This article was downloaded by: [185.107.56.

83] On: 13 March 2024, At: 10:00

Publisher: Institute for Operations Research and the Management Sciences (INFORMS)
INFORMS is located in Maryland, USA

Manufacturing & Service Operations Management

Publication details, including instructions for authors and subscription information:
http://pubsonline.informs.org

Inclusive Manufacturing: The Impact of Disability

Diversity on Productivity in a Work Integration Social
Enterprise
Sriram Narayanan, Ed Terris

To cite this article:

Sriram Narayanan, Ed Terris (2020) Inclusive Manufacturing: The Impact of Disability Diversity on Productivity in a Work
Integration Social Enterprise. Manufacturing & Service Operations Management 22(6):1112-1130. https://doi.org/10.1287/
msom.2020.0940

Full terms and conditions of use: https://pubsonline.informs.org/Publications/Librarians-Portal/PubsOnLine-Terms-and-

Conditions

This article may be used only for the purposes of research, teaching, and/or private study. Commercial use
or systematic downloading (by robots or other automatic processes) is prohibited without explicit Publisher
approval, unless otherwise noted. For more information, contact permissions@informs.org.

The Publisher does not warrant or guarantee the article’s accuracy, completeness, merchantability, fitness
for a particular purpose, or non-infringement. Descriptions of, or references to, products or publications, or
inclusion of an advertisement in this article, neither constitutes nor implies a guarantee, endorsement, or
support of claims made of that product, publication, or service.

Copyright © 2020, INFORMS

Please scroll down for article—it is on subsequent pages

With 12,500 members from nearly 90 countries, INFORMS is the largest international association of operations research (O.R.)
and analytics professionals and students. INFORMS provides unique networking and learning opportunities for individual
professionals, and organizations of all types and sizes, to better understand and use O.R. and analytics tools and methods to
transform strategic visions and achieve better outcomes.
For more information on INFORMS, its publications, membership, or meetings visit http://www.informs.org
 MANUFACTURING & SERVICE OPERATIONS MANAGEMENT
Vol. 22, No. 6, November–December 2020, pp. 1112–1130
http://pubsonline.informs.org/journal/msom ISSN 1523-4614 (print), ISSN 1526-5498 (online)

Inclusive Manufacturing: The Impact of Disability Diversity on

Productivity in a Work Integration Social Enterprise
Downloaded from informs.org by [185.107.56.83] on 13 March 2024, at 10:00 . For personal use only, all rights reserved.

Sriram Narayanan,a Ed Terrisa

a
Department of Supply Chain Management, Michigan State University, East Lansing, Michigan 48824
Contact: sriram@msu.edu, https://orcid.org/0000-0003-2126-4227 (SN); terrised@msu.edu (ET)

Received: October 4, 2017
Revised: September 11, 2018; July 15, 2019;
August 13, 2020; August 27, 2020
Accepted: August 27, 2020
Published Online in Articles in Advance:
October 19, 2020
https://doi.org/10.1287/msom.2020.0940
Copyright: © 2020 INFORMS
Abstract. Problem definition: This study examines the impact of disability diversity on the
productivity of apparel manufacturing teams in the context of a work integration social
enterprise. Two measures of disability diversity are examined: the number of disability
categories employed in a production line and the evenness of disability category dispersion
among workers employed in an apparel production line. Academic/practical relevance:
The problem has both academic and practical relevance. From an academic standpoint, the
current literature does not study the implications of employing individuals with dis-
abilities and their impact on productivity with microdata. The issue also has practical
relevance because it ties into recent managerial interest to employ individuals with dis-
abilities in organizations. According to statistics, only 35% of workers with disabilities have
some form of significant full-time employment. Methodology: The study uses panel re-
gression analyses to test the impact of disability diversity (number of disability categories
and evenness of disability category dispersion) of workers in a production line using
detailed multiyear data on the productivity of apparel manufacturing cells. Results: Two
key insights emerged from the analyses. First, productivity can be enhanced by increasing
the diversity of workers with disabilities in the workforce within a garment-manufacturing
cell. Specifically, productivity is best at moderate levels of disability categories in a team.
Second, team productivity is higher at greater levels of evenness of disability category
dispersion. Managerial implications: The analysis in this paper sheds light on the potential
benefits of integrating individuals with disabilities into organizations and its implications
on productivity. Specifically, the study finds evidence that having moderate levels of
disability categories on a team with higher levels of evenness in disability category dis-
persion is associated with better productivity rather than having a concentrated team
focused on a specific disability. The implications of the results and limitations for the study
as well as its potential insights into the context of social enterprises that employ individuals
with disabilities are discussed.

History: This paper has been accepted for the Manufacturing & Service Operations Management Special
Issue on Responsible Operations and Supply Chain Management.
Funding: The authors acknowledge grant funding from the Peckham Foundation that facilitated the
data collection for the project.
Supplemental Material: The online appendix is available at https://doi.org/10.1287/msom.2020.0940.

Keywords: disability diversity • disability dispersion • disability inclusion • manufacturing productivity

1. Introduction may be as important as environmentalism (The

Individuals with (dis)abilities are a vital part of any
economy. In the United States, according to the 2017
disability statistics released by Cornell University,
about 12.7% of the population has some form of
disability (Erickson et al. 2017). People with disabil-
ities are an important source of talent given the labor
shortage (Gaudiano and Hunt 2017). The issue of
employing individuals with disabilities is also so-
cially critical as identified by the United Nations Sus-
tainable Development Goals (SDGs) (United Nations
2015). The Economist notes the employment of indi-
viduals with disabilities as the “new green,” stating
that the issue of managing individuals with disabilities
Economist 2012). Despite its importance, only 35% of
individuals with disabilities have some form of sig-
nificant full-time employment in the United States
(Gaudiano and Hunt 2017). The Bureau of Labor
Statistics reports that the workplace participation rate
of individuals with disabilities is at 20.6% compared
with 68.6% for those without (Bureau of Labor Statistics
2017). With the tightening labor market, opportuni-
ties to employ individuals with disabilities have sub-
stantially expanded. Yet practitioners note that substantial
challenges remain (Kennedy and Gaul-Stigge 2019).
Disability is frequently seen as a barrier to em-
ployment. A recent Deloitte consulting report on the

