. In addition to exchanging our life in paradise for a tiny

Silicon Valley apartment and a brutal daily commute, he constantly lent me his eyes, ears, and

encouragement as I moved through this project—without ever appearing to feign interest or

understanding.

Page 5: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

Abstract

The rollout of India’s Right to Education Act of 2009 coincided with large declines in

math and reading proficiency in public schools. These declines have led scholars and

policymakers to question the Act and the broader project of building a common education system

in India, with some calling for alternatives to state schooling such as private school vouchers.

This paper uses multivariate logistic regression to explore extra-school causes of these learning

declines, and finds that in the years post-Act, student- and peer-level socioeconomic status have

a) declined and b) become stronger predictors of reading and math proficiency. These findings

suggest the need to consider shifting student demographics when interpreting trends in

achievement data, both within India and beyond. Further, because public school learning

declines were precipitated by the out-migration of high socioeconomic status students to private

schools, state-sponsored privatization schemes stand to exacerbate learning declines in public

schools. In sum, attributing achievement declines to deteriorating public school quality may be

misdirecting public educational policy.

Keywords

Input output analysis; Privatization; India; Peer effects; Achievement; Right to Education Act

Page 6: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

1

1. Introduction

The Right of Children to Free and Compulsory Education Act of 2009 (the “Act”) is the

most sweeping and ambitious piece of education legislation in the history of India. Guaranteeing

free and compulsory primary schooling to the nation’s 233 million children in the 6 to 14 age

group, the Act seeks to arrest the cycle of poverty for India’s indigent populations, while

strengthening their bargaining power in the political, economic, and household spheres. With

India housing one-fifth of the world’s primary school-aged children—as many as 20 million of

whom are out of school (UNICEF and the UNESCO Institute for Statistics, 2012)—such

legislation holds the potential to make or break progress toward the Education for All goals

embraced by the international community in Jomtien over 20 years ago.

The Act lays the practical groundwork for ensuring that the constitutionally-enshrined

right to education is delivered to all of India’s school children. In addition to establishing

protocols on school availability and hours, minimum infrastructure, and student-teacher ratios,

the Act offers provisions on funding and monitoring of its implementation. The central

government of India has devoted significant financial resources to implementing the Act, and has

established a center-state cost sharing ratio of 70:30 for most states, with a ratio of 90:10 for the

poorest states in the northeast of the country (Garg, 2013).

Since the Act went into effect in 2010, the nation has witnessed improvements in the

public school system, including increased accessibility, expanded school infrastructure, and

reduced class sizes (ASER Centre, 2013c). The three years following the Act also revealed

considerable declines in average student achievement in rural areas, where the majority of the

population resides. Though the recent nature of these events limits the extent of scholarship on

the topic, critics have contended that post-Act declines in student achievement are endogenous to

Page 7: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

2

the school system itself (Chavan, 2012; Kaul, 2013). This emphasis on school-level factors

ignores the effects of socioeconomic status (SES) and peer composition, which, taken together,

have been shown to eclipse the contribution of school-level inputs in the production of student

achievement (Coleman, 1966; Somers, McEwan, & Williams, 2004). Further, the expansion of

India’s education system coincided with the migration of millions of children from public to

private schools1 and rural to urban areas, which may likewise affect the average achievement of

students left behind through general equilibrium effects.

The purpose of this study is to explore the underlying reasons for declines in student

achievement in rural areas post-Act. Using household- and school-level survey data gathered by

Pratham’s Annual Status of Education Report (ASER) between 2009 and 2012, I attempt to shed

light on the following questions: (i) is there a relationship between the “decline” in student

achievement post-Act and the growing proportion of lower SES children in the public school

system? and (ii) will controlling for these demographic shifts change the relationship between

achievement and school inputs? In answering these questions, I hope to fill a critical gap in

research on the extent to which extra-school factors influence recent student achievement trends

in India. Such understanding can contribute to the public-private school debate and inform

analysis of the Act and the broader project of building a functional common education system for

the world’s second largest nation. Further, in light of the growing shift in development

paradigms from “Education for All” to “Learning for All” (World Bank Group, 2011), this study

can offer valuable insight into the short-term effects of mass education expansion on average

1 “Public” schools refer to government-run schools whereas “private” schools include both

government aided and unaided privately operated schools. As of 2005, only 3.4 percent of

primary and 10.1 percent of upper primary schools were classified as private aided, making the

distinction relatively insignificant for the purpose of this study (Kingdon, 2007).

Page 8: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

3

student achievement, informing criticisms of global education policies that prioritize equitable

access over quality.

The roadmap of this paper is as follows. In the background section, I briefly trace the

history and state rationale underlying the right to education movement in India followed by a

discussion of trends in student achievement and private school migration that have coincided

with the more recent Right to Education Act. I then review the extensive literature on variables

that effect student achievement—including socioeconomic status, peer effects, and school

inputs—and relate these variables to the context of Indian public schooling. I then present a

conceptual framework that highlights the relationship between student achievement trends and

the changing socioeconomic and peer composition across public and private schools. Using this

conceptual framework, I formulate and present my central hypotheses and empirical strategy.

Finally, I present key findings followed by a discussion of their implications for future education

policy in India and beyond.

2. Background and context

2.1 The right to education in India

The right to education was first expounded in India’s original 1950 constitution with the

caveat that such a right could only be delivered “within the limits of economic capacity” (Law

Commission of India, 1998, p. 23). Despite this caveat, Indian state practice remained tightly

coupled with this goal, and the nation saw drastic expansion in education access in the decades

following independence. The mass expansion and near universalization of primary education was

precipitated by three major initiatives beginning in 1994: the World Bank-funded District

Primary Education Programme (DPEP), Sarva Shiksha Abhiyan (Education for All Movement),

Page 9: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

4

and more recently, the Right of Children to Free and Compulsory Education Act of 2009. As a

result of these programs, from 1994 to 2010 the primary gross enrollment ratio (GER) increased

from 95 to 112 percent, the secondary GER increased from 45 to 63 percent, and today, 99

percent of the nation’s children are located within 1 kilometer of a school (UNESCO-IBE, 2011;

Hill & Chalaux, 2011).2

In 2002, after several decades and endless parliamentary debates and revisions, the 86th

amendment to the constitution lifted the “economic capacity” caveat, placing the burden of

education funding for children in the 6 to 14 age group squarely on the state. The specifics on

how this constitutional right would be delivered were enumerated in the Right of Children to

Free and Compulsory Education Act which entered into law in 2009 and came into effect in

April of 2010.

2.2 The state rationale for a common school system

India’s rationale for providing free and compulsory state education is influenced by

human capital theory, social capital theory, and the view of education as a catalyst for equality.

The 1998 Law Commission Report, which provided the core rationale for the 86th amendment,

directly references the logic of human capital theorists including Theodore Schultz, Gary Becker,

and Alfred Marshall (Law Commission of India, 1998, p. 2). While Indian policy documents

make only limited references to social capital, the state has long resisted the creation of a private

school voucher program—despite research claiming vouchers could increase efficiency and

equity (Kingdon, 1996; Muralidharan & Kremer, 2009)—on the grounds that a segregated

2 Figure 1 shows more recent enrollment trends in rural areas which are the focus of the present

analysis. Overall enrollments have fluctuated between roughly 94 and 95 percent since 2007,

which suggests that India is suffering from a “last mile problem” in achieving universal access, a

common phenomenon in developing countries.

Page 10: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

5

system would “weaken social cohesion” (National Council of Educational Research and

Training, 1966). The view that privatization would pose a threat to the social capital necessary to

build an effective common school system is convincingly articulated by Jean Drèze and Amartya

Sen in their recent book, An Uncertain Glory: India and its Contradictions:

The most hidden penalty of greater reliance on private schools is that it tends to take

away from state schools the children of precisely those parents that could make state

schools more responsible and accountable. Reliance on private schools can make the

problems of state schools much greater, by providing a way out for the more prosperous

and more vocal families. (Drèze & Sen, 2013, p. 138)

This position is reflected in the 1964-66 Education Commission report, which argues that an

integrated common education system “compels the privileged and powerful to bring about the

early improvement of the system of public education” (Seshadri, 1976, p. 229). Finally, India

views public schooling as the primary vehicle for equalizing inequalities “transmitted through

the social environment and the accident of birth” (Government of India, 1992, p. 3-6) among the

nation’s historically marginalized groups, including Scheduled Castes, Scheduled Tribes,

Muslims, girls, and the rural poor, which make up the great majority of the population.

2.3 From a common school dream to growing privatization

Despite the state’s goal of fostering equality and social cohesion through education, the

mass expansion of the public school system in India has coincided with a rapid acceleration in

student migration to private schools (Chavan, 2012). Between 2006 and 2012, rural primary

private enrollments increased from 19.6 to 27.8 percent (see Figure 2), with even swifter

migration in urban areas and at the secondary level (Kingdon, 2007). As of 2012, 36 percent of

India’s 1.4 million schools were private, and it is estimated that an astonishing 41 percent of

primary school children will be in private schools by the end of 2014 (Chavan, 2012).

[Figure 2 about here]

Page 11: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

6

This growing public school exodus has been attributed to perceived declines in the

quality of education in government schools (Kingdon, 2007), and it is widely acknowledged that

the quality in public schools is remarkably poor in India. The Indian Human Development

Survey, for example, revealed that only 50 percent of 8 to 11 year-olds in public schools can read

a three-sentence paragraph, and only 43 percent can subtract two-digit numbers (Drèze & Sen,

2013). Further, as this paper will explore in detail, student achievement appears to be worsening

since the Act went into effect. Indeed, the ratio of lowest performing children in both reading and

math (i.e., when presented with 5 letters or 5 numbers, they can only identify 4 or fewer) has

increased considerably since 2010. Notably, these declines are not restricted to public schools,

and private schools have experienced similar declines in student achievement since the Act went

into effect, a critical point to which I will return below.

3. Student achievement in relevant literature

To better understand the factors contributing to recent student achievement declines in

India, I review the extensive research on the relative weights of socioeconomic status, peer

effects, and school inputs in education production function models. Following this critical

literature review, I discuss how these findings may be applicable to the largely under-explored

Indian context.

