The limited impact of reference groups' symbolic gender representation on willingness to coproduce

• Published date: 01 May 2023
• Author: Martin Sievert
• Date: 01 May 2023
• DOI: http://doi.org/10.1111/puar.13619

RESEARCH ARTICLE

The limited impact of reference groups’symbolic gender

representation on willingness to coproduce

Martin Sievert

1,2

1

Institute of Public Administration, Leiden

University, Den Haag, Netherlands

2

Business School, University of Mannheim,

Mannheim, Baden-Württemberg, Germany

Correspondence

Martin Sievert, Leiden University, Institute of

Public Administration, Turfmarkt 99, 2511 DC

Den Haag, Netherlands.

Email: m.c.g.sievert@fgga.leidenuniv.nl

Funding information

Joachim Herz Foundation

Abstract

Previous literature presents a strong rationale for the positive impact of symbolic

representation in coproduction contexts. However, empirical studies yield incon-

clusive findings indicating that meaningful effects are limited if citizens face high

levels of uncertainty. This article combines symbolic representation with signaling

theory, suggesting that the representativeness of central reference groups might

reduce uncertainty. The theoretical framework suggests that the representation of

supervisors and existing coproducers might positively affect citizens’willingness

to coproduce. Contrary to the theoretical expectations, the empirical results from

two preregistered factorial survey experiments (n=2979), situated in prisoner

rehabilitation and refugee integration, indicate that the symbolic gender represen-

tation of these reference groups has a limited impact. Only a balanced representa-

tion of coproducers exhibits a positive treatment effect on citizens’willingness to

coproduce. The results oppose central arguments in the representative bureau-

cracy literature. At least for gender categories, symbolic representation is less

important than expected.

Evidence for practice

•Increased representation of female coproducers and supervisors does not gener-

ally increase citizens’willingness to coproduce.

•Coproduction practitioners should not expect positive effects based solely on

displaying an increased representation of these groups.

•If any, a balanced representation of coproducers may be desirable to increase

willingness to coproduce.

•More importantly, public organizations should try to make citizens feel comfort-

able by reducing the uncertainty related to coproduction processes.

Citizens’perceptions of public organizations profoundly

impact citizen-state interactions (Jakobsen et al., 2019). In

essence, how citizens experience public organizations

and public encounters shapes their attitudes toward

the organizations and employees. Attitudes such as

legitimacy should affect citizens’behavior within citizen-

state interactions (Headley et al., 2021; Riccucci & van

Ryzin, 2017). Furthermore, citizens with favorable atti-

tudes toward public organizations should be more willing

to coproduce (Riccucci et al., 2016), referring to service

delivery “not only (…) by professional and managerial

staff in public agencies but also (…) by citizens and

communities”(Brandsen & Honingh, 2016, 427). This

proposition is relevant because public organizations often

depend on coproduction (Uzochukwu & Thomas, 2018).

A pressing question following these elaborations is

whether delimitable factors can positively affect citizens.

Scholars suggest coproduction initiatives might benefit if

an organization represents the society it serves, so-called

“passive representation”(Riccucci & van Ryzin, 2017). The

theoretical arguments indicate that passive representa-

tion might positively affect citizens’attitudes through

symbolic effects (Theobald & Haider-Markel, 2009). This

symbolic representation originates from the salience of

Received: 16 November 2021 Revised: 11 January 2023 Accepted: 18 January 2023

DOI: 10.1111/puar.13619

This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribu tion and reproduction in any medium, provided the

original work is properly cited.

© 2023 The Author. Public Administration Review published by Wiley Periodicals LLC on behalf of American Society for Public Administration.

Public Admin Rev. 2023;83:587–602. wileyonlinelibrary.com/journal/puar 587

passive representation related to social categories, such as

gender, among bureaucrats (Riccucci & van Ryzin, 2017).

For instance, representation increases if women make up a

relevant share of the organizations’employees (Riccucci

et al., 2016; van Ryzin et al., 2017). Still, empirical research

remains inconclusive. Previous studies on gender represen-

tation provide mixed findings, with some supporting the

hypothesized claims about symbolic effects (Meier &

Nicholson-Crotty, 2006; Riccucci et al., 2016; Schuck

et al., 2021; Xu & Meier, 2022). Still, others present findings

that disregard them (Miller et al., 2022; Sievert, 2021;van

Ryzin et al., 2017).