1112
 Narayanan and Terris: Inclusion, Disability Diversity, and Productivity
Manufacturing & Service Operations Management, 2020, vol. 22, no. 6, pp. 1112–1130, © 2020 INFORMS
inclusion of individuals with disabilities states that
substantial barriers exist regarding their employment
(Allen and Cohen 2010). These include negative at-
titudes toward individuals with disabilities, hiring
practices that exclude individuals with disabilities,
lack of technological assistance, and lack of education
with respect to disabilities within firms (Allen and
Downloaded from informs.org by [185.107.56.83] on 13 March 2024, at 10:00 . For personal use only, all rights reserved.
Cohen 2010). These barriers need to be actively man-
aged (Allen and Cohen 2010); if they fail to be man-
aged, they can result in organizations not utilizing
individuals’ talents. Both accommodations and tech-
nologies are used to overcome these barriers.
Although commercial organizations are rising to
the challenge of creating an integrated workforce in
which individuals with and without disabilities can
work together, an important source of employment
for individuals with disabilities are social enterprises
(Kerlin 2006). With the primary goal of working for
societal causes, social enterprises play a pivotal role in
several economies (Santos et al. 2015). Within the
United States and Europe, a set of organizations
dedicated to the cause of employing individuals with
serious disabilities is also referred to as “sheltered
workshops.” The aim of these workshops is to focus
on assisting individuals with significant disabilities
(Mallender et al. 2015) with the idea of transitioning
them into the open labor market (Hoffman 2013). In
2014, sheltered workshops alone employed 136,000
adults in the United States (Cooney 2016).
Also called work integration social enterprises
(WISEs) (InWork 2020), these organizations primarily
focus on solving social challenges (Battilana et al.
2015) while prioritizing the larger social mission of
efficiency. For example, several organizations em-
ploy individuals with intellectual disabilities while
serving their medical and social needs (Migliore et al.
2008). Although these organizations likely provide
superior services to individuals (Migliore et al. 2008),
they also have significant operational challenges in
making their production efficient (Neumann 2019).
These organizations face challenges related to bal-
ancing work-related operational constraints (e.g., fre-
quent interruptions in work), including the need to
serve clients with specific disabilities (e.g., intellectual
disabilities) who may need specific accommodations
(e.g., additional breaks). Accommodations are pro-
vided to individuals for them to perform work to the
best of their abilities. The Americans with Disabilities
Act (ADA) notes that reasonable accommodations
are a right to employment for individuals with dis-
abilities (Stamps 1995). In addition to accommoda-
tions, these firms also work on tailoring jobs to fit
individual abilities—also referred to as “job carving”
(Griffin et al. 2007, Scoppetta et al. 2019).
Overall productivity improvement—within the con-
fines of the organizational mission—is a key challenge
1113
within these WISE organizations and is important
to examine. Improving productivity can result in
greater commercial viability for these enterprises, and
can help serve their social mission, for example,
supporting individuals with disabilities (Neumann
2019). Furthermore, examining approaches to im-
prove productivity may also enable these organiza-
tions to effectively perform their role of integrating
individuals with disabilities into equal opportunity
environments (Hoffman 2013). This is a critical issue
given that these WISEs have recently been under
pressure to integrate these individuals into the open
labor market in both Europe (Mallender et al. 2015)
and the United States (Hoffman 2013). This study
focuses on such a setting in examining the implica-
tions of employing workers (also referred to as in-
dividuals) with disabilities on productivity.
Focusing on work settings, individuals with dis-
abilities work in individual and team settings. In
individual settings, workers can better pace their
work within the realm of their accommodations.
However, any variability in productivity is restricted
to the individual. In contrast, for team settings, the
tasks performed must be coordinated with other
upstream and downstream workers, and the skills
of one employee may benefit the other employees.
For example, employing individuals across disability
categories within a coordinated production line may
facilitate firms’ ability to carve out jobs that can
leverage a specific individual’s abilities (Narayanan
and Terris 2020) while facilitating task coordination.
This can improve team productivity and promote in-
clusion (e.g., Miralles et al. 2007). Parallels can be
seen in the Toyota Sewing System (Black and Kohser
2011), in which workers share tasks in a bucket bri-
gade (Bartholdi and Eisenstein 1996), or work-sharing
systems in a serial assembly line in which employees
share tasks (McClain et al. 2000). A bucket brigade
may also promote inclusion by leveraging workers’
abilities across disability categories in a group and by
leveraging assignment flexibility within steps. Nev-
ertheless, promoting inclusion within these lines is
nontrivial (Scoppetta et al. 2019). It is accomplished
through investing effort to carve out jobs to match
individual skills, provide necessary accommodations,
implement job redesign, and develop programmed
instruction (Mithaug 1979). Drawing from this dis-
cussion, this study seeks to document the impact of the
number of disability categories employed in a team on
its productivity.
Furthermore, compared with individual work, man-
aging a team of workers with disabilities is also chal-
lenging from a social justice standpoint. In addition
to the operational complexity of managing several
workers with disabilities in a team, there are social
challenges of managing employee perceptions with