3.1 Socioeconomic status and student achievement

Across the literature, family background or student socioeconomic status (SES) is

consistently found to be a strong predictor of student achievement. In his landmark study on

post-segregation era schooling in the U.S., Coleman (1966) found that after controlling for

socioeconomic factors, school-level characteristics account for very little of the variation in

Page 12: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

7

student achievement. In the U.S. context, Rothstein (2004) argues that social class greatly

overshadows the role of school inputs and thus views broader social reform as a necessary

condition for equitable education outcomes. In developing countries, studies similarly find that a

considerable portion of variance in student achievement can be explained by socioeconomic

status, with some estimates placing the figure at over 50 percent (Fuller & Clarke, 1994). The

extensive findings on the influence of SES on student achievement carry over to the Indian

context and indeed, an internal (though unpublicized) analysis of ASER’s data found that

correcting for SES reduced the learning differential between public and private school students

from 8.6 to 2.9 percent (Wadhwa, 2009).

3.2 Peer effects and student achievement

Peer effects are widely considered important by scholars, policymakers, and consumers

of education alike, though measuring their role in student achievement presents many

challenges.3 Despite this, many researchers have attempted to measure peer effects and

consistently find that they matter for student achievement. A more nuanced debate in the peer

effects literature, however, is whether peer effects are linear for all ability groups. Hoxby (2000)

and Hanushek et al. (2003) find only limited evidence of non-linearity. Sacerdote (2001), who

conducted a peer effects study using random assignment, finds strong evidence of non-linearity,

with high ability peers having a significant, positive impact on lower ability peers whereas high

ability peers are unaffected by low ability peers. Likewise, Summers and Wolfe (1977) and

Zimmer and Toma (2000) find that high ability peers greatly benefit low ability students, while

high ability students do not seem to be hurt by low ability peers.

3 Foremost among these is the complexity of the task of disentangling peer effects from SES and

school-level inputs, all of which are highly correlated and thus introduce imprecision into peer

effect econometric estimates (Hanushek, Kain, Markman, & Rivkin, 2003; Hoxby, 2000).

Page 13: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

8

The findings that peer effects matter more for low ability groups have important

implications for the school choice debate. If choice policies remove higher ability students from

public schools, they will fundamentally alter the peer composition in public schools and thus

negatively impact those left behind. Epple and Romano (1998), through a theoretical model on

the interaction between private and public schools, demonstrate that high ability, low income

students benefit from voucher policies whereas the majority low ability, low income students are

indeed hurt by them. Ladd (2002) argues that peer composition plays a significant role in school

choice, with parents choosing schools based on the SES composition of peers. If school choice

were truly universalized, she argues, private schools would have little incentive to accept low

SES students, as the schools have a vested interest in maintaining a high SES peer composition.4

Further, to the extent that “the stock of good peers is finite” (Somers et al., 2004, p. 69), it is

logically impossible for all disadvantaged students to benefit from a choice model.

The discussion on the relationship between peer effects and student achievement has

important implications for the Indian context. India’s Right to Education policies have

historically favored equalization over improving average student achievement, with national

education policy frameworks developed with the explicit goal of “[correcting] unfair distribution

of educational resources and effort” (Seshadri, 1976). These policy statements suggest that,

confronted with a tradeoff between equality and quality in education, the nation would prioritize

social equity goals.

4 In the Indian context, the Act sought to address this tendency by mandating private schools

reserve 25 percent of seats for marginalized children. Consistent with Ladd’s theory, Srivastava

and Noronha (2014) provide evidence that private schools have either ignored the reservation

requirement or intentionally segregated lower SES students from the general private school

population.

Page 14: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

9

While research in developing countries on the interaction between school choice and peer

effects has been limited, the findings are nevertheless consistent with the literature on developed

countries. In looking at ten countries across Latin America, for example, Somers et al. (2004)

find that including peer effects in an education production function reduces the differential

between public and private student achievement to zero and, in some cases, even favors public

schools. It is therefore worth considering whether or not the inclusion of peer effects in an

education production function model would reduce the public-private achievement gap in India.

Indeed, there is some preliminary evidence that suggests that this would be the case. Since the

Right to Education Act went into effect, declines in public school student achievement have been

roughly mirrored by declines in private school achievement (ASER Centre, 2013c). One possible

explanation for this is that as lower SES students enter the public system, higher SES students

shift to private schools which results in a corresponding downward pull on student achievement

within private schools. While the declining achievement in private schools might be explained by

decreasing average SES of private school students, the declining achievement in public schools

may be caused by both declines in average SES and the loss of higher ability peers.

Unfortunately, the large body of quantitative research on the relative effectiveness of private

versus public schools in India has consistently ignored the potentially large role peer effects play

in determining student achievement (see Muralidharan & Kremer, 2009; Tooley, Dixon, &

Gomathi, 2007; Gouda, Das, Goli, & Pou, 2013; and Kingdon, 1996).

3.3 School inputs and student achievement

In the developed world, the debate on the role of school inputs in student achievement is

contentious (Hanushek, 2002). In the context of developing countries, however, there tends to be

more agreement that school-level inputs matter (Gamoran & Long, 2007). In education

Page 15: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

10

production function models, such school-level inputs vary but tend to include basic

infrastructure, teacher attendance, teacher education and experience, student-teacher ratios, and

time-on-task (USAID, 2008). Interestingly, as schools approach a certain baseline level of school

inputs, socioeconomic status begins to weigh more heavily in determining student achievement

(Baker, Goesling, & LeTendre, 2002). This finding is particularly relevant to the Indian context,

where the Act and its policy predecessor Sarva Shiksha Abhiyan (or ‘Education for All

Movement’) have enabled the public school system to approach this baseline level. As a

consequence, school inputs may carry even less weight today than they have historically.

Despite solid reasons for believing that non-school variables play a dominant role in

student achievement, most research on student outcomes in India focuses on school inputs, with

particular attention given to teacher attendance and time-on-task (Muralidharan & Kremer,

2009). In public schools, teacher attendance and time-on-task are consistently poor, with 25

percent of teachers found absent on a given day; of those present, only half were engaged in

teaching activities (Kingdon, 2007). The correlation between low teacher inputs and low student

achievement has generated what is widely considered to be a crisis in Indian public education,

which has only been made worse by the expansion of the system as a result of the Act (Chavan,

2012). While teacher attendance and time-on-task are undoubtedly important in India, recent

demographic shifts in public schools offer reason to believe that their role in achievement trends

is overstated. Further, there is reason to suspect that, in controlling for SES and peer effects, the

effect of teacher-level inputs on student achievement may actually be improving in public

schools. This latter point is in direct opposition to the general belief that India’s government

schools are declining in their ability to deliver quality education to the majority of India’s school

children.

Page 16: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

11

4. Conceptual framework

The prevailing opinion in India is that the post-Act declines in student achievement are

endogenous to the school system itself (Chavan, 2012; Kaul, 2013). This theory parallels an

education production function in which school inputs are the main predictors of student

achievement. While scholars have included socioeconomic indicators in earlier education

production function models demonstrating the superiority of private school quality in India

(Kingdon, 1996), there appears to be a gap in the research on both the inclusion of peer effects in

such models as well as the analysis of more recent data on student achievement.

I theorize that an accurate model of post-Act student achievement should give careful

consideration to shifting student socioeconomic status and peer composition in public schools.

Adapting education production function models of Coleman (1966), Summers and Wolfe,

(1977), and Zimmer and Toma (2000), I contend that the Act may have contributed to the

acceleration of changes in socioeconomic and peer composition in public schools in two ways.

First, by retaining a growing number of first generation learners and previously out-of-school

children, the Act has potentially lowered the average socioeconomic status and student

preparedness in government schools and consequently, lowered average achievement levels.

Second, this lowering of average achievement and socioeconomic status in public schools may

have expedited the migration of higher ability and higher socioeconomic status students, who are

flocking to private schools in ever-increasing numbers. Possible reasons for this public school

exodus include both the perception that public schools are declining in quality (Kingdon, 2007)

Page 17: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

12

as well as India’s version of “white flight”5 resulting from the integration of historically

marginalized groups.

Per the concept map in Figure 3, this conceptual framework can be modeled with two

separate production functions—one for public schools and one for private schools—which

interact with each other. The migration of high SES students from public to private schools alters

the SES and peer composition in both public and private schools and, consequently, alters

student achievement. The concept map shows how the migration from public to private schools

impacts variables within both production function models. Note that this model assumes that

migrating students represent both the highest SES groups in the public schools that they leave

and the lowest SES groups in the private schools that they join. This concept map illustrates that

when higher socioeconomic students migrate from public to private schools, the result is a net

decline in average achievement in both public and private schools. Consistent with theories on

the non-linearity of peer effects demonstrated by Summers and Wolfe (1977), Zimmer and Toma

(2000), and others, higher SES student migration to private schools will have a net negative

effect on student achievement in public schools but a null effect on student achievement in

private schools. Overall, the net effect on student achievement is negative for both public and

private schools but more negative for public schools due to the nonlinearity of peer effects. This

model appears to be consistent with the data that show student achievement is declining in both

public and private schools, but declining at a faster rate in public schools (ASER Centre, 2013b).

Critically, these changes in average student achievement occur regardless of school-level factors,

5 In the U.S. post-segregation context, “white flight” refers to student “exit or avoidance of

racially mixed public schools” which risks “[undermining] the racial contact that desegregation

policy seeks to enhance” (Clotfelter, 2001, p.199).

Page 18: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

13

challenging the notion that school quality is the main culprit in declining student achievement in

Indian public schools.

[Figure 3 about here]

While the change in student achievement in private schools is beyond the scope of this

paper, considering the shifts in private-public school composition has allowed me to develop a

more nuanced portrait of changing student achievement in public schools. The production

function model for public school student achievement, now taken in isolation, offers a

springboard for my key research questions which enable me to explore the dynamics of changing

variables affecting student achievement within public schools pre- and post-Act. Specifically, my

central hypotheses are as follows: (i) since the Act went into effect, average socioeconomic

status and peer quality have declined in public schools; (ii) as average socioeconomic status and

peer quality decline in public schools, average student achievement declines as well. Because

socioeconomic status and peer composition are stronger predictors of student achievement than

school inputs, this leads to a misleading overall average decline in student achievement; and (iii)

since the Act went into effect, school inputs have improved in public schools. As a result,

controlling for socioeconomic status and peer composition will improve the odds ratios on

school inputs post-Act, where the outcome variable is student achievement. This final

hypothesis, if confirmed, will challenge the view that India’s public schools are in a state of

accelerating crisis. This latter point makes intuitive sense given the central government’s

credible commitment to expanding opportunity to learn through increased school accessibility,

improved infrastructure, and reduced student-teacher ratios.

