Do Street‐Level Bureaucrats Discriminate Based on Religion? A Large‐Scale Correspondence Experiment among American Public School Principals

• Published date: 01 March 2021
• Author: Steven Pfaff,Charles Crabtree,Holger L. Kern,John B. Holbein
• Date: 01 March 2021
• DOI: http://doi.org/10.1111/puar.13235

244 Public Administration Review • March | Apri l 202 1

Abstract: Although public administration scholars have long studied discrimination on the basis of race/ethnicity,

class, and gender, little to no research exists on whether street-level bureaucrats provide differential services based on the

religious identity of their constituents. This article reports the results from a large-scale correspondence study of street-

level bureaucrats in the American public school system. The authors emailed the principals of a large sample of public

schools and asked for a meeting, randomly assigning the religious (non)affiliation of the family. To get at potential

causal mechanisms, religious belief intensity was also randomly assigned. The findings show evidence of substantial

discrimination against Muslims and atheists on a par with, and sometimes larger than, the racial discrimination

found in previous studies. These individuals are substantially less likely to receive a response, with discrimination

growing when they signal that their beliefs are more intense. Protestants and Catholics face no discrimination unless

they signal that their religious beliefs are intense.

Evidence for Practice

• Although legal protections and social norms are in place to protect public school students from religious

discrimination, religious discrimination is large and widespread in American public schools. Groups such as

Muslims and atheists face substantial barriers to equal treatment in the public domain.

• Public school officials and other policy makers should prioritize testing and implementing new policies and

practices that protect both religious believers and nonbelievers. These policies should be designed with the

understanding that unequal treatment is driven, in part, by differences in perceived costs attached to students

from social minority groups.

• In a context of heightened scrutiny toward social out-groups at all levels of government, public schools play a

vitally important role in ensuring that social minorities actually receive the equal treatment they are afforded

under the law.

Street-level bureaucrats (SLBs) have substantial

discretion in their actions as “ultimate

policymakers” (Lipsky1971, 1980, 2010; see

also Brodkin2011a, 2011b; Maynard-Moody and

Musheno2012; Pressman and Wildavsky1984;

Sowa and Selden2003).1 Despite being a core

focus of public administration research over the

past four decades (e.g., Brodkin2012; Jensen and

Vestergaard2016; Jilke and Tummers2018; Jilke,

Van Dooren, and Rys2018; Keiser2010; Brodkin

and Majmundar2010; Maupin1993; Monnat2010;

Moynihan and Herd2010; Prottas1979;

Riccucci2005a, 2005b; Tummers et al.2015;

Wenger and Wilkins2009), whether SLBs choose

to use their discretion in ways that aide or impede

disadvantaged clients remains a vitally important

open question.2 Indeed, there are many important

gaps in this literature (Brodkin2011a, 2011b,

2012). For example, most of the literature on the

decision-making processes of SLBs has focused on

three client characteristics: race (e.g., Adida, Laitin,

and Valfort2010; Butler and Broockman2011;

Costa2017; Einstein and Glick2017), gender (e.g.,

Kalla, Rosenbluth, and Teele2018), and class (e.g.,

Carnes and Holbein2018). The search for bias

induced by these characteristics has yielded many

valuable insights into the drivers and consequences of

bureaucratic action (e.g., Dubois2016; Harrits2019;

Herd, Mitchell, and Lightman2005; Riccucci and

Saidel1997; Watkins-Hayes2011). At the same

time, the relatively narrow focus has restricted

our understanding of the breadth of bureaucratic

discretion’s influence. As a result, we are left with a

limited understanding of the extent of discriminatory

practices among SLBs.

In addition, despite being a core area of observational

research, comparatively little experimental research

has been conducted on the extent, nature, and

mechanisms of potential biases among frontline

Steven Pfaff

Charles Crabtree

Holger L. Kern

John B. Holbein

Do Street-Level Bureaucrats Discriminate Based on Religion?

A Large-Scale Correspondence Experiment among American

Public School Principals

University of Washington

Dartmouth College

Florida State University

University of Virginia

Research Article

John B. Holbein is assistant professor

of public policy, politics, and education in

the Frank Batten School of Leadership and

Public Policy at the University of Virginia.

He studies political participation, political

inequality, democratic accountability,

political representation, and education

policy.

Email: holbein@virginia.edu

Holger L. Kern is associate professor of

political science at Florida State University.

His main research areas are comparative

politics and quantitative methodology. He

studies the politics of authoritarian regimes

with a focus on transitions to and from

democracy, populism, mass media, public

opinion, and information control.

Email: hkern@fsu.edu

Charles Crabtree is assistant professor

of government at Dartmouth College. His

research focuses on fairness in politics, with

applications to several areas, including the

study of repression, human rights, policing,

and immigration.

Email: crabtree@dartmouth.edu

Steven Pfaff is professor of sociology

at the University of Washington. He is an

expert in the sociology of religion and

has published on topics such as collective

action, Eastern European historical

sociology, and social movements.

Email: pfaff@uw.edu

Public Administration Review,

Vol. 81, Iss. 2, pp. 244–259. © 2020 by

The American Society for Public Administration.

DOI: 10.1111/puar.13235.