1114 Manufacturing & Service Operations Management, 2020, vol. 22, no. 6, pp. 1112–1130, © 2020 INFORMS
respect to accommodations (Colella 2001). Thanem
(2008, p. 584) notes, “Initiatives aimed at enabling
one disabled person (e.g., a wheelchair user) may
disable another (e.g., a visually impaired person).”
Accordingly, managing a diverse group of individ-
uals with disabilities is more challenging in team-
based settings (Thanem 2008). One approach to en-
Downloaded from informs.org by [185.107.56.83] on 13 March 2024, at 10:00 . For personal use only, all rights reserved.
suring fairness is the idea of allocative fairness or
ensuring statistical parity by focusing on the pro-
portion of allocated individuals within a team that
belong to a specific category (Trewin 2018). This
statistical parity relates to the concept of “evenness”
in the diversity literature, which emphasizes the relative
representation of subcategories within a larger group
(Harrison and Klein 2007). An even distribution of
individuals across disability categories may likely
contribute to better perceptions of fairness and improve
between-individual affinity, a critical aspect of social
support for individuals with disabilities (Dunn and
Burcaw 2013). However, the empirical implications
regarding the evenness of disability category dis-
persion (or simply “disability dispersion”) on the
productivity of a team is not documented.
Overall, drawing from this discussion, this study
seeks to answer two research questions: (a) What is
the impact of increasing the number of disability cate-
gories (referred to as disability categories, henceforth)
in a team on its productivity? (b) What is the impact of
the evenness of disability category dispersion (referred
to as disability dispersion) on a team’s productivity?
These two measures are attributes of diversity as
variety (Harrison and Klein 2007). The idea of dis-
abilities as distinct categories has been used within
the literature in predicting the employment of indi-
viduals with disabilities (Martz and Xu 2008) and in
the incidence reporting on the nature of disabilities
(Elkind 1990). Williams and Mavin (2012) provide a
discussion on the epistemological roots of disability
categorization. Further, the relative evenness con-
cerning the distribution of different categories is re-
flective of within-category variations, particularly
when individuals with disabilities work in a team and
produce coordinated output.
This investigation is primarily exploratory. Fol-
lowing Nicolae et al. (2017), this study does not imply
causality on the impact of disability categories or
disability dispersion on team productivity. Rather, it
explores these variables as potential drivers of pro-
ductivity and documents their association. Further,
this study does not intend to establish the mecha-
nisms; instead, it raises questions that elicit the need
to understand the mechanisms influencing the impact
of disability diversity on productivity.
Overall, this study contributes to the literature by
being the first to document the empirical relationship
Narayanan and Terris: Inclusion, Disability Diversity, and Productivity
between disability diversity and team productivity
with microdata. Two different literature reviews point
to this gap. Shore et al. (2009, p. 122) note that, al-
though there are several streams of theoretical work
on disabilities, “empirical research is somewhat sparse.”
Similarly, in a comprehensive review of more than
20 years of research (from 1997 to 2017) on the pro-
ductivity benefits of employing individuals with dis-
abilities, Lindsay et al. (2018) note that only eight
studies document increased productivity from employ-
ing individuals with disabilities. These studies consist
of self-reported data and do not capture the types of
disabilities involved. Thus, unlike past studies that
focus on self-reported outcomes or case studies, this
study uses archival operational data to examine pro-
ductivity outcomes and captures the specific types of
disabilities involved. Capturing the type of disability
for individuals in the team is critical to studying the
impact of disability diversity (Lindsay et al. 2018).
The relationship between productivity and diver-
sity is of interest in the operations management and
the economics literature as well. For example, the
operations management literature examines the issue
of diversity on productivity at both the individual
and team levels in various contexts. At the individual
level, the implications of diversity on productivity is
examined in the contexts of commercial banking (Staats
and Gino 2012), software maintenance (Narayanan
et al. 2009), and surgical procedures (KC and Staats
2012). At the team level, the implications of diversity
of employee experience on productivity have been
examined in healthcare service delivery (Akşin et al.
2020), software development (Huckman and Staats
2011), and software maintenance (Narayanan et al.
2014). Further, the diversity of employee abilities is
found to positively impact productivity in apparel
manufacturing (Hamilton et al. 2012). The positive
impact of workforce diversity on productivity is
also found at the firm level (e.g., Garnero et al. 2014,
Parrotta et al. 2014). None of these studies examines
the relationship between disability diversity and team
productivity, an issue of interest in this study.
The empirical analysis of this study is based on pro-
prietary data collected from a single firm, Peckham, Inc.,
and is embedded in the apparel-manufacturing con-
text. Despite the limitations of the specific context
(i.e., apparel production), it is worth noting that the
garment/apparel industry employs more than 75 million
individuals worldwide (United Nations Economic
Commission for Europe 2018). Hiring individuals
with disabilities within this industry is promoted
internationally as seen in a recent collaboration be-
tween German and Bangladesh governments, such as
the inclusive job center initiative (Green Economy
Coalition 2019) or the recent hiring of more than
 Narayanan and Terris: Inclusion, Disability Diversity, and Productivity
Manufacturing & Service Operations Management, 2020, vol. 22, no. 6, pp. 1112–1130, © 2020 INFORMS
250 workers for a garment factory in El Paso, Texas
(Kolenc 2017). The industry also relies on teamwork,
particularly when factories pursue a modular pro-
duction approach (Berg et al. 1996, Hamilton et al.
2003) as is the case in the data-collection context of this
study (described in Section 2). Finally, informal in-
terviews were conducted with senior executives from
Downloaded from informs.org by [185.107.56.83] on 13 March 2024, at 10:00 . For personal use only, all rights reserved.
two other smaller WISEs to complement the data
analysis in addition to collecting quotes that cor-
roborate key findings from within the firm as po-
tential validation of the insights.
The remainder of the paper is organized as follows.
In Section 2, the study context is detailed. In Section 3,
an overview of the literature on disability, task analysis,
and team productivity is presented. Exploratory argu-
ments are offered on the impact of disability diversity
on productivity. In Section 4, the data and empirical
strategy are detailed. In Section 5, the results of the
analysis are detailed and additional post hoc analyses
are conducted. In Section 6, the implications of the
study are discussed along with its limitations.
2. Study Context
The study site for this research is Peckham, Inc.
Founded in 1976, Peckham, Inc., is a not-for-profit
WISE that pursues vocational rehabilitation for in-
dividuals with disabilities. According to the website
(www.peckham.org), the organization’s mission is
to “provide a wide range of opportunities to maxi-
mize human potential for persons striving for inde-
pendence and self-sufficiency” (Peckham 2020). The
primary mission of the organization is to serve in-
dividuals with disabilities to transition them into
an open labor environment. It is one of the largest
community rehabilitation programs in the state of
Michigan. As a key supplier in the AbilityOne net-
work, Peckham adheres to stipulations that more than
75% of hours should be billed by individuals with
significant disabilities. A substantial source of the
manufacturing organization’s revenue comes from
contracts to manufacture apparel under the Abil-
ityOne Program, which is the largest source of em-
ployment for people with significant disabilities within
the United States. The program employed more than
45,000 people in 2018, many of whom worked on
contracts with the Department of Defense (AbilityOne
2019). It facilitated procurement of more than $3.3
billion of goods in 2017 (AbilityOne 2019).
Peckham has multiple key business areas: apparel
manufacturing, environmental services, farms, business
services, packaging, and supply chain solutions. Apparel
manufacturing was the largest business line at the time
of the data collection. In 2016, the firm assisted about
789 individuals in gaining community employment.
This study is situated in a manufacturing organization
that specializes in high-value and high-quality apparel.
1115
The manufacturing operations are based out of two
factories located approximately 50 miles apart. The
firm’s manufacturing organization had 700+ pro-
duction team members at the time of the data col-
lection. Peckham pursues a philosophy of providing
opportunities for diverse disability categories and
adopts an inclusive approach to the recruitment,
job assessment, training, and deployment of indi-
viduals across the disability spectrum into opera-
tional environments. This stands in contrast to some
of the organizations that support individuals with
specific disability conditions (e.g., Migliore et al. 2008,
Neumann 2019). The firm’s business configuration
diversity and approach to inclusion allow more work-
related flexibility to provide individuals with dis-
abilities opportunities for employment that fits their
needs. For example, within apparel manufacturing,
the individuals employed had a range of physical and
mental disabilities.
In addition to the Department of Defense engage-
ments via AbilityOne, the firm also works with the
commercial manufacturers of some leading athletic
brands that focus on corporate social responsibility
initiatives. The firm cut more than 2.1 million garment
pieces in its facilities in 2016. The firm’s sophisticated
garment-manufacturing operation is comparable with
several leading apparel manufacturers with digital
pattern design machines, die cuts, and embroidery
machines, to name a few. The firm also relies on a
sophisticated manufacturing production system that
focuses on lean process implementations. It provides
six-sigma training to its employees with black belt and
green belt individuals on its staff. Peckham, Inc., is
International Organization for Standardization: 9001
2015 and Commission on Accreditation of Rehabili-
tation Facilities (CARF) certified. CARF is a commu-
nity accreditation program that evaluates programs
such as Peckham, Inc., for the quality of rehabilitation
services provided to associates within the social en-
terprise. The CARF certification attests to Peckham’s
significant emphasis on workforce development pro-
grams, which is consistent with its mission of tran-
sitioning the workforce into an open labor environ-
ment. The firm’s apparel operations are registered
with the U.S. Department of Labor for apprenticeship
training programs, a significant accomplishment for
garment manufacturers.
The firm prides itself on paying attention to matching
specific employees to roles that fit their skills and
abilities to help them perform optimally. This is done
with the help of dedicated recruiting teams and vo-
cational service specialists (VSSs). The recruiting team
focuses on the intake, eligibility, and onboarding of
workers. The VSS team works with the recruiting
team in assessing individual disabilities. Further, the
VSS team determines for which specific business line