To restate, this study does not posit that school inputs are unimportant for student

achievement; indeed they are by far the most accessible education production function variables

Page 19: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

14

for policymakers in India and around the globe. Rather, I argue that the singular emphasis on the

influence of school-level factors in determining learning outcomes is reductionist and short-

sighted, and the changing school landscape in India—which has involved rapid shifts in student

ability, preparedness, and SES composition between public and private schools—deserves much

deeper consideration.

5. Data and methods

In order to make causal inferences about the effect of Right to Education Act legislation

on learning outcomes, a quasi-experimental or “natural experiment” approach is warranted

(Murnane & Willett, 2011). For the purposes of this study, this would require exploiting

exogenous variation in the Act’s implementation; unfortunately, there is no evidence that this

legislation was implemented in random phases across villages, districts, or states in India.

Consequently, this paper is observational in nature. The ultimate goal of this observational study,

however, is not to use correlational data to make causal inference—an approach fraught with

internal validity threats such as selection and omitted variable bias (Stock & Watson, 2011).

Rather the purpose is to question the validity of causal claims drawn by others from the Annual

Status of Education Report data. This is achieved by introducing previously omitted SES and

peer effects variables into an education production functions to better understand the factors

underlying recent achievement declines in India. As such, this paper aims to expose—rather than

commit—the oft-derided research transgressions of correlational studies.

Page 20: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

15

5.1 Data description

Established in 2005, the Annual Status of Education Report (ASER) is the largest

independent survey to assess student learning in rural India.6 Targeting approximately 300,000

households each year, ASER employs a two-stage probability proportional to size (PPS) random

sampling methodology to ensure cross-sectional, representative coverage of India’s rural

population. The survey is administered to every rural district in the country, and each district is

represented by a rotating panel of 30 villages. Enumerators7 then randomly select 20 households

via a village-level customized sampling strategy which accounts for variation in household

distribution. While enumerators make every effort to visit households while tenants are home,

households with no one home are excluded from the survey which may potentially introduce

selection bias into the household sample (ASER Centre, n.d.; ASER, 2013).

The ASER household survey contains simple reading and math instruments—

administered in the language selected by the child—which measure the highest grade level at

which the child can read or perform math.8 The ASER dataset also includes extensive household

6 The exclusive focus on rural districts is borne of Pratham’s belief that—in contrast to urban

areas—rural areas in India suffer from “an enormous shortage of people able to design,

implement, and understand the nuts and bolts of assessment” (Establishment of ASER Centre,

n.d.). 7 ASER ensures data quality by requiring each volunteer enumerator to attend an annual three-

day training session which includes testing and field practice. Approximately 28 percent of the

sampled villages are monitored by a Master Trainer during survey administration, and any

questionable village-level results are verified or re-checked. Finally, ASER conducts periodic

annual field audits to ensure the survey is being administered per the training protocols (Ensuring

ASER Data Quality, n.d.). 8 The reading instrument has four levels including a grade level 2 paragraph, a grade level 1

sentence, simple word identification, and simple letter identification. If the child is unable to

identify simple letters, he or she is marked as having no reading ability. Each child in the survey

is given a dummy equal to 1 for the highest level at which he or she can read. The math

instrument similarly presents four problems, and a child is given a dummy equal to 1 for highest

level at which he or she can perform math among following categories: ‘nothing,’ ‘single digit

Page 21: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

16

variables on family size, household assets, parental education, and other socioeconomic status

indicators such as access to electricity, books, and newspapers. Finally, beginning in 2007 ASER

began sampling one government school within each village (chosen by the enumerator) to gather

data on teacher and student absences; school infrastructure, resources, and amenities; and

enrollments. Following the passage of the Act, the school-level dataset was expanded to include

indicator variables that are aligned with the legislation’s core mandates.

5.2 Empirical strategy

This study aims to measure changes in math and reading proficiency in rural public

schools pre- and post-Act (2009 to 2012). Because of shifting demographics in public schools in

recent years—driven by rural out-migration and the exodus of millions of students from public to

private schools (Chavan, 2012)—a simple comparison of mean reading and math proficiency

over time may be considerably biased. I thus begin by tabulating descriptive statistics to confirm

parallel post-Act trends in the ASER data, including public school out-migration, the lowering of

student- and peer-level SES in public schools, increasing school inputs, and declining math and

reading performance.

I then regress both reading and math performance on three categories of independent

variables including student socioeconomic status, peer effects, and school characteristics

(including Act-mandated school inputs):

number recognition,’ ‘double digit recognition,’ ‘subtraction with carry over,’ and ‘division.’

The latter four categories correspond to grade levels 1, 2, 3, and 4 in mathematics, respectively.

Page 22: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

17

Where Yis is a binary indicator for whether or not child i in government schools s is proficient in

reading or math at a grade-appropriate level;9 SESi is a vector of socioeconomic status controls

for child i including binary indicators for mother’s schooling status, toilet ownership, and

television ownership; Peerss-i is a vector of SES variable means for children in school s

excluding child i's status; Schools is a vector of school inputs including the teacher absenteeism

rate, student-teacher ratio, and dummy indicators for school infrastructure and amenities; δ is a

time fixed effect for year t; and ε is the error term. While the selected SES proxies are imperfect

measures of both student SES and peer quality, my analysis is limited by the availability of

variables both within the dataset and across years. The purpose of this model is to observe the

relative magnitude of the odds ratios on each of the three categories of education production

function inputs.

Next, I use the above variable definitions to explore the relationship between math and

reading proficiency and student SES, peer effects, and school inputs in each of the years pre- and

post-Act (2009 to 2012):

9 For the purpose of this study, I restrict my analysis to children’s math scores for grades 2-5 and

their reading scores for grades 2-3 (see footnote 8 for more detail). Because the test is

administered mid-year, it is unreasonable to expect the child to have obtained his or her grade-

appropriate math or reading level by the time the test was administered. I therefore generate two

new Y variables called proficient_reading and proficient_math, and assign a dummy equal to 1 if

the child is able to read or perform math at the grade directly below his or her current grade or

better, and equal to 0 if he or she is unable to read or perform math at the grade level from which

he or she most recently advanced.

Page 23: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

18

These models enable exploration of the extent to which the magnitude of the odds ratios for each

of the three categories of variables—student SES, peer effects, and school inputs—are changing

by year and can thus lend credence to my hypotheses that non-school factors have played a larger

role in learning outcomes post-Act.

6. Findings

6.1 Overall trends in rural Indian education

Table 1 provides summary and descriptive statistics for all ASER observations in rural

India for 2009 to 2012. The overall enrollment rate for 6-14 year olds exceeds 96 percent across

all years in the sample. Consequently, enrollment increased by only 0.4 percent between 2009

and 2012. As reported in Figure 2, public versus private enrollment shares have shifted

dramatically, with approximately 7 percent of surveyed students moving from public to private

schools between 2009 and 2012.

[Table 1 about here]

Overall student SES—measured by mothers’ schooling status and household assets—

remained relatively stable in rural areas overall, though asset ownership has slightly declined

since 2010, which may be suggestive of rural out-migration. Government school students

experienced sharper declines in SES indicators, with the percentage of student mothers with at

least 1 year of education dropping over 4 percentage points to 45.79 percent between 2010 and

2012. Over the same time period, television and toilet ownership among public school students

declined by 2.55 and 3.87 percentage points respectively, whereas they increased slightly among

private school households. For the sample of class 2-5 public school students used in the

analyses below, asset declines are slightly higher at 5 percent each (see Table 2).

Page 24: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

19

[Table 2 about here]

Consistent with widespread media reports, student reading and math performance

declined considerably post-Act, with math proficiency in public schools dropping nearly 17

percentage points between 2009 and 2012. While declines in reading proficiency in government

schools were not as drastic (and appear to have improved somewhat since 2011), they remain

well below 20 percent for public school students. While private schools fare better in both math

and reading, they too have experienced overall declines since 2009 which is consistent with the

conceptual framework outlined in section 4.

6.2 SES and student achievement in government schools

Tables 3a and 3b present odds ratios for math and reading proficiency regressed on

student-level SES dummy variables—including mother’s schooling status, household television

ownership, and household toilet ownership—controlling for year, student age, grade level,10

and

gender. As shown in Model 1 in Table 3a, all three SES variables are strong predictors of math

proficiency, with higher SES students 30 to 71 percent more likely to be math proficient than

their low SES counterparts. Reading proficiency (Model 1, Table 3b) can be similarly predicted

by SES proxies, though the odds ratio on mother’s schooling is higher for reading than math.

[Tables 3a and 3b about here]

10

Through my research I have found that, regardless of the model, student grade level is the

strongest predictor of reading and math proficiency (see Table 1 in Appendix II for a simple

model). Such a finding suggests that learning deficiencies are additive and, as such, post-Act

shifts in student preparedness may explain some of the observed declines in reading and math

performance. Table 2 in Appendix II presents math and reading proficiency odds ratios on year

dummies interacted with age-appropriate grade status (=0 if the student is over-aged). These

results suggest that student over-aging in rural India has become an increasingly strong predictor

of math and reading deficiency in each year post-Act. While beyond the scope of this paper, this

interesting finding warrants further analysis.

Page 25: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

20

To determine whether the odds ratio on student socioeconomic status changes by year, I

regress math and reading proficiency on the interaction between the SES dummies and dummies

for each year in the dataset, controlling for peer- and school-level characteristics.11

According to

the results in Table 4, the reading proficiency odds ratio on all three SES indicators increases

each year from 2010 to 2012, suggesting that higher SES has become a stronger predictor of

reading proficiency post-Act. Similarly, the odds ratio for math proficiency increases for each of

the SES indicators after 2010, however it drops between 2011 and 2012 for toilet and television

ownership.