Do Street-Level Bureaucrats Discriminate Based on Religion? 245

bureaucrats.3 This is unfortunate since observational research often

struggles to fully address concerns about endogeneity and reverse

causation (Jilke, Van de Walle, and Kim2016; Van de Walle and

Bouckaert2003) and to identify causal mechanisms (Bullock,

Green, and Ha2010). Finally, when taken as a whole, research in

this area suggests that when legal institutions (e.g., Brodkin2012),

professional norms (e.g., Weissert1994), established managerial

hierarchies (e.g., Keiser2010; Maynard-Moody and Portillo2010;

Scott1997), and active public oversight (e.g., Holbein2016;

Holbein and Hassell2018; Moynihan2008; Moynihan and

Ingraham2003; Moynihan, Pandey, and Wright2011) are present,

equitable representation is more likely to follow. However, most

studies have examined the role of each of these factors in isolation;

as a result it remains unclear is whether discriminatory behavior can

still occur even when all of these factors are present.

In this article, we begin to fill these gaps. To do so, we conducted

a large-scale correspondence experiment that tested whether

SLBs discriminate on the basis of religion—a core individual

characteristic that plays a key role in modern American society

(as we describe later) but has yet to be explored thoroughly in the

representative bureaucracy literature.4 We focus on public schools

in the United States, for three reasons. First, this is a social context

in which all of the forces thought to promote equal treatment are

present. In public schools, legal protections, professional norms

of equitable treatment, established managerial hierarchies, and

active public oversight are present to a degree not often seen in

other contexts.5,6 Second, public schools play a key role in local

communities, where they serve as one of the most common touch

points between citizens and their government (Holbein2016;

Holbein and Hassell2018; McDonnell2013; Soss1999; Soss and

Schram2007). Third, as we describe later, public schools are at the

center of fundamental debates about how the state and religion can,

do, and should interact. Studying the behavior of SLBs in public

schools thus allows us to expand our understanding of whether

the factors thought to promote equitable treatment are actually

sufficient to do so in this vitally important policy arena.

In our experiment, we emailed the principals of more than

45,000 public PK–12 schools in 33U.S. states. In our emails, we

manipulated the identity of a parent who was considering enrolling

his or her child in that school and asked for a meeting with the

principal. We randomly assigned the religious affiliation/nonaffiliation

of the family (no information given, Protestant, Catholic, Muslim, or

atheist). We held the names of the parent and child constant so that

we could separate the effects of race/ethnicity and class—which have

been studied previously in other contexts—from the effect of religious

affiliation. To go an extra step beyond previous correspondence

experiments, we explored a potential causal mechanism by also

randomizing the intensity of the signaled beliefs: low (identification

only), medium (identification + compatibility inquiry), or high

(identification + accommodation request). This allows us to

experimentally explore a key mechanism that might be driving any

discriminatory effect: the perceived costs attached to the enrollment of

religious adherents or atheists. We then observed whether principals

replied to our email. This constitutes our outcome of interest.

Compared with baseline emails, which provide no information

about religious background, we found high levels of discrimination

against Muslims and atheists. We found that Muslim and atheist

parents are discriminated against for merely revealing their beliefs

in the signature part of their emails. Signaling membership in

these groups decreased the probability of a reply by 4.6 and 4.7

percentage points, respectively. This difference is statistically and

substantively meaningful; it is only slightly smaller than (but

not statistically distinct from) the discriminatory effects of race/

ethnicity shown in previous correspondence studies (Butler and

Broockman2011). For Muslims and atheists, discriminatory effects

are present regardless of whether enrollment costs are explicitly

signaled. Moreover, discrimination against Muslim and atheist

parents increases dramatically if they inquire about the compatibility

of the school with their beliefs or ask for religious accommodations,

with such signals reducing response rates by 8.7 and 13.8 percentage

points, respectively.

These strikingly lower response rates for higher levels of request

intensity suggest that an important mechanism behind the

discriminatory effects we find is the degree to which these

individuals’ belief systems are seen to be imposing a cost on public

officials. Response rates for Protestant and Catholic parents are

indistinguishable from the no information baseline in the low-

intensity condition; discrimination only appears when parents

inquire about the compatibility of the school with their beliefs

or ask for accommodation of their beliefs. Finally, exploiting

our purposefully large sample size, we show that discriminatory

effects are systemic in the public education system. Many tests for

treatment effect heterogeneity are woefully underpowered (Blair

et al.2018; Fink, McConnell, and Vollmer2014), but our large

sample size allows us to show with a great degree of precision that

discrimination is remarkably consistent across the racial/ethnic

composition of the school, the school type (primary, middle, or

high), the median household income/poverty rates, the share of

adults holding a bachelor’s degree, Republican vote shares in the

2012 presidential elections, and the religious adherence rates of

the surrounding community. Discrimination against citizens with

nonmainstream beliefs about religion seems to be widespread in the

American public school system.

Although research on SLBs has proliferated since Lipsky’s seminal

research on this topic (Lipsky1971, 1980, 2010), important

questions remain unanswered. Our article contributes in at least

four important ways to this core area of public administration

research. First, it improves our knowledge about the scope of

inequitable treatment in the public domain. A large majority of

studies on discriminatory behavior by SLBs focus narrowly on race,

class, and gender (Maynard-Moody and Portillo2010). While these

social dimensions are vitally important, until now, the literature has

(largely) ignored clients’ religious affiliation. In performing their

duties, the race, gender, and class of the client are not the only—

and perhaps not even the most important—heuristics that SLBs use

in deciding whether to respond to a request for help. As such, our

research shows that public administration scholars would do well to

expand the list of potential social cleavages reinforced by unequal

treatment in the public domain.

Second, our research constitutes an important step forward in

how we study the potential biases of SLBs. While a vast literature

in public administration has studied the roots, nature, and extent