1116 Manufacturing & Service Operations Management, 2020, vol. 22, no. 6, pp. 1112–1130, © 2020 INFORMS
(e.g., manufacturing, call center, etc.) each worker is best
suited and coordinates with the respective business lines
to place them. Should the firm not find a match for a
specific individual within any of its business lines, it
works with other external agencies to place these indi-
viduals into job roles that fit the individual’s interests
and skills. After placement, the VSS team also works
Downloaded from informs.org by [185.107.56.83] on 13 March 2024, at 10:00 . For personal use only, all rights reserved.
in coordination with the operations to manage the
worker’s day-to-day accommodation needs and other
work-related challenges. Individuals are assigned to a
VSS staff member who works closely with the individual
to provide insights into the individual’s special needs.
This aspect is unique within the setting studied,
wherein many of the workers require active, ongo-
ing support. Having dedicated, ongoing support for
workers also ensures that the firm learns better about
the employees’ abilities and facilitates informed de-
cisions about assignments, reassignments, and accom-
modations. The firm also has an internal department
that focuses on designing automation solutions to
accommodate individuals’ disabilities within its pro-
cesses. Once individuals are assigned to a specific
business line, the operational staff trains them to
perform specific jobs (e.g., sewing in the case of ap-
parel) and focuses on on-the-job deployment in op-
erational settings. On-the-job training of employees
requires constant experimentation on the part of
supervisors, operational staff, and the VSS team.
The study data were collected from the firm’s
apparel-manufacturing unit. Apparel manufacturing
has two broad production methodologies: a module
system and a progressive bundling system (Berg et al.
1996). Although the progressive bundle system is
designed to minimize contact between workers, the
module system, in contrast, requires a substantial
degree of teamwork and helps improve productivity.
The modular production system is primarily used to
make the production of apparel more flexible and
responsive, thereby requiring groups of operators to
work together to assemble the garment (Berg et al.
1996). The plant operates under the ideology of a
single piece flow that minimizes inventory and waste
in manufacturing and, thus, applies a modular pro-
duction strategy. This approach also empowers workers
to engage with one another better and take ownership
in their jobs, building “respect for one another”—a
key tenet of lean production (Womack 2007). Workers
are not paid based on piece rates; rather, they are
paid productivity bonuses as a team based on target
achievement. The team bonus complements indi-
vidual compensation, which is based on the worker’s
skill level. Workers have four skill levels with 1 being
the lowest and 4 being the highest.
The apparel setting at Peckham, Inc., is ideally
suited for studying the impact of disability diversity
Narayanan and Terris: Inclusion, Disability Diversity, and Productivity
on team productivity for multiple reasons. First, gov-
ernment policies within the WISE sector have in-
creasingly moved toward the model of competitive
integrated employment (CIE) (Hoffman 2013). Under
this model, individuals with and without disabilities
frequently work side by side with no distinction in
pay or privileges (Allison et al. 2017). For example, the
14(c) rule for many organizations that employ indi-
viduals with disabilities allows for subminimum wages
to be paid to these individuals, which a CIE model
aims to supersede. Given the diverse business lines
within the firm, Peckham, Inc., offers CIE opportu-
nities for individuals by leveraging the inherent va-
riety in the jobs across various business lines that do
not distinguish between workers with and without
disabilities for specific jobs. The apparel-manufacturing
division is a setting with CIE, in which individuals
with and without disabilities work side by side in
teams. This setting provides a unique opportunity to
study disability diversity while controlling for situa-
tions in which individuals have no recorded disabil-
ities. Anecdotal evidence of interactions with practi-
tioners suggests that the scale for employing individuals
with disabilities in a typical corporate environment
is not similar, thus enhancing the value of studying
disability diversity in this environment.
Second, individuals in teams come from diverse
categories of disabilities, performing the same set of
tasks, that is, sewing, to achieve a team goal. This
consistency of tasks across employees contrasts with
other possible settings in which individuals may
perform tasks that are not comparable. The variety of
employees with disabilities found on the manufacturing
lines is possible because of the firm’s philosophy of
supporting individuals across a range of disability
categories in their business operations. Further, var-
iations in team size across garments and disability
dispersion within (and across) lines, in addition to
variation in the disability categories, allows for a
study of disability diversity. A discussion on dis-
ability diversity, task analysis, and productivity are
presented next.
3. Disability, Task Analysis,
and Productivity
Employing individuals with disabilities in an envi-
ronment can be seen through the lens of a socio-
technical system (Trist 1981). The sociotechnical view
argues that it is necessary to find an ideal fit between
the technical and social subsystems to allow em-
ployees to work effectively. According to Trist (1981),
sociotechnical systems theory views people as an
important resource and puts considerable emphasis
on creating an appropriate environment for effective
 Narayanan and Terris: Inclusion, Disability Diversity, and Productivity
Manufacturing & Service Operations Management, 2020, vol. 22, no. 6, pp. 1112–1130, © 2020 INFORMS
task execution. The sociotechnical system recom-
mends that work practices facilitate the effective in-
tegration of human resources and technology (Cherns
1976). Within the context of employing people with
disabilities, sociotechnical system design includes
matching the individual’s ability to the job’s design,
the accommodations provided to individuals, and
Downloaded from informs.org by [185.107.56.83] on 13 March 2024, at 10:00 . For personal use only, all rights reserved.
technical tools necessary to facilitate work. Techni-
cal tools include automation, task analysis, and
methods of programmed instruction and training
(Mithaug 1979).
Accommodations are an integral part of employing
individuals with disabilities (Thanem 2008) and are
unique to an individual. As a right to employment,
per ADA guidelines, accommodations allow workers
to perform to the best of their abilities and overcome
restrictions that they experience in the effective per-
formance of work. Accommodations may include
work modifications, changes to shift schedules, and
task automation that varies by the nature of the
disability. For example, accommodations could in-
clude allowing workers time during work hours to
seek counseling; providing roles that can fit indi-
vidual constraints, such as avoiding walking or stand-
ing; and longer break hours, etc., among others (Colella
2001). Further, the reasons for accommodations are
confidential and are usually not disclosed without
the explicit consent of the individual. Finally, co-
workers are not entitled to know the reason for ac-
commodation for privacy reasons. These tenets not
only impact the importance of the accommodations
provided to a specific individual to help the person
perform better, but also influence the supervisor’s
approach to managing accommodations (Williams-
Whitt et al. 2016). Beyond the individual–supervisor
relationships, the tenets of accommodations described
also create differential fairness perceptions regarding
others’ accommodations within a team, which can
influence the team’s work design and motivation
(Colella 2001). Thus, management must not only
provide accommodations for specific individuals, but
must also ensure that perceptions of others’ accom-
modations are effectively managed.
In addition to providing workers with accommo-
dations, firms may also consider task redesign to
leverage specific individual abilities. The redesign of
tasks to fit individual abilities is also referred to as job
carving (Scoppetta et al. 2019). In its broadest defi-
nition, job carving refers to the customized devel-
opment of positions around individual skills. Job
carving is a term used in the context of employing
individuals with severe disabilities (Nietupski and
Hamre-Nietupski 2000). For example, the innovative
use of jigs and fixtures in plants can help overcome
specific individual challenges (Narayanan et al. 2019).
1117
Job carving is frequently practiced in settings in
which supporting the employment of individuals
with disabilities is the primary goal (Ho 2018). Job
carving may involve adapting roles within firms to
employees’ skills or the creation of new roles (Nietupski
and Hamre-Nietupski 2000).
Moreover, job carving may involve “the rearrange-
ment of work tasks” that “can lead to new job op-
portunities and add value to all involved” (Scoppetta
et al. 2019). It is a process that is undertaken with the
help of specialized consultants who focus on voca-
tional rehabilitation (Scoppetta et al. 2019). Job carving
efforts also require a systematic approach to work-
force development (Fesko et al. 2008), including col-
laborations with individuals and accommodating spe-
cific worker preferences, such as seeking additional
rotations across tasks. Carving jobs may also involve
experimentation to assess the best fit between task
needs and employee abilities (Williams-Whitt et al.
2016). Effective job carving needs to integrate rigor-
ous task analysis with effective programmed in-
struction and training.
Modern scientific work methodologies emphasized
by Frederick Taylor, as well as time and motion
studies, have emphasized task analysis. Although
traditional task analysis focuses on finding the most
efficient way of doing things (Barnes 1949), in the
context of individuals with disabilities, traditional
task analysis can be complemented with programmed
instruction that can be designed to suit individual
abilities (Mithaug 1979). Compared with task analysis
in traditional operations management settings, pro-
grammed instruction focuses on developing employee
skills. This type of instruction refers to the presentation
of instructional cues and arrangements. Moreover, it
facilitates a high level of expected responses and al-
lows for the reinforcement of these responses, which
can successfully improve employee behavior on the
job (Mithaug 1979). These instructional cues may also
involve specific training for workers to perform the
task and may consider their specific challenges to
make the job more meaningful. The collective goal
of both task analysis and programmed instruction
is to reduce process variance and improve output.
Overall, managing workers with disabilities within a
line requires a careful juxtaposition of job carving,
accommodation management, task analysis, task reas-
signment, and programmed instruction. Table (i) of the
online appendix presents an example set of accom-
modations and restrictions.
Based on this discussion, the complexity of managing
a team of workers with disabilities makes it chal-
lenging to develop the theory and hypotheses on the
impact of disability diversity in teams on productivity.
Further, isolating a specific mechanism to assess the