[Table 4 about here]

Most strikingly, the lowest SES students—defined as those with uneducated mothers and

no television and toilet at home—are considerably less likely to be reading and math proficient

as the years progress. In 2010 for example, the lowest SES students were 25 percent less likely to

be math proficient than their 2009 counterparts; in 2012, they were nearly 70 percent less likely

to be math proficient (see Table 4). Similarly, the lowest SES students in 2010 were 28 percent

less likely to be reading proficient than their 2009 counterparts; in 2012, they were over 52

percent less likely to be reading proficient.

Table 5, which provides descriptive statistics on reading and math proficiency for high

and low SES government school students12

by year, sheds light on these findings. While math

proficiency drops for both high and low SES students, declines were largely absorbed by

students in higher grade levels and students of low SES. Math proficiency for high SES second

graders, for example, remained virtually unchanged whereas math proficiency for low SES fifth

11

2009 is omitted due to multicolinearity 12

Here I define “high SES” students as those who have a household television and toilet and

whose mothers have at least 1 year of schooling. Conversely, “low SES” students have no

household television and toilet and their mothers never attended school.

Page 26: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

21

graders dropped over 20 percentage points. Reading proficiency rates improved overall for high

SES students whereas they dropped 6.75 percentage points to a shocking 6.85 percent for low

SES students. These higher grade declines may be due to the large influx of first generation

learners in years prior to 2009 (see Figure 1), a point which will be further explored in the next

section.

[Table 5 about here]

Table 6 provides additional support for the hypothesis that declines in student learning

can be explained by parallel declines in SES. Because the ASER survey is representative at the

district-level, I collapse math and reading proficiency and SES proxies into district-level means

and reshape the data into a longitudinal panel. I then subtract district-level mean math and

reading proficiency, mother’s education status, television ownership, and toilet ownership in

2009 from their 2012 values and regress the differences in mean proficiency on the differences in

SES means over the same time period. While the results for reading proficiency are not

statistically significant, the results for math proficiency are both significant (p<0.01 for mother’s

schooling and television ownership) and large, with an adjusted R-squared of 0.066. This

suggests that—across all rural India—at least 6.6 percent of the variation in district-level math

proficiency rates between 2009 and 2012 can be explained by variations in district-level SES.

[Table 6 about here]

6.3 Peer effects and student achievement in government schools

Per the summary and descriptive statistics in Table 2, peer SES indicators—defined as

mean SES status for all students in a given child’s school (excluding the child’s own status)—

roughly mirror SES indicators for the overall sample population. Table 3a presents odds ratios

for math proficiency including student SES (Model 1), peer SES (Model 2), and student and peer

Page 27: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

22

SES combined (Model 4). Because student and peer SES are positively correlated, the odds

ratios on student and peer SES variables in Model 4 are reduced relative to Models 1 and 2 with

the exception of peer television ownership. Comparing the odds ratios for all student- and peer-

level SES variables in Model 4 reveals that being in a school with high SES peers has a greater

effect on math proficiency than being in a high SES household. These results hold for reading

proficiency (Table 3b) with the exception of toilet ownership, which has a higher odds ratio at

the household level.

The findings that peer effects matter for the overall sample is unsurprising, given the

extensive literature on the topic. More central to this study, however, is whether the impact of

peer effects differs by year. Table 7 presents math and reading proficiency odds ratios by peer

SES interacted with year dummies, controlling for student- and school-level characteristics.

While the effect of having high SES peers does not appear to be trending in a particular direction

post-Act, the effect of having low peers does. In 2010, for example, children with the lowest SES

peers were 26.6 percent less likely to be math proficient than their 2009 counterparts; in 2012,

they were over 68 percent less likely to be math proficient. Results are similar for reading

proficiency, although the shift is less dramatic. This finding is unsurprising, however, given that

average reading proficiency appears to have stabilized or even slightly improved13

in public

schools in recent years. Overall, the data support the story that the growing proportion of low

SES peers is negatively impacting learning outcomes in public schools.

[Table 7 about here]

13

It is important to reiterate that—despite rate stabilization or even modest improvement—

reading proficiency remains abysmally low in rural India, with only 15 percent of 2-3 grade

students capable of reading at a grade-appropriate level.

Page 28: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

23

6.4 School inputs and student achievement in government schools

Table 2 presents summary and descriptive statistics on school inputs—including whether

the school has a toilet, drinking water, boundary wall, and provides a midday meal as well as the

student-teacher ratio and teacher attendance rate. Across all domains, school inputs have been

steadily improving, with a few exceptions (see e.g., boundary wall in 2010 and school toilet in

2011) which may suggest uneven school-level data collection across years.14

Teacher

attendance—the ratio of teachers present on the day of the survey to government appointed

teachers—dropped roughly 3 percent between 2009 and 2010 and remained steady at 85 percent

beyond 2010.

Tables 3a and 3b present odds ratios for overall math and reading proficiency regressed

on school inputs. In both tables, Model 3 presents the odds ratio on school inputs alone with

Model 5 including student- and peer-level controls. Relative to Model 3, the inclusion of student-

and peer-level controls in Model 5 increases the coefficients on school inputs with negative odds

ratios and decreases the coefficients on school inputs with positive odds ratios, suggesting that

different school inputs are differentially correlated with the omitted student- and peer-level SES

factors. In general, school inputs appear to be slight predictors of deficiency in math whereas

they are slight predictors of proficiency in reading.

Table 8 presents math and reading proficiency odds ratios for school inputs interacted

with year dummies, controlling for student- and peer-level characteristics. Overall, there appear

to be no consistent patterns across years and variables and, regardless of year, school inputs

14

The ASER village sampling strategy is designed to ensure representation of students at the

district-level. School-level data is drawn from a non-randomly selected school from each

sampled village.

Page 29: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

24

appear to have a mixed or slightly negative effect on student learning outcomes (with the

exception of toilets and boundary walls).

[Table 8 about here]

To summarize, this study finds that since the Act went into effect, average socioeconomic

status and peer quality have declined in public schools, which is mirrored by declines in student

achievement. Indeed, at least 6.6 percent of the variation in district-level math proficiency rates

between 2009 and 2012 can be predicted by variations in district-level SES. This, combined with

the evidence presented here that peer SES is an even stronger predictor of subject proficiency

than household SES, provides support for the hypothesis that extra-school factors account for a

significant proportion of post-Act learning declines. Overall enrollments increased only slightly

after 2009, suggesting only a marginal effect of the inclusion of first generation learners post-Act

on learning outcomes. Finally, including student- and peer-level SES variables in an education

production function reduces the magnitude of both positive and negative effects of school inputs,

though school inputs still have a mixed or slightly negative effect on student learning outcomes.

7. Discussion

The view that post-Act declines in math and reading proficiency are the product of

deteriorating quality in public schools is pervasive (Balasubramanian, 2014; Banerjee & Duflo,

2013; Nanda, 2014; Chavan, 2012). This simple narrative ignores the important role that shifting

student SES and peer composition have played in learning declines. Driven by both rural-urban

migration and public-private school migration, these shifts have led to a deterioration of overall

SES and peer composition in public schools. As this study has demonstrated, both student- and

peer-level socioeconomic status have become increasingly strong predictors of math and reading

Page 30: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

25

deficiencies post-Act, with the gap between low and high SES students widening each successive

year. The following section will summarize this study’s implications for year-wise

interpretations of ASER data and close with a discussion of what more accurate interpretations

mean for the project of building a common education system in India.

7.1 Public school out-migration

The descriptive statistics in Table 1 show that across the entire ASER sample, mothers’

schooling and asset ownership declined between 2009 and 2012. While possibly due to sampling

variation across the rotating panels of villages, this unexpected finding is suggestive of a larger

pattern of rural-urban migration of higher SES families (Khan, Hassan, & Shamshad, 2011). In

India, urban migration is accelerating. According to census data, the share of Indians living in

urban versus rural areas increased from 27.81 to 31.16 percent between 2001 and 2011, and this

figure is expected to increase to 42.5 percent by 2025 (Tulloch, 2011).15

Given that the ASER

assessment is administered exclusively in rural areas, interpretations of year-wise learning trends

should account for the effects that rural out-migration will have on district averages. The

relationship between rural out-migration and ASER learning declines is an important area for

future research.

Public to private school migration has likewise been rapid, with the private school share

of total enrollments increasing by 7 percentage points to 30 percent between 2009 and 2012 (see

Table 1). This public school exodus has led to an ever-expanding gap between the

socioeconomic status and peer composition of public and private school students. As this study

demonstrates, student- and peer-level SES in public schools have declined in parallel with the

15

This overall shift does not take into account differential population growth between urban and

rural areas.

Page 31: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

26

public school exodus and as such, both factors have become increasingly strong predictors of

academic performance. Consequently, interpretations of year-wise ASER learning trends should

account for the effects that public to private school migration will have on public school

averages.

7.2 Increased enrollment of first generation learners

As presented in Section 6.1, overall enrollments increased only slightly (0.4 percent),

which suggests the enrollment of first generation learners between 2009 and 2012 plays only a

marginal role in recent learning declines. It is important to consider, however, the effect that pre-

2009 cohorts might have on achievement levels as they move into higher grades. Between 2006

and 2007, for example, overall enrollments increased 3 percentage points (ASER Centre, 2013a).

To the extent that learning deficiencies are additive and new enrollments consisted of first

generation learners entering the system in earlier grades, pre-2009 enrollment trends may explain

some of the post-2009 declines among higher grade students.

7.3 The growing performance gap in government schools

The widening performance gap between low and high SES students demonstrated in this

study is a topic for further research. In the U.S. context, Reardon (2011) suggests that such an

achievement gap may be due in part to a growing gap between high and low SES families’ time

and resource investment in their children’s cognitive development. This hypothesis may carry

over to the Indian context, where the perceived value of, and investment in, education is

increasing dramatically. A recent survey conducted by Credit Suisse, for example, shows that

parents in India now spend, on average, 7.5 percent of their earnings on their children’s

education (“Education: A Billion Brains,” 2012). While this is a plausible explanation for the

Page 32: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

27

achievement gap in India, it does not account for the fact that higher SES students also

experienced declines in math proficiency, particularly in the higher grades (see Table 5). While

potentially explained by rural-urban migration, public-private migration, the loss of higher

ability peers, and the increased enrollment of first generation learners, this trend may be

suggestive of the reallocation of school resources toward improving reading and focusing on

building a solid base of learning in earlier grade levels.