Recently, Sievert (2021) suggested that the effects of

symbolic representation are context-dependent. Positive

effects might decrease in uncertain coproduction contexts

due to information asymmetries. In these settings, uncer-

tainty includes a lack of knowledge about coproduction

processes and activities. Simultaneously, a lack of predict-

ability regarding the bureaucracies’actions constitutes

uncertainty. Citizens often cannot anticipate how bureau-

crats react in public encounters (Headley et al., 2021). Thus,

generic information about aggregated levels of an organi-

zation’s representativeness seems ineffective in reducing

information asymmetries. Thus, decreasing uncertainty and

the resulting information asymmetries seems necessary.

This article combines symbolic representation and signal-

ing theory (Connelly et al., 2011), aiming to reduce citizens’

uncertainty. Instead of focusing on abstract representative-

ness on an aggregated level, this article focuses on the

symbolic representation of identifiable reference groups.

First, as coproduction requires citizens’active participation,

coproduction initiatives exhibit a pool of existing copro-

ducers. Second, the participation of citizens requires pro-

fessional staff that supervises and guides the coproducers.

Information about the representativeness of these groups

should help citizens develop salientexpectations regarding

the procedures (Headley et al., 2021) because they better

understand with whom they will work. Two factorial survey

experiments test these expectations. Study 1 (n=1510) is

set in the policy context of prisoner rehabilitation, while

Study 2 (n=1469) applies refugee integration. The

research designs manipulated the gender distribution of

supervisors and coproducers (Study 1: 3 3 factorial

design; Study 2: 6 6 factorial design). Afterward, partici-

pants indicated willingness to coproduce.

This article complements the empirical literature addres-

sing symbolic representation and coproduction (Riccucci

et al., 2016;Sievert,2021; van Ryzin et al., 2017). The theo-

retical framework combines representative bureaucracy

theory (Riccucci & van Ryzin, 2017) with signaling theory

(Connelly et al., 2011), suggesting that representation sig-

nals must reduce information asymmetries. Contrary to the

theoretical expectations, the empirical results indicate that

the symbolic representation of coproduction supervisors

did not matter. Furthermore, only a balanced gender distri-

bution of existing coproducers had a positive treatment

effect. Overall, symbolic representation of these reference

groups had a limited causal impact. The discussion indi-

cates that coproduction contexts deviate from other public

encounters, which may explain the limited relevance of

both interventions. For instance, citizens cannot observe

bureaucratic behavior and lack contextual information.

Overall, symbolic representation seems to play a minor r ole

in coproduction. Still, practitioners should foster representa-

tion in public organizations. However, the primary rationale

should be to increase active representation since the sym-

bolic effects of representation seem to constitute a costly

yet, overall, weak signal.

THEORETICAL FRAMEWORK

Theory of representative bureaucracy

Representative bureaucracy is at the core of public admin-

istration research because it addresses mechanisms shap-

ing citizen-state interactions. This includes focusing on

employees responsible for service delivery and policy

implementation ( Meier, 2019). In addition, representative

bureaucracy focuses on citizens facing public organizations

and employees (Headley et al., 2021; Jakobsen et al., 2019).

The theory of representative bureaucracy has seen consid-

erable development throughout the last decades (Ding

et al., 2021). Based on a descriptive perspective, scholars

outlined the assumption that public organizations’staff

should resemble society, that is, those the organization is

supposed to serve (Bishu & Kennedy, 2020). This passive

representation indicates whether a public organization’s

workforce mirrors the characteristics of the citizenry

based on an analysis of the demographic composition of

employees (Meier & Bo hte, 2001;Selden,1997). Still,

scholars suggest that passive representation alone cannot

change how well an organizatio n serves its citizen s.

Instead, it requires tangible decisions and the behavior of

public employees. This active representation entails policy

decisions and discretion when bureaucrats “press for the

interests and desires of those whom he is presumed to

represent”(Mosher, 1968, p. 12).

Recently, attention shifted to symbolic repres-

entation (Riccucci & van Ryzin, 2017; Theobald & Haider-

Markel, 2009), suggesting that public employees’

demographics are relevant. In particular, citizens’atti-

tudes may change if they encounter representative public

organizations (Headley et al., 2021; Miller & Keiser, 2021).

This reasoning originates from the need to explain earlier

findings. For instance, some studies showed congruence

effects in schools, where girls perform better in the pres-

ence of female teachers (Keiser et al., 2002). Apart from

the teacher’s behavior, a “potential response to what

female teachers represent for female students”may drive

these results (Theobald & Haider-Markel, 2009, p. 412).

Similarly, other findings attributed to active representa-

tion may instead result from clients’behavior (Meier

et al., 1999; Meier & Nicholson-Crotty, 2006). Early

588 THE LIMITED IMPACT OF SYMBOLIC GENDER REPRESENTATION