1118 Manufacturing & Service Operations Management, 2020, vol. 22, no. 6, pp. 1112–1130, © 2020 INFORMS
causal implications of the impact of disability diversity
on productivity is difficult. In practice, accommoda-
tions need not be consistently implemented for the
same individual across time given changing individ-
ual needs. Managing accommodations may also ne-
cessitate task redesign and task reassignment. An
example of a typical line with its accompanying re-
Downloaded from informs.org by [185.107.56.83] on 13 March 2024, at 10:00 . For personal use only, all rights reserved.
strictions and accommodations for each worker in the
team is shown in Online Table (ii). Online Table (ii)
details restrictions and accommodations that em-
ployees were provided, the ongoing involvement of
VSS (if needed), and potential automation solutions/
advocacy for workers embedded in a line. An im-
portant point to note is that, for each worker in the
line, the accommodations and restrictions can change
depending on their medical needs at the time. A lack
of specific accommodations/restrictions for an indi-
vidual can also change over time based on the worker’s
needs. In recognition of this challenge, in the para-
graphs that follow, “exploratory arguments” are of-
fered on the impact of disability categories and dis-
ability dispersion on team productivity.
3.1. Disability Categories and Productivity
In modular apparel production, teams play a prom-
inent role (Berg et al. 1996). Opportunities to match an
individual’s disability to the skill and ability needs of
the task are likely higher when the team has workers
that span a range of disability categories. A range of
distinct disabilities within a team may provide flex-
ibility to line supervisors in allocating workers to
tasks that enable them to match accommodation, task,
and task sequencing that better fit individual skills.
For example, workers with visual challenges may
navigate their environments differently compared
with workers with hearing impairments. Similarly,
workers with visual challenges who are involved in
sewing tasks may potentially benefit from additional
support in the form of automation (Narayanan et al.
2019). Cognitive or physical challenges may impede
an employee from working effectively. For example, a
worker with poor bodily function may be effective in a
less strenuous step. In general, these natural variations
across individual skills may be used within a larger line
to facilitate productivity improvements.
In support of this argument, Henry Ford, an em-
ployment visionary, classified jobs within the Ford
assembly line and stated, “Out of the 7882 different
jobs classified at the time within Ford assembly lines,
949 were classified as heavy work, 3595 required no
physical exertion, and 3338 required ordinary phys-
ical exertion” (Ford and Crowther 1922, p. 108). The
authors also state, “Out of 7882 kinds of jobs, 4034—
although some of them required strength—did not
require full physical capacity” (Ford and Crowther
1922, p. 108) and suggest that, if appropriately applied,
Narayanan and Terris: Inclusion, Disability Diversity, and Productivity
a worker with a physical disability should theoretically
be able to perform as well as anyone else, provided that
the task setting is appropriate. Reflecting this logic,
Miralles et al. (2007, p. 190) note that assembly lines
that employ individuals with significant disabilities
“can provide many advantages since the traditional
division of work in single tasks may become a perfect
tool for making certain worker disabilities invisible.”
In practice, accommodating natural variations in
worker skills requires experimentation as it is not
feasible to predetermine which worker will effec-
tively accomplish a specific step (Williams-Whitt et al.
2016). For example, Williams-Whitt et al. (2016) argue
that implementing task-level accommodation is a
result of work/medical restrictions, available accom-
modation options, and worker abilities. A dominant
theme that emerges in their research is the notion of
“trial and error” in finding and implementing accom-
modations that work for a specific individual. Managing
these accommodations requires the engagement of
human resources, vocational rehabilitation, and op-
erations personnel (i.e., supervisors) in addition to a
significant degree of brainstorming and adaptation
(Williams-Whitt et al. 2016, Narayanan et al. 2019).
Williams-Whitt et al. (2016), note that, in all 23 of their
case studies related to identifying worker accom-
modations, supervisors describe trial and error as a
salient factor. One of the supervisors in the study
facility states that assigning workers to specific steps
is mostly trial and error. They also note, “If I notice
someone who has a noticeable physical disability and
they are struggling on a machine/step, I’m not going
to keep them there. I will try and find something with
less movement/easier.”
Finally, the embodied approach to managing dis-
ability argues that individuals with disabilities cope
with their restrictions differently, and by this logic,
they are also impacted differently (Thanem 2008). As
teams add workers from a range of disability cate-
gories, it is likely that the demands associated with
managing accommodations also increase (see Online
Table (ii)). These demands are also tied to peer con-
flict, employee motivation, and engagement and,
consequently, team productivity.
3.2. Disability Dispersion and Productivity
Even when a team has individuals from multiple
disability categories, they may differ in the relative
dispersion of individuals across categories. Disability
dispersion may present trade-offs in allocating tasks
to workers. A concentrated pool of individuals from a
specific disability category may minimize the flexi-
bility that supervisors have in assigning individuals
to a diverse range of tasks within a line. Additional
effort to find task matches and carve out jobs in a
team is needed within a specific work environment
 Narayanan and Terris: Inclusion, Disability Diversity, and Productivity
Manufacturing & Service Operations Management, 2020, vol. 22, no. 6, pp. 1112–1130, © 2020 INFORMS
(Williams-Whitt et al. 2016). With higher disability
dispersion, it is likely that the team is not only more
flexible, but also more receptive to accommodations
by virtue of multiple individuals within a group
across a range of disabilities.
Individuals with disabilities working in a team
may also have different perceptions of justice in terms
Downloaded from informs.org by [185.107.56.83] on 13 March 2024, at 10:00 . For personal use only, all rights reserved.
of how the other person is treated (Thanem 2008),
which likely impacts team productivity. Specifically,
a worker’s efforts may be driven by coworker reac-
tions to specific accommodations (Colella 2001). Co-
worker perceptions of accommodation fairness are
likely to engender cohesiveness within the group and
encourage a positive team response (Colella 2001).
Alternatively, perceptions of a lack of fairness in
accommodations can result in workers perceiving
outcomes as distributively unfair, that is, a lack of
returns for their effort (Colella et al. 2004). This may
lead workers to develop resentment toward other
individuals with a different set of accommodations
(Williams-Whitt et al. 2016). When there is relative
evenness in the representation of workers across
disability categories, these workers are exposed to
others’ accommodations, which likely have a similar
pattern. With a relative evenness in the distribution of
disability categories, such resentment may be lower.
Variable accommodations across a range of disabil-
ities are likely to foster a more inclusive environment
with no single set of accommodations standing out
(Colella 2001). Workers may feel that others truly
“need” the accommodations (Colella 2001), thus pro-
moting distributive fairness.
When the dispersion of workers across disability
categories is even, supervisors may more likely learn
to deal with the current challenges of others in the
team through repetition of worker assignments and
accommodations. Williams-Whitt et al. (2016) state
that supervisors learn about accommodations through
frequent exposure to them. Even dispersion of dis-
ability categories may likely promote such exposure
given the repetition of accommodations across the
range of disability categories. Such frequent exposure
may facilitate better management of worker accom-
modations and consequent worker responses.
4. Data Compilation, Variables, and
Empirical Strategy
A detailed data set was compiled drawing from
manufacturing and human resources departments at
the firm. Production data were entered by the su-
pervisors each day and compiled in spreadsheets for
each garment. These data contain the daily produc-
tion, total hours logged in by the employees, and
target hours for each garment. Further, work infor-
mation, such as the number of hours worked, worker
1119
skill level, other identifiers (e.g., leader/sewer), and
employee disability category was obtained from the
human resources department. All employee infor-
mation was masked and deidentified to maintain
confidentiality. Overall, the compiled data spanned
49 unique garments and their variants (a total of 82
garments, including color variations). The production
and human resource data were matched for each
garment, and the team composition was built from
time-stamp data to construct a panel data set that
had a total of 16,951 garment-days of data compiled
from October 2011 through June 2017. Two filtering
criteria were applied to the original data. First, some
garments for which a specific individual’s disability
status was not known were coded as missing. These
data were subsequently dropped from the analysis
even if there was one worker’s disability count missing.
Next, only teams with two sewers or more were in-
cluded in the final analyses. Instances in which the
firm engaged only one worker for a garment were
dropped from the data.
4.1. Variable Description
The key variables used as part of the analysis are
detailed as follows.
4.1.1. Dependent Variable.
4.1.1.1. Productivity (Ln(Labor Hours per Garment)). The
primary measure of productivity used in the analyses
is the log of labor hours per garment. This measure
was calculated using the hours entered by supervi-
sors in each day’s production sheet and dividing them
by the total day’s production.
4.1.2. Independent Variables. The following indepen-
dent variables are used.
4.1.2.1. Number of Disability Categories (Disability
Categories). This variable captures the range of dis-
ability categories of the workers employed in the
specific garment line. The firm classifies 10 types
of disabilities (other than those with no disability).
These are (i) mental, (ii) cognitive impairment/de-
velopmental, (iii) traumatic brain injury, (iv) learning,
(v) emotional/behavioral impairment, (vi) posttrau-
matic stress disorder, (vii) chemical dependence/
substance abuse, (viii) physical, (ix) deaf or hard of
hearing, and (x) visual impairment. Workers could
have disabilities that are primary, secondary, or ter-
tiary. The presence of each disability (irrespective of
whether the disability is primary, secondary, or ter-
tiary) is recorded to account for the presence/absence
of each disability. Next, the presence (absence) of
each disability is coded as one (zero). The number of
disability categories is a count based on whether a