7.4 Implications for school inputs

The mixed or negative effect of school inputs on learning outcomes may be due to

problems with the data collection and sampling methodology, but there are many other possible

stories that might fit with these results. Some scholars suggest, for example, that the Act has

diverted teachers’ time from teaching and toward other tasks, such as construction projects and

midday meal preparation (Muralidharan, 2013). Equally plausible, these unexpected findings

may be due to the fact that school inputs are correlated with omitted student and peer

characteristics that are not adequately captured by the variables in this dataset. Finally, there is

some evidence that households in India respond to anticipated improvements in school inputs by

reducing their own spending on school-related expenses. In such cases, school inputs have a null

effect on learning outcomes (Das, Dercon, Habyarimana, Krishnan, Muralidharan, &

Sundararaman, 2011). In general, this counterintuitive result deserves much deeper investigation,

including the exploration of other more comprehensive school data sets such as those produced

by India’s District Information System for Education (DISE), which—unlike the ASER data—is

representative at the school-level.

While I cannot rule out the possibility that schools—particularly those serving low SES

children—have declined in quality since 2009, it is interesting to note the finding that teacher

Page 33: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

28

absenteeism has become a progressively weaker predictor of math proficiency and there are no

statistically significant year-wise changes on the reading proficiency odds ratio for teacher

attendance (see Table 8). This result suggests that changes in teacher attendance explain little of

the post-Act declines in learning outcomes, which in part challenges the commonly held view

that the “alarming degeneration” in learning outcomes in recent years is the product of

deteriorating teacher performance (Chavan, 2012, p. 1).

8. Conclusion

The finding that SES and peer variables have become stronger determinants of student

achievement has important implications for the contentious debate surrounding the Right to

Education Act and the future of Indian public education. First, they showcase the need to temper

brusque interpretations of year-wise ASER findings, which frequently operate on the assumption

that schools themselves are the singular force behind learning declines. Such interpretations have

led to widespread calls by many preeminent scholars of Indian education for privatization

schemes such as vouchers. To the extent that learning declines are a product of deteriorating SES

and peer quality in public schools, voucher schemes may exacerbate learning differentials

between lower and higher SES students, thereby undermining India’s dream of “neutralizing the

distortions of the past” (Government of India, 1998, p. 3). Unless the public system is eliminated

entirely,16

privatization threatens to shift critical financial resources and social capital away from

government schools as well as lead to ‘cream skimming’ of higher ability peers. Such schemes

risk contributing to the ghettoization of public schools, upon which the vast majority of Indian

16

Few, if any, scholars advocate for the complete elimination of the government system given

the role that “state schools are meant to play and have played in the educational transformation

of most countries in the world” (Dreze & Sen, 2013, p. 136). In addition, overreliance on private

schools is risky, given their ability to exit the market at any time (Mcloughlin, 2013).

Page 34: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

29

children—and marginalized children in particular—depend. While many scholars argue that

privatization schemes support equity by enabling poorer children to access educational resources

previously reserved for upper classes, they often ignore the possibility that the types of

households that self-select into privatization programs may be fundamentally different—in terms

of motivation, resources, and time—than those who do not. As such, privatization policies risk

“[stratifying] the poor into the better-off, who go private, and the worse-off, who go public”

schools (Mcloughlin, 2013, p. 5).

Second, the findings that SES and peer composition matter suggest the need for patience

in India and other developing countries that have undergone rapid expansion of their public

education systems—particularly in countries such as Nepal and Pakistan, which have their own

pending Right to Education legislation. Because rapid system expansion involves the inclusion

of more low SES students, short-term declines in learning outcomes are to be expected. While

this call for patience does not in any way suggest that issues such as pervasive corruption and

absenteeism be swept under the rug, I posit that there may be a shorter-term equality-quality

trade off in building a functional common education system. Policymakers in India should not

lose sight of their goal of building an equitable system in exchange for superficial and potentially

misleading improvements in learning outcomes facilitated by privatization schemes.

Given the equity concerns outlined above, post-Act reform policies should be focused on

enhancing the performance of the public education system rather than inadvertently contributing

to its decline. While this study offers limited insight on the role that school factors play in

changing educational outcomes, few would dispute that minimum resource requirements such as

classrooms, toilets, and sufficient teachers—i.e., the very items mandated by the Right to

Education Act—are necessary to sustain regular attendance (see e.g., Hanushek & Woessmann,

Page 35: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

30

2007). Meanwhile, many initiatives for improving learning outcomes in public schools show

promise, including remedial education projects such as Pratham’s Balsakhi program (Banerjee,

Cole, Duflo, & Linden, 2007) as well as teacher monitoring and incentive schemes (Duflo,

Hanna, & Ryan, 2012).

Sixty years in the making, the Right to Education Act seeks to deliver on the promise to

educate approximately 20 percent of the world’s primary school-aged children. Borne of India’s

desire to provide the opportunity to flourish for all children regardless of caste, gender, or creed,

the Act formulated the practical steps necessary to bring the vision of a common school system

to life. Yet the Act continues to endure considerable scrutiny from those who believe it

multiplied the inefficiencies of government schools. As this study has shown, a deeper analysis

of the factors contributing to public school student achievement declines is warranted. Because

the Act coincided with the flight of less marginalized groups to urban areas and private schools,

socioeconomic status and peer composition declined in rural public schools resulting in a

downward pull on average student achievement. Including these factors in an analysis of student

achievement trends thus paints a clearer picture of how the government school system is doing in

serving India’s children. It is my hope these findings will help sensitize the scholarly community

to the complexities underlying achievement data trends as well as help salvage the noble project

of building an equitable state education system for one-fifth of the world’s children.

Page 36: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

References

ASER Centre. (2013a). Annual Status of Education Report [Data file and codebook].

Unpublished raw data.

ASER Centre. (2013b). Annual Status of Education Report (rural) 2013 – provisional. New

Delhi, India: ASER Centre.

ASER Centre. (2013c). How far has India come in guaranteeing education? The Right to

Education Act & ASER findings 2010-2012. Retrieved from

http://www.unicef.org/india/ASER_6_pager_on_status_of_RTE_implementation_30th_

Mar_2013.pdf

ASER Centre. (n.d.) Sample design of rural ASER. Retrieved from

http://img.asercentre.org/docs/Aser%20survey/Sampling/Sample_Design_of_Rural_ASE

R_1.pdf

Baker, D. P., Goesling, B., & Letendre, G. K. (2002). Socioeconomic status, school quality, and

national economic development: A cross-national analysis of the "Heyneman-Loxley

effect" on mathematics and science achievement. Comparative Education Review, 46(3),

291-312.

Balasubramanian, S. (2014, February 1). Education policy: Where is it in the elections 2014

debate? The Economic Times. Retrieved from

http://articles.economictimes.indiatimes.com/2014-02-01/news/46897971_1_higher-

education-primary-education-system-education-report

Banerjee, A. & Duflo, E. (2013, March 1). Wrong lessons from experts. India Today. Retrieved

from http://indiatoday.intoday.in/story/wrong-lessons-from-experts/1/252218.html

Banerjee, A. V., Cole, S., Duflo, E., & Linden, L. (2007). Remedying education: Evidence from

two randomized experiments in India. The Quarterly Journal of Economics, 122(3),

1235-1264.

Chavan, M. (2012). Uphill battle ahead as outcomes go downhill. Retrieved October 14, 2013,

from http://www.pratham.org/file/Article-by-DrMadhav-Chavan.PDF

Coleman, J. S. (1988). Social capital in the creation of human capital. American Journal of

Sociology, 94, S95–S120.

Coleman, J. S. (1966). Equality of educational opportunity. Washington, DC: Government

Printing Office.

Das, S. K. (2013). India's rights revolution: Has it worked for the poor? New Delhi, India:

Oxford University Press.

Das, J., Dercon, S., Habyarimana, J., Krishnan, P., Muralidharan, K., & Sundararaman, V.

(2011). School inputs, household substitution, and test scores (NBER Working Paper No.

16830). Retrieved from the National Bureau of Economic Research website:

http://www.nber.org/papers/w16830

DeVotta, N. (2010). Understanding contemporary India (2nd ed.). Boulder, CO: Lynne Rienner

Publishers.

Drèze, J., & Sen, A. (2013). An uncertain glory: India and its contradictions. Princeton, NJ:

Princeton University Press.

Duflo, E., Hanna, R., & Rya, S. P. (2012). Incentives work: Getting teachers to come to

school. The American Economic Review, 102(4), 1241-1278.

Education: A billion brains. (2012, September 29). The Economist. Retrieved from

http://www.economist.com/node/21563418

Page 37: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

Ensuring ASER data quality (n.d.). Retrieved February 1, 2014 from

http://www.asercentre.org/p/136.html#fa4qr

Epple, D., & Romano, R. (1998). Competition between private and public schools, vouchers, and

peer-group effects. The American Economic Review, 88(1), 33–62.

Establishment of ASER Centre (n.d.). Retrieved June 28, 2014 from

http://www.asercentre.org/establishment/p/156.html

Fuller, B., & Clarke, P. (1994). Raising school effects while ignoring culture? Local conditions

and the influence of classroom tools, rules, and pedagogy. Review of Educational

Research, 64(1), 119–157.

Gamoran, A., & Long, D. A. (2007). Equality of educational opportunity: A 40 year

retrospective. In R. Teese, S. Lamb, M. Duru-Bellat, & S. Helme (Eds.), International

Studies in Educational Inequality, Theory and Policy (pp. 23–47). Netherlands: Springer.

Garg, P. K. (2013). Fundamental Right of Children to Free and Compulsory Education - its

implementation and future strategies. International Journal of Basic and Advanced

Research. 2(1), 4-7.