1120 Manufacturing & Service Operations Management, 2020, vol. 22, no. 6, pp. 1112–1130, © 2020 INFORMS
disability category is present among any worker in
the team.
4.1.2.2. Disability Dispersion (Disability Dispersion). The
evenness of disability dispersion in the team is cap-
tured using the Blau index (Harrison and Klein 2007),
which is used in past literature to characterize and
Downloaded from informs.org by [185.107.56.83] on 13 March 2024, at 10:00 . For personal use only, all rights reserved.
measure diversity (Harrison and Klein 2007); this
∑
index is given by 1 − ​ p2j . Here, the subscript j in-
dicates the worker’s disability category. The number pj
indicates the proportional representation of disability
category j in the team relative to the total number of
disabilities recorded within the team (across primary,
secondary, and tertiary disabilities). The Blau index
was computed for each garment, each day based on
the total number of recorded disabilities across all
workers present in the team. The Blau index values
ranged from zero when the team had individuals
from only one disability category to one when the
team had no workers with disabilities (these were
relatively few).
4.1.3. Control Variables. Several control variables are
used in the analyses. These include team/worker and
garment-specific control variables. They are detailed
as follows.
4.1.3.1. Facility Dummy (Site Dummy). This variable
accounts for the facility in which the garment was
manufactured and controls for facility-specific effects.
4.1.3.2. Number of Leaders. This variable controls
for supervisory oversight (most lines had a single
supervisor). A greater number of supervisors is likely
to help better coordinate accommodations and be
responsive to workers’ needs.
4.1.3.3. Number of Workers. This variable accounts for
team size to control for the total number of workers
engaged in sewing operations.
4.1.3.4. Ratio of Workers with No Disabilities (Ratio\
Nondisabled). This variable controls for the proportion
of workers in the team with no recorded disabilities.
4.1.3.5. Average Skill Level. This variable accounts for
the average worker’s skills on the line. The workers’
sewing skills are determined during training based on
the type of stitch that they can perform in production
conditions and vary from levels 1 to 4 for any worker.
The garments manufactured have different steps,
each of which is classified as basic, medium, or
difficult. A worker assessed at a skill level 3 can
perform basic, medium, and high-difficulty tasks; a
worker assessed at skill level 2 can perform basic or
medium difficulty tasks; and a worker assessed at
Narayanan and Terris: Inclusion, Disability Diversity, and Productivity
skill level 1 can perform only basic tasks. Skill level 4
workers are considered master sewers and perform
the most complex sewing operations, such as proto-
typing new products.
4.1.4. Other Controls. Several fixed effects are used
in the analysis. To account for garment complexity,
dummy variables are used for each garment. In this
regard, some garments have multiple part numbers.
However, the only difference is the garment’s color.
Given the lack of production complexity difference,
these garments are treated the same with respect to
the inclusion of fixed effects. To account for time
variations, the year, month, and weekday of pro-
duction fixed effects are used.
4.2. Empirical Strategy
The study data cover several days for each garment
line and historical information on the production of
several garments. Accordingly, the data contain both
cross-sectional and time-series elements. Given that
the median number of days that each garment ran
was about 158, residuals may be autocorrelated.
Accordingly, the analysis approach used by Staats
and Gino (2012) was followed. This estimation ap-
proach accounts for panel-wide first order autocor-
relation and heteroskedasticity in using a Prais–Winsten
regression (XTPCSE in STATA 15.1). This also follows
other studies, such as Beck (2001). The full model is
shown in Equation (1).
Ln(Laborhoursperpart)it  β0 + β1 Disability categoriesit
+ β2 Disability categories2it
+ β3 Disability dispersionit
+ β4 Disability dispersion2it
+ β5 Site dummy
+ β6 Number of leadersit
+ β7 Number of workersit
+ β8 Ratio nondisabled it
+ β9 Average skill levelit + αi
+ λ t + εit (1)
The dependent variable for Equation (1) is produc-
tivity as described by the logged values of labor hours
per garment i in time t. Disability categories and its
square term indicate the linear and quadratic effects
of the number of disability categories among the
workers employed in the line. Disability dispersion
and its square term indicate the linear and quadratic
effects of disability dispersion. Site dummy indicates
a dummy variable for the plant in which the line
is located. Number of leaders indicates the number
of supervisory staff that oversaw the line. Number
of workers indicates the total number of sewers.
Ratio nondisabled is the ratio of workers in the line with
 Narayanan and Terris: Inclusion, Disability Diversity, and Productivity
Manufacturing & Service Operations Management, 2020, vol. 22, no. 6, pp. 1112–1130, © 2020 INFORMS 1121

no disabilities. Average skill level indicates the average
skill level of workers in the line. Finally, αi is a vector
of garment fixed effects, λt is a vector of time fixed
effects (year, month, and day of the week), and
it is
the time-varying error term.
5. Results
Downloaded from informs.org by [185.107.56.83] on 13 March 2024, at 10:00 . For personal use only, all rights reserved.
Focusing on Table 2, Model 1, the control variable
results suggest that, first, the site dummy has a sig-
nificant impact on productivity (b = −0.290, p < 0.01)
and captures the differential productivity levels across
the two facilities. This result corroborates with the in-
tuition of the managers in the firm. The total number of
supervisory staff (number of leaders) did not influence