Gouda, J., Das, K. C., Goli, S., & Pou, L. M. A. (2013). Government versus private primary

schools in India: An assessment of physical infrastructure, schooling costs and

performance. The International Journal of Sociology and Social Policy, 33(11/12), 708–

724.

Gourishankar, V., & Lokachari, P. S. (2012). Benchmarking educational development

efficiencies of the Indian states: A DEA approach. International Journal of Educational

Management, 26(1), 99–130.

Government of India. (1992). National policy on education, 1986. Retrieved from

http://mhrd.gov.in/sites/upload_files/mhrd/files/NPE-1968.pdf

Hanushek, E. A. (2002). Publicly provided education (NBER Working Paper No. 8799).

Retrieved from the National Bureau of Economic Research website:

http://www.nber.org/papers/w8799

Hanushek, E. A., Kain, J. F., Markman, J. M., & Rivkin, S. G. (2003). Does peer ability affect

student achievement? Journal of Applied Econometrics, 18(5), 527–544.

Hanushek, E. & Woessmann, L. (2007). The role of education quality for economic growth

(World Bank Policy Research Working Paper No. 4122). Retrieved from the World Bank

Group website:

https://openknowledge.worldbank.org/bitstream/handle/10986/7154/wps4122.pdf?sequen

ce=1

Hill, S., & Chalaux, T. (2011). Improving access and quality in the Indian education system

(OECD Economics Department Working Paper No. 885). Retrieved from the

Organisation for Economic Co-operation and Development website: http://www.oecd-

ilibrary.org/content/workingpaper/5kg83k687ng7-en

Hoxby, C. (2000). Peer effects in the classroom: Learning from gender and race variation

(NBER Working Paper No. 7867). Retrieved from the National Bureau of Economic

Research website: http://www.nber.org/papers/w7867

Jenkins, R., & Barr, E. (2006). Social exclusion of scheduled caste children from primary

education in India. Retrieved from

http://www.unicef.org/files/Social_Exclusion_of_Scheduled_Caste_Children_from_Prim

ary_Education_in_India.pdf

Page 38: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

Kaul, V. (2013, July 23). Amartya Sen is right about India’s education system. Firstpost.

Retrieved from http://www.firstpost.com/economy/the-menace-of-rte-amartya-sen-is-

right-about-indias-education-system-978207.html

Kingdon, G. (2007). The progress of school education in India. Oxford Review of Economic

Policy, 23(2), 168–195.

Kingdon, G. (1996). The quality and efficiency of private and public education: A case-study of

urban India. Oxford Bulletin of Economics and Statistics, 58(1), 57–82.

Ladd, H. F. (2002). School vouchers: a critical view. The Journal of Economic Perspectives,

16(4), 3–24.

Law Commission of India. (1998). 165th report on free and compulsory education for children.

Retrieved from http://lawcommissionofindia.nic.in/101-169/Report165.pdf

Mcloughlin, C. (2013). Low cost private schools: Evidence, approaches and emerging issues.

Retrieved from http://r4d.dfid.gov.uk/Output/194091/

Muralidharan, K. (2013). Priorities for primary education policy in India's 12th five year plan. In

India Policy Forum 2013 (pp. 1-46). Retrieved from

http://econweb.ucsd.edu/~kamurali/papers/Published%20Articles/Priorities%20for%20Pr

imary%20Education%20Policy%20in%20India.pdf

Muralidharan, K., & Kremer, M. (2009). Public and private schools in rural India. In P. Peterson

& R. Chakrabarti (Eds.), School Choice International: Exploring public-private

partnerships. Cambridge, MA: MIT Press.

Murnane, R. J, & Willett, J. B. (2011). Methods matter: Improving causal inference in

educational and social science research. New York, NY: Oxford University Press.

Nanda, P. (2014, January 15). Education quality worsens under UPA: ASER. Live Mint & The

Wall Street Journal. Retrieved from

http://www.livemint.com/Politics/db5Bfb8qSTFhO70WEafBpJ/Indias-education-quality-

has-failed-to-improve-finds-ASER.html

National Council of Educational Research and Training. (1966). Report of the education

commission, 1964-66. Retrieved from

http://www.teindia.nic.in/Files/Reports/CCR/KC/KC_V1.pdf

Right of Children to Free and Compulsory Education Act 2009, no. 35 (India 2009).

Rothstein, R. (2004). Class and schools: Using social, economic, and educational reform to

close the black-white achievement gap. New York, NY: Columbia Teachers College.

Sacerdote, B. (2001). Peer effects with random assignment: Results for Dartmouth roommates.

The Quarterly Journal of Economics, 116(2), 681–704.

Seshadri, C. (1976). Equality of educational opportunity: Some issues in Indian education.

Comparative Education, 12(3), 219–230.

Somers, M.A., McEwan, P. J., & Williams, J. D. (2004). How effective are private schools in

Latin America? Comparative Education Review, 48(1), 48.

Srivastava, P., & Noronha, C. (2014). Early private school responses to India’s Right to

Education Act: Implications for equity. In I. Macpherson, S. Robertson, & G. Walford

(Eds.), Education, privatization and social justice: case studies from Africa, South Asia

and South East Asia (pp.179-198). Oxford, United Kingdom: Symposium Books.

Stock, J. H., & Watson, M. W. (2011). Introduction to econometrics (3rd ed.). Boston, MA:

Addison-Wesley.

Summers, A. A., & Wolfe, B. L. (1977). Do schools make a difference? American Economic

Review, 67(4), 639–52.

Page 39: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

Tooley, J., Dixon, P., & Gomathi, S. V. (2007). Private schools and the Millennium

Development Goal of universal primary education: A census and comparative survey in

Hyderabad, India. Oxford Review of Education, 33(5), 539–560.

Tulloch, J. (2011, October 18). India’s massive migration crisis. Open Knowledge. Retrieved

from http://knowledge.allianz.com/demography/migration/?454/indias-massive-

migration-crisis

UNESCO-IBE (2011). India: World data on education, 2010/11. Retrieved from

http://www.ibe.unesco.org/fileadmin/user_upload/Publications/WDE/2010/pdf-

versions/India.pdf

UNICEF and the UNESCO Institute for Statistics (2012). Global initiative on out-of-school

children. Retrieved from

http://www.unicef.org/education/files/OOSCI_flyer_EN_Sept.pdf

USAID (2008). Opportunity to learn: A high impact strategy for improving educational

outcomes in developing countries. Retrieved from

http://toolkit.ineesite.org/toolkit/INEEcms/uploads/1038/Opportunity_To_Learn.pdf

Vagh, Shaher Banu (n.d.). Validating the ASER testing tools: Comparisons with reading fluency

measures and the Read India measures. Retrieved from

http://img.asercentre.org/docs/Aser%20survey/Tools%20validating_the_aser_testing_too

ls__oct_2012__2.pdf

Wadhwa, W. (2009). Are private schools really performing better than government schools?

Retrieved from http://images2.asercentre.org/Resources/Articles/Pvt_Sch_Final.pdf

World Bank Group (2011). Learning for all: Investing in people's knowledge and skills to

promote development. Retrieved from

http://siteresources.worldbank.org/EDUCATION/Resources/ESSU/Education_Strategy_

4_12_2011.pdf

Zimmer, R., & Toma, E. (2000). Peer effects in private and public schools across countries.

Journal of Policy Analysis and Management, 19(1), 75–92.

Page 40: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

Appendix I

Figure 1: Primary and secondary combined enrollment rate in rural areas. India, 2006-2013.

Source: ASER Centre (2013a).

0.89

0.9

0.91

0.92

0.93

0.94

0.95

0.96

2006 2007 2008 2009 2010 2011 2012 2013

En

roll

men

t ra

te

Year

Page 41: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

Figure 2: Public versus private school enrollment shares in rural areas. India, 2006-2012. Source:

ASER Centre (2013a).

0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

2006 2007 2008 2009 2010 2011 2012

Per

cen

tag

e o

f to

tal

enro

llm

ents

Year

Government

Private

Other

Post RTE

Page 42: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

Figure 3: Concept map: education production function model including socioeconomic status, peer quality, and school inputs for

public and private schooling in post-Right to Education Act India

Page 43: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

Table 1: Summary and descriptive statistics – all rural India, 6-14 year olds, 2009-2012

2009 2010 2011 2012

School status % n % n % n % n

Enrolled 96.30 453,994 96.77 435,081 96.87 420,213 96.69 390,961

Government schools 75.61 343,254 74.32 323,340 72.00 302,578 68.89 269,335

Girls 46.41 154,990 46.88 151,329 48.27 145,361 49.47 131,392

Boys 53.59 178,947 53.12 171,471 51.73 155,796 50.53 134,219

Private schools 23.32 105,857 24.76 107,722 27.00 113,464 30.03 117,392

Girls 42.09 43,224 42.03 45,232 42.38 47,880 42.15 49,105

Boys 57.91 59,482 57.97 62,391 57.62 65,110 57.85 67,408

Other schools 1.08 4,883 0.92 4,019 0.99 4,171 1.08 4,234

Out of school 3.70 17,463 3.23 14,527 3.13 13,556 3.31 13,393

Dropped out 51.93 9,068 56.75 8,244 53.58 7,263 57.40 7,688

Never enrolled 48.07 8,395 43.25 6,283 46.42 6,293 42.59 5,705

Socioeconomic status

Mother has some education 49.23 223,042 52.80 233,718 51.16 216,096 50.18 200,145

Government 46.57 151,390 50.06 156,999 47.53 137,972 45.79 119,300

Private 62.44 62,510 65.80 69,342 64.85 71,270 64.52 73,906

Household has television 47.79 223,749 48.08 214,919 48.01 206,085 47.55 192,136

Government 43.83 146,847 43.97 139,348 42.85 126,501 41.42 109,570

Private 65.23 67,943 65.01 68,911 65.65 73,220 65.89 76,017

Household has toilet 40.45 186,579 41.08 184,730 40.92 177,266 39.71 161,635

Government 35.78 118,110 36.29 115,686 34.90 104,014 32.42 86,360

Private 58.84 60,371 58.68 62,527 59.41 66,754 58.94 68,478

Student performance

Math proficiency (grades 2-5) 67.24 156,404 65.13 145,369 58.13 119,132 54.49 96,152

Government schools 64.99 118,404 62.56 107,498 53.77 81,443 48.09 59,817

Private schools 75.85 36,280 74.19 36,529 71.17 36,704 70.48 35,512

Reading proficiency (grades 2-3) 21.76 25,446 20.68 23,423 20.18 21,181 21.69 19,480

Government schools 18.60 16,845 17.27 14,912 15.36 11,715 15.55 9,663

Private schools 33.39 8,255 32.38 8,278 33.68 9,294 36.37 9,710

Page 44: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

Table 2: Summary and descriptive statistics – 2-5 graders in govt. schools, 2009-2012

2009 2010 2011 2012

mean n mean n mean n mean n

Math proficiency

(grades 2-5)

0.66 127,769 0.64 106,023 0.54 182,274 0.48 80,143

Reading proficiency

(grades 2-3)

0.19 12,089 0.16 8,828 0.15 13,760 0.15 5,995

Student characteristics

Mother's years of school 2.99 123,121 3.33 104,706 3.12 185,148 3.05 84,831

Household has television 0.40 127,612 0.43 107,334 0.39 189,080 0.38 86,568

Household has toilet 0.32 125,627 0.34 107,882 0.31 190,936 0.29 87,294

Child’s grade 3.49 130,367 3.48 109,233 3.48 193,568 3.49 88,175

Child is male 0.54 126,934 0.53 108,978 0.52 192,632 0.51 86,999

Peer characteristics

Peer mothers’ class 2.98 122,037 3.32 103,489 3.12 185,148 3.03 83,421

Peers own television 0.40 126,552 0.43 106,166 0.39 189,080 0.38 85,198

Peers own toilet 0.32 124,572 0.34 106,698 0.31 190,936 0.29 85,945

School characteristics

School has a toilet 0.88 121,864 0.90 101,360 0.89 181,462 0.93 85,574

School has water 0.91 124,724 0.91 104,602 0.92 185,304 0.92 87,070

Midday meals served 0.82 121,304 0.82 103,758 0.85 184,328 0.86 86,665

Boundary wall 0.51 123,287 0.48 103,964 0.51 183,448 0.52 86,744

Student-teacher ratio 56.25 106,037 52.52 92,209 56.73 165,746 49.86 80,055

Teacher attendance rate 0.88 90,286 0.85 80,702 0.85 134,104 0.85 67,437

Page 45: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

Table 3a: Logistic regression: math proficiency by student-, peer-, and school-level

characteristics (odds ratio) – 2-5 graders in govt. schools, 2009-2012

Dependent variable:

math proficiency

Student

SES

Peer

SES

School

Inputs

Student +

Peer SES

All

(1) (2) (3) (4) (5)

Student characteristics

Mother has schooling 1.712*** 1.392*** 1.391***

(0.0124) (0.0117) (0.0145)

Household has television 1.300*** 1.145*** 1.145***

(0.0102) (0.0111) (0.0135)

Household has toilet 1.667*** 1.280*** 1.292***

(0.0137) (0.0137) (0.0169)

Grade 5 0.0411*** 0.0392*** 0.0360***

(0.000576) (0.000558) (0.000650)

Grade 4 0.117*** 0.113*** 0.105***

(0.00143) (0.00139) (0.00165)

Grade 3 0.288*** 0.282*** 0.276***

(0.00320) (0.00317) (0.00394)

Child age 1.109*** 1.112*** 1.107***

(0.00284) (0.00288) (0.00368)

Female 0.876*** 0.879*** 0.884***

(0.00601) (0.00608) (0.00762)

Peer characteristics

Mothers have schooling 1.962*** 1.778*** 1.758***

(0.0210) (0.0246) (0.0300)

Peers have television 1.038*** 1.167*** 1.179***

(0.0109) (0.0168) (0.0211)

Peers have toilet 1.886*** 1.555*** 1.583***

(0.0199) (0.0235) (0.0290)

School characteristics

Toilet 1.082*** 0.942***

(0.0142) (0.0148)

Drinking water 0.897*** 0.967**

(0.0124) (0.0160)

Midday meal 0.870*** 0.906***

(0.00912) (0.0115)

Student-teacher ratio 0.998*** 0.999***

(9.24e-05) (0.000101)

Boundary wall 1.042*** 1.009

(0.00772) (0.00899)

Teacher attendance 1.296*** 1.244***

(0.0188) (0.0216)

Page 46: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

Year fixed effect

2012 0.398*** 0.467*** 0.462*** 0.394*** 0.391***

(0.00437) (0.00456) (0.00523) (0.00438) (0.00535)

2011 0.530*** 0.589*** 0.575*** 0.522*** 0.504***

(0.00477) (0.00473) (0.00556) (0.00474) (0.00587)

2010 0.837*** 0.846*** 0.874*** 0.814*** 0.803***

(0.00856) (0.00780) (0.00955) (0.00842) (0.0104)

Constant 3.050*** 1.200*** 2.093*** 2.444*** 2.728***

(0.0688) (0.00921) (0.0473) (0.0565) (0.108)

Observations 458,073 460,280 323,293 454,637 298,125

Standard error in parentheses

*** p<0.01, ** p<0.05, * p<0.1

Page 47: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

Table 3b: Logistic regression: reading proficiency by student-, peer-, and school-level

characteristics (odds ratio) – 2-5 graders in govt. schools, 2009-2012

Dependent variable:

reading proficiency

Student

SES

Peer

SES

School

Inputs

Student +

Peer SES

All

(1) (2) (3) (4) (5)

Student characteristics

Mother has schooling 1.848*** 1.464*** 1.457***

(0.0228) (0.0210) (0.0257)

Household has television 1.382*** 1.212*** 1.217***

(0.0177) (0.0191) (0.0232)

Household has toilet 1.392*** 1.241*** 1.250***

(0.0180) (0.0213) (0.0259)

Grade 3 0.753*** 0.743*** 0.742***

(0.00930) (0.00927) (0.0114)

Child age 1.115*** 1.121*** 1.119***

(0.00475) (0.00486) (0.00616)

Female 0.991 0.994 1.016

(0.0114) (0.0115) (0.0144)

Peer characteristics

Mothers have schooling 2.578*** 1.926*** 1.874***

(0.0507) (0.0439) (0.0523)

Peers have television 1.378*** 1.231*** 1.113***

(0.0261) (0.0289) (0.0321)

Peers have toilet 1.381*** 1.083*** 1.034

(0.0250) (0.0258) (0.0297)

School characteristics

Toilet 1.212*** 1.074***

(0.0312) (0.0296)

Drinking water 1.032 1.090***

(0.0261) (0.0295)

Midday meal 1.162*** 1.162***

(0.0228) (0.0247)

Student-teacher ratio 0.996*** 0.999***

(0.000213) (0.000199)

Boundary wall 1.082*** 1.033**

(0.0148) (0.0151)

Teacher attendance 1.414*** 1.281***

(0.0411) (0.0396)

Year fixed effect

2012 0.757*** 0.748*** 0.719*** 0.750*** 0.720***

(0.0139) (0.0137) (0.0149) (0.0139) (0.0159)

Page 48: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

2011 0.737*** 0.729*** 0.684*** 0.727*** 0.676***

(0.0108) (0.0106) (0.0118) (0.0107) (0.0124)

2010 0.792*** 0.776*** 0.829*** 0.773*** 0.781***

(0.0130) (0.0127) (0.0156) (0.0128) (0.0156)

Constant 0.0630*** 0.115*** 0.155*** 0.0489*** 0.0349***

(0.00232) (0.00166) (0.00673) (0.00187) (0.00233)

Observations 234,105 235,296 163,279 232,425 150,758

Standard error in parentheses

*** p<0.01, ** p<0.05, * p<0.1

Page 49: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

Table 4: Logistic regression: math and reading proficiency by student-level characteristics

interacted with year (odds ratio) – 2-5 graders in govt. schools

Math proficiency Reading proficiency

Mother has schooling 1.287*** 1.269***

(0.0254) (0.0377)

Household has television 1.017 1.149***

(0.0215) (0.0363)

Household has toilet 1.093*** 1.120***

(0.0251) (0.0375)

Grade 5 0.0357***

(0.000646)

Grade 4 0.104***

(0.00163)

Grade 3 0.274*** 0.743***

(0.00393) (0.0114)

Child age 1.107*** 1.119***

(0.00369) (0.00617)

Female 0.883*** 1.015

(0.00763) (0.0144)

Mother has schooling x 2012 1.206*** 1.491***

(0.0343) (0.0713)

Mother has schooling x 2011 1.111*** 1.170***

(0.0270) (0.0459)

Mother has schooling x 2010 1.022 1.109**

(0.0277) (0.0473)

Household has television x 2012 1.180*** 1.183***

(0.0354) (0.0571)

Household has television x 2011 1.263*** 1.106**

(0.0325) (0.0448)

Household has television x 2010 1.040 0.999

(0.0300) (0.0441)

Household has toilet x 2012 1.258*** 1.234***

(0.0399) (0.0603)

Household has toilet x 2011 1.299*** 1.189***

(0.0354) (0.0489)

Household has toilet x 2010 1.152*** 1.091*

(0.0350) (0.0488)

2012 0.312*** 0.478***

(0.00650) (0.0191)

2011 0.402*** 0.548***

(0.00710) (0.0175)

2010 0.754*** 0.717***

(0.0150) (0.0250)

Page 50: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

Peer controls

Mothers have schooling 1.752*** 1.865***

(0.0300) (0.0522)

Peers have television 1.163*** 1.102***

(0.0208) (0.0319)

Peers have toilet 1.576*** 1.033

(0.0289) (0.0298)

School controls

Toilet 0.942*** 1.076***

(0.0149) (0.0296)

Drinking water 0.970* 1.094***

(0.0161) (0.0296)

Midday meal 0.906*** 1.165***

(0.0115) (0.0247)

Student-teacher ratio 0.999*** 0.999***

(0.000101) (0.000199)

Boundary wall 1.012 1.035**

(0.00903) (0.0152)

Teacher attendance 1.243*** 1.278***

(0.0216) (0.0396)

Constant 3.148*** 0.0405***

(0.129) (0.00277)