productivity (b = −0.007, p > 0.1). Further, the more

A summary of the descriptive statistics is shown
in Table 1. Recall that productivity is defined by
labor hours per part. A higher number for labor
hours per part indicates lower productivity. A neg-
ative correlation coefficient between productivity and
any variable indicates that productivity improves
when the variable in question increases. The corre-
lation between the disability categories (disability
dispersion) employed in the line and productivity is
negative. As seen in Table 1, the correlation between
labor hours per garment and the disability cate-
gories employed in the line is −0.21. Further, the
correlation between disability dispersion and pro-
ductivity is negative (−0.20). These findings suggest
that increasing the representation of the number of
disability categories in the line and increasing the
disability dispersion in a team may improve pro-
ductivity. Finally, the ratio of nondisabled workers
in a team and the team’s average skill level are pos-
itively correlated with productivity at 0.16 and 0.60,
respectively. These suggest that a greater number of
individuals with no disability do not substantially
improve productivity.
Table 2 summarizes the estimation results. The
results of the control variables are presented in Model 1.
In Model 2, the impact of the linear terms of dis-
ability categories and disability dispersion are pre-
sented. Model 3 (Model 4) presents the nonlinear
influence of disability categories (disability disper-
sion) on productivity. Finally, the full model is pre-
sented in Model 5. Recall that the dependent variable
used in Table 2 is a logged measure of the hours per
garment. Thus, a negative coefficient indicates that
productivity is higher.
workers there were in the line (i.e., sewers), the less
productive the team was (b = 0.003, p < 0.01). Focusing
on the fraction of nondisabled workers, we find that
this fraction in the team did not significantly impact
team productivity (b = −0.026, p > 0.1). Finally, the
group’s average skill level had no significant impact
on productivity (b = 0.024, p > 0.1).
Focusing on the impact of the disability categories
on productivity, the linear effect of the disability
categories employed in the line is positive and sig-
nificant on productivity (Model 2, Table 2). When the
nonlinear term is introduced for disability categories,
the results show that the impact of the disability
categories is nonlinear and U-shaped and that the
effects remain stable in the full model (Model 5, Ta-
ble 2). As seen in Table 2, Model 5, the linear (non-
linear) impact of the disability categories involved
in the line is negative (positive) and significant: b =
−0.044, p < 0.01 (b = 0.007, p < 0.01) on productivity.
The marginal plot is shown in Figure 1. As seen in
Figure 1, productivity improves until about three cate-
gories and drops afterward with a further increase in
the disability categories.
Focusing on the results of the impact of the dis-
ability dispersion on productivity as seen in Table 2,
Model 2, the linear effect of the disability dispersion is
negative and insignificant. However, when the nonlinear
term is added, the effect of the disability dispersion on
productivity is significant and has an inverted-U shape
with the linear (nonlinear) term indicating b = 0.410, p <
0.01 (b = −0.460, p < 0.01) as seen in Table 2, Model 5.
This plot is shown in Figure 2. As seen in Figure 2, at
high levels of disability dispersion, team productivity
is higher.

Table 1. Correlations and Descriptive Statistics

Mean Standard deviation (a) (b) (c) (d) (e) (f) (g) (h)

Ln(labor hours per garment) (a) −0.56 1.45

Target achievement (b) −0.44 0.50 −0.53
Disability categories (c) 4.11 1.94 −0.21 0.03
Disability dispersion (d) 0.60 0.19 −0.20 0.01 0.68
Site dummy (e) 0.90 0.29 0.39 0.04 −0.12 −0.21
Number of leaders (f) 1.16 1.13 0.06 −0.09 0.31 0.17 0.06
Number of workers (g) 12.49 8.79 0.07 −0.03 0.66 0.35 0.01 0.51
Ratio nondisabled (h) 0.31 0.19 0.16 0.01 −0.37 −0.19 −0.01 0.02 −0.08
Average skill level (i) 2.72 0.36 0.60 −0.13 −0.35 −0.39 0.40 −0.03 −0.14 0.17

Note. Correlations are bivariate; coefficients rounded to two decimal places.

 Narayanan and Terris: Inclusion, Disability Diversity, and Productivity
1122 Manufacturing & Service Operations Management, 2020, vol. 22, no. 6, pp. 1112–1130, © 2020 INFORMS

5.1. Robustness Checks

Notes. Coefficients and standard errors rounded to three decimal places. N = 13,266. All models estimated with fixed effects corresponding to garments, year, month, and day of operation.
Standard
Several robustness checks were performed. First, an

(0.010)
(0.001)
(0.073)
(0.080)
(0.021)
(0.004)
(0.001)
(0.029)
(0.017)
(0.085)
error
additional alternative dependent variable was used
to measure productivity. Table 3 presents the impact

139,062.9
Model 5

of all key variables on the alternate productivity

0.390
0.92
measure. This alternative productivity measure is
Coefficient

−0.044***
0.007***
0.410***
−0.460***
−0.300***

1.610***
estimate

defined as target achievement (target hours/actual

−0.005
−0.000
0.023
0.014
Downloaded from informs.org by [185.107.56.83] on 13 March 2024, at 10:00 . For personal use only, all rights reserved.

hours per garment). The plant uses this measure to

evaluate teams, and the team bonus is paid against
productivity achievement. These measures are based
on internal productivity targets. This measure is
Standard

used as an alternative because hours per garment is

(0.064)
(0.070)
(0.021)
(0.004)
(0.001)
(0.025)
(0.017)
(0.085)
error

considered as a direct productivity metric. For ex-

ample, as seen in Table 1, the correlation between the
134,995.2
Model 4

labor hours per garment and target achievement

0.395
0.92

is −0.53. Note that the target achievement mea-

Coefficient

−0.290***

0.003***

1.540***
estimate

sure implies that higher achievement is better com-

0.140**
−0.130*

−0.008*

−0.009
0.021

pared with the primary productivity measure of la-

bor hours per garment. Thus, the sign of the coefficients
is expected to be the opposite. As seen in Table 3, the
effects of both disability categories and disabil-
Table 2. Impact of Disability Categories and Disability Dispersion on Labor Hours per Garment (Logged)

Standard

ity dispersion continue to be similar and in the ex-

(0.009)
(0.001)

(0.021)
(0.004)
(0.001)
(0.026)
(0.016)
(0.085)
error

pected direction. Specifically, a moderate level of

disability categories in a line is associated with the
137,453.8
Model 3

highest target achievement.

0.393
0.92

Second, garments that were in production for

Coefficient

−0.024***
0.004***

−0.300***

1.620***
estimate

fewer than 60 and 90 days were dropped. This was

0.029*
−0.004
0.001
−0.024

done with the idea of eliminating the effects of

garments that were in operation for a smaller num-
ber of days. The results are identical (see Table 4).
Third, least square dummy variable (LSDV), fixed-
Standard

effect regressions were used. These results, presented

(0.004)

(0.024)

(0.021)
(0.004)
(0.001)
(0.026)
(0.016)
(0.084)
error

in Table 4, are consistent.