Observations 298,125 150,758

Standard error in parentheses

*** p<0.01, ** p<0.05, * p<0.1

Page 51: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

Table 5: Math and reading proficiency for high and low SES students in government schools, 2009-2012

2009 2010 2011 2012

% n % n % n % n

High SES Students

Math proficiency 75.57 17,313 74.28 18,427 72.58 26,446 65.69 10,564

Grade 2 94.91 3,952 95.37 4,304 94.64 6,636 93.38 2,495

Grade 3 86.59 4,430 85.46 4,834 84.17 6,848 79.31 2,659

Grade 4 73.16 4,612 72.90 4,657 68.56 6,806 57.95 2,830

Grade 5 49.13 4,319 44.38 4,632 40.35 6,156 33.37 2,580

Reading proficiency 28.14 8,403 24.49 9,193 25.71 13,566 30.05 5,168

Grade 2 30.88 3,966 28.46 4,333 28.37 6,696 29.75 2,501

Grade 3 25.69 4,437 20.96 4,860 23.11 6,870 30.33 2,667

Low SES Students

Math proficiency 60.43 40,679 55.72 31,443 43.24 59,498 37.21 26,853

Grade 2 85.20 10,643 83.09 8,200 73.90 15,288 70.03 6,811

Grade 3 67.26 10,761 63.76 8,190 48.59 15,626 39.94 6,922

Grade 4 52.16 9,518 45.36 7,364 31.72 13,932 23.95 6,600

Grade 5 33.94 9,757 27.88 7,689 16.47 14,652 13.42 6,520

Reading proficiency 13.60 21,483 9.92 16,489 8.16 31,176 6.85 13,757

Grade 2 14.67 10,683 11.06 8,253 8.41 15,434 6.52 6,820

Grade 3 12.55 10,800 8.77 8,236 7.91 15,742 7.16 6,937

Page 52: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

Table 6: OLS: district-level math and reading proficiency differences between 2012 and 2009

regressed on SES variable difference –2-5 graders in govt. schools

Math proficiency Reading proficiency

Mothers have schooling (2012 – 2009) 0.225*** 0.0401

(0.0609) (0.0488)

Households have television (2012 – 2009) 0.229*** 0.0611

(0.0590) (0.0473)

Households have toilet (2012 – 2009) -0.0732 0.0439

(0.0600) (0.0480)

Constant -0.156*** -0.0272***

(0.00798) (0.00639)

Observations 560 560

R-squared 0.071 0.014

Adjusted R-squared 0.066 0.009

Standard errors in parentheses

*** p<0.01, ** p<0.05, * p<0.1

Page 53: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

Table 7: Logistic regression: math and reading proficiency by peer-level characteristics

interacted with year (odds ratio) – 2-5 graders in govt. schools

Math proficiency Reading proficiency

Peer mothers have schooling 1.691*** 1.616***

(0.0559) (0.0789)

Peers have television 0.966 1.061

(0.0323) (0.0523)

Peers have toilet 1.217*** 0.882***

(0.0404) (0.0421)

Peer mothers have schooling x 2012 1.064 1.341***

(0.0495) (0.101)

Peer mothers have schooling x 2011 1.060 1.145**

(0.0427) (0.0721)

Peer mothers have schooling x 2010 1.012 1.244***

(0.0455) (0.0854)

Peers have television x 2012 1.254*** 1.161**

(0.0564) (0.0830)

Peers have television x 2011 1.482*** 1.158**

(0.0585) (0.0705)

Peers have television x 2010 1.054 0.884*

(0.0470) (0.0596)

Peers have toilet x 2012 1.358*** 1.376***

(0.0605) (0.0932)

Peers have toilet x 2011 1.543*** 1.251***

(0.0601) (0.0726)

Peers have toilet x 2010 1.248*** 1.152**

(0.0543) (0.0730)

2012 0.314*** 0.505***

(0.00801) (0.0234)

2011 0.362*** 0.537***

(0.00794) (0.0205)

2010 0.734*** 0.700***

(0.0183) (0.0298)

Student controls

Mother has schooling 1.391*** 1.459***

(0.0146) (0.0258)

Household has television 1.140*** 1.214***

(0.0134) (0.0231)

Household has toilet 1.287*** 1.251***

(0.0168) (0.0259)

Grade 5 0.0357***

(0.000645)

Page 54: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

Grade 4 0.104***

(0.00163)

Grade 3 0.274*** 0.743***

(0.00393) (0.0114)

Child age 1.107*** 1.119***

(0.00369) (0.00617)

Female 0.884*** 1.016

(0.00763) (0.0144)

School controls

Toilet 0.942*** 1.076***

(0.0149) (0.0296)

Drinking water 0.972* 1.097***

(0.0162) (0.0297)

Midday meal 0.903*** 1.164***

(0.0115) (0.0248)

Student-teacher ratio 0.999*** 0.999***

(0.000101) (0.000199)

Boundary wall 1.012 1.035**

(0.00903) (0.0152)

Teacher attendance 1.241*** 1.276***

(0.0216) (0.0395)

Constant 3.283*** 0.0406***

(0.138) (0.00283)

Observations 298,125 150,758

Standard error in parentheses

*** p<0.01, ** p<0.05, * p<0.1

Page 55: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

Table 8: Logistic regression: math and reading proficiency by school-level characteristics

interacted with year (odds ratio) – 2-5 graders in govt. schools

Math proficiency Reading proficiency

School characteristics

School has toilet 0.742*** 0.817***

(0.0238) (0.0371)

School has drinking water 0.901*** 1.057

(0.0314) (0.0522)

School serves midday meal 1.010 1.279***

(0.0253) (0.0497)

Student-teacher ratio 1.001*** 1.000

(0.000212) (0.000281)

School has boundary wall 0.847*** 0.918***

(0.0160) (0.0258)

Teacher attendance 1.030 1.357***

(0.0442) (0.0904)

School has toilet x 2012 1.257*** 1.138

(0.0683) (0.102)

School has toilet x 2011 1.473*** 1.635***

(0.0599) (0.111)

School has toilet x 2010 1.324*** 1.667***

(0.0607) (0.128)

School has drinking water x 2012 0.945 0.883

(0.0491) (0.0705)

School has drinking water x 2011 1.217*** 1.094

(0.0537) (0.0752)

School has drinking water x 2010 1.050 1.173**

(0.0529) (0.0924)

School serves midday meal x 2012 0.743*** 0.791***

(0.0291) (0.0508)

School serves midday meal x 2011 0.773*** 0.767***

(0.0256) (0.0413)

School serves midday meal x 2010 1.088** 1.056

(0.0388) (0.0617)

Student-teacher ratio x 2012 0.998*** 0.997***

(0.000344) (0.000630)

Student-teacher ratio x 2011 0.998*** 0.998***

(0.000258) (0.000430)

Student-teacher ratio x 2010 0.998*** 0.997***

(0.000327) (0.000557)

School has boundary wall x 2012 1.190*** 1.187***

(0.0331) (0.0537)

School has boundary wall x 2011 1.367*** 1.156***

Page 56: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

(0.0325) (0.0438)

School has boundary wall x 2010 1.138*** 1.211***

(0.0302) (0.0498)

Teacher attendance x 2012 1.158** 0.880

(0.0668) (0.0860)

Teacher attendance x 2011 1.194*** 0.902

(0.0609) (0.0755)

Teacher attendance x 2010 1.412*** 0.976

(0.0789) (0.0907)

2012 0.382*** 1.019

(0.0330) (0.142)

2011 0.291*** 0.546***

(0.0208) (0.0622)

2010 0.423*** 0.430***

(0.0330) (0.0549)

Student controls

Mother has schooling 1.391*** 1.456***

(0.0146) (0.0257)

Household has television 1.145*** 1.217***

(0.0135) (0.0232)

Household has toilet 1.291*** 1.250***

(0.0169) (0.0259)

Grade 5 0.0358***

(0.000648)

Grade 4 0.105***

(0.00164)

Grade 3 0.275*** 0.742***

(0.00393) (0.0114)

Child age 1.106*** 1.118***

(0.00369) (0.00616)

Female 0.883*** 1.015

(0.00762) (0.0144)

Peer controls

Peer mothers have schooling 1.754*** 1.865***

(0.0300) (0.0522)

Peers have television 1.180*** 1.110***

(0.0211) (0.0321)

Peers have toilet 1.580*** 1.026

(0.0290) (0.0295)

Constant 4.038*** 0.0408***

(0.261) (0.00396)

Observations 298,125 150,758

Standard error in parentheses

*** p<0.01, ** p<0.05, * p<0.1

Page 57: A RIGHT WORTH FIGHTING FOR? UNDERSTANDING...A Right Worth Fighting For? Understanding Declines in Student Achievement Following India’s Right to Education Act Erika V. Keaveney July

Appendix II

Table 1: Logistic regression: math and reading proficiency by grade (odds ratio) – govt. schools,

2009-2012

Math proficiency Reading proficiency

School class

Grade 5 0.0696***

(0.000714)

Grade 4 0.170***

(0.00168)

Grade 3 0.346*** 0.857***

(0.00347) (0.00929)

Constant 5.937*** 0.206***

(0.0480) (0.00155)

Observations 496,209 253,472

Standard error in parentheses

*** p<0.01, ** p<0.05, * p<0.1

Table 2: Logistic regression: math and reading proficiency by age-appropriate grade status

interacted with year (odds ratio) - 2-5 graders in govt. schools

Math proficiency Reading proficiency

Age-appropriate grade x 2012 0.961** 0.937*

(0.0179) (0.0327)

Age-appropriate grade x 2011 1.047*** 0.891***

(0.0160) (0.0245)

Age-appropriate grade x 2010 0.978 0.925**

(0.0172) (0.0287)

Age-appropriate grade 0.873*** 0.909***

(0.0104) (0.0185)

2012 0.486*** 0.794***

(0.00688) (0.0208)

2011 0.586*** 0.801***

(0.00669) (0.0163)

2010 0.910*** 0.890***

(0.0121) (0.0206)

Constant 2.110*** 0.241***

(0.0188) (0.00363)

Observations 495,170 252,939

Standard error in parentheses

*** p<0.01, ** p<0.05, * p<0.1