Fourth, identification in this setting is challenging.
135,885.3
Model 2

For example, to capture data on why individuals

0.394
0.92

may move across stations along with other allocation

Coefficient

0.014***

−0.300***

1.550***

nuances that may happen regularly within the line is

estimate

−0.007*
−0.014

0.001
0.012
0.024

costly. Hartmann et al. (2008, p. 295) note that, in a

setting in which instruments are not available, “[a]n
alternative exclusion restriction imposes a temporal
ordering.” Other studies, such as Nam et al. (2010)
follow such an approach as well. In this setting,
Standard

(0.021)
(0.004)
(0.001)
(0.023)
(0.016)
(0.083)
error

employee diversity for each day is likely to influence

the productivity on that day. However, changes in
135,317.3
Model 1

0.395

the team structure made on the previous day may

0.92

have an impact on productivity the following day

*p < 0.10; **p < 0.05; ***p < 0.01.
Coefficient

−0.290***

0.003***

1.570***
estimate

because employees may be adjusting to a new rou-

−0.007

−0.026
0.024

tine. Thus, key diversity variables are lagged by one

and two days, respectively. The results continued to be
the same.
Variable description

Disability dispersion2
Disability categories2

Fifth, to account for any past changes, lagged

Disability dispersion
Disability categories

Wald chi-squared
Number of workers
Number of leaders

Average skill level

Ratio nondisabled

productivity was used as a control variable, that

is, productivity of the line in the previous day. The
Site dummy

Constant

results continued to hold. Finally, only groups that

had more than three and four workers, respec-
R2

tively, are included in the analysis to examine the

ρ
 Narayanan and Terris: Inclusion, Disability Diversity, and Productivity
Manufacturing & Service Operations Management, 2020, vol. 22, no. 6, pp. 1112–1130, © 2020 INFORMS
Figure 1. (Color online) Predictive Plot of Disability
Categories and Line Productivity
Downloaded from informs.org by [185.107.56.83] on 13 March 2024, at 10:00 . For personal use only, all rights reserved.
sensitivity of results to variations in team size. These
results also continue to be the same. In the interest of
brevity, these results are not presented.
6. Discussion of Implications
Disability diversity is an important and understudied
topic in the current literature (Shore et al. 2009, Boehm
and Dwertmann 2015). With the passing of the ADA
and the Workforce Innovation Opportunities Act, the
need to understand the influence of employing in-
dividuals with disabilities within the United States
has substantially grown, yet empirical work in this space
is sparse (Shore et al. 2009, Boehm and Dwertmann
2015). The salience of studying this topic also arises
from several contextual factors that societal envi-
ronments (and, by extension, firms) face today. For
example, the United Nations articulates inclusion
as a key facet of its SDGs (United Nations 2015) and
pledges to leave no one behind, particularly indi-
viduals with disabilities in the workplace (UCLG 2018,
Figure 2. (Color online) Predictive Plot of Disability
Dispersion and Line Productivity
1123
Mohammed 2019). The aging workforce also exac-
erbates the urgency of this challenge in Western
economies (World Health Organization 2011). There
is a substantial correlation between age and disability
(World Health Organization 2011). Firms, such as
BMW, are increasingly looking to employ an aging
workforce (Loch et al. 2010). In Europe, individuals
with disabilities continue to face low employment
rates (Shima et al. 2008). Several European countries,
such as Austria, Germany, and Spain, have instituted
quota systems for firms to employ individuals with
disabilities to increase worker participation (Shima
et al. 2008, Boehm and Dwertmann 2015). Yet sub-
stantial gaps remain in disability employment across
several countries of the world, including the United
States. A key factor among them involves understanding
the productivity implications of employing individuals
with disabilities. This study takes a first exploratory step
in this direction by examining the issue of disability
diversity and its implications for team productivity
outcomes. Productivity is particularly important to study
because the economics of a product are also determined
by manufacturing productivity.
Within the United States, policymakers increas-
ingly favor of the concept of competitive integrated
employment (Novak 2015). This contrasts with other
programs, such as the 14(c), in which employers can
pay below minimum wage if serious disabilities prevent
individuals from performing any work (Hoffman 2013).
Most recently, in 2011, the Fair Labor Standards Act
was passed, which disallows employers from offering
subminimum wages (Hoffman 2013). The passage of
such acts has ensured that firms (including many
erstwhile sheltered workshops) initiate a move to
CIE. In the CIE model, individuals with and without
disabilities work side by side with no differences in
wages (Novak 2015). This issue is nontrivial and
difficult to execute as it emphasizes the importance of
cultivating a diverse workforce with disabilities and
building operational strengths to build not only better
processes, but also better employee support systems.
Thus, WISE firms can continue to become commer-
cially viable as social enterprises. As a result, strat-
egies to improve productivity within the confines of
the organizational mission are valuable in improving
an organization’s commercial viability.
This study focuses on a WISE that pursues the
model of CIE. Within Peckham, Inc., individuals with
and without disabilities work side by side. One ap-
proach to such a model is to cultivate employees’
strengths to serve individuals across a range of dis-
ability categories. The firm’s commitment to employ-
ing individuals across disability categories enables
it to be more competitive, according to interactions
with the firm’s management. The pursuit of CIE also
raises the importance of studying disability diversity
 Narayanan and Terris: Inclusion, Disability Diversity, and Productivity
1124 Manufacturing & Service Operations Management, 2020, vol. 22, no. 6, pp. 1112–1130, © 2020 INFORMS

Notes. Coefficients and standard errors rounded to three decimal places. N = 13,224. All models estimated with fixed effects corresponding to garments, year, month, and day of operation.
and its implications within operations. Against this

Standard

(0.010)
(0.001)
(0.073)
(0.080)
(0.021)
(0.004)
(0.001)
(0.028)
(0.017)
(0.085)
backdrop, this study empirically examines the im-

error
pact of disability diversity—as measured by the
disability categories and disability dispersion employed
Model 5

9,336.1
in a line—on team productivity.

0.374
0.46
In this context, the results of this study suggest that
Coefficient

0.036***
−0.005***
−0.330***
0.380***
0.290***

−0.003***
estimate

−1.46***
having moderate levels of disability categories in a

−0.001

0.044
0.010
Downloaded from informs.org by [185.107.56.83] on 13 March 2024, at 10:00 . For personal use only, all rights reserved.

team is associated with better productivity. In ad-

dition, the exploratory analysis also suggests that a
higher level of disability dispersion helps improve
team productivity. It is likely that, on the one hand,
Standard

increasing the disability categories in a line may

(0.064)
(0.070)
(0.022)
(0.004)
(0.001)
(0.025)
(0.017)
(0.084)
error

provide management some flexibility in managing

employee allocations and improving the line flow.
Model 4

9,294.7 On the other hand, increasing disability dispersion

0.376
0.46

may help better manage the accommodations within

Coefficient

0.280***

−0.005***
estimate

−0.130**

0.063**

a line. Such an approach to managing the allocation

−1.40***
0.130*

0.001

0.005

of employees with disabilities may strike a balance

between getting the best out of an individual’s abili-
ties by providing allocation flexibility to managers.
Moreover, it may likely cause fewer conflicts as the
Standard

presence of multiple individuals within the same

(0.009)
(0.001)

(0.021)
(0.004)
(0.001)
(0.026)
(0.016)
(0.084)
error

disability category may provide a reasonable safety

net for the individuals within the group for advocacy
Model 3

9,282.9

(cf., von Schrader et al. 2014). Furthermore, evenly

0.376
0.46

dispersing individual disability categories within a

Table 3. Impact of Disability Categories and Disability Dispersion on Target Achievement

Coefficient

−0.003***

0.290***

−0.003***
0.082***
estimate

line may help supervisors learn better about specific

0.019**

−1.47***
−0.002

−0.003

employees and their skill sets.

A caveat to the results is that the individuals who
are assigned to the apparel-sewing tasks are trained
and deemed capable of performing them. Despite the
Standard

early determination of employee ability, the orga-

(0.004)

(0.024)

(0.021)
(0.004)
(0.001)
(0.025)
(0.016)
(0.084)
error

nization also focuses on managing and actively

supporting the employees in the line. To better un-
derstand the firm’s approaches to allocation flexi-
Model 2

9,291.0
0.376
0.46

bility on the line and accommodations, some anec-

Coefficient

0.290***

−0.004***

dotal supervisor observations were obtained. One

estimate

−0.009**

0.056**

−1.42***
0.012

0.000

0.001

supervisor stated, “Someone not good with body

functions, we place on easier jobs like bartack [stitches
used to reinforce areas of garment that are heavy load
bearing] or trimming threads. If [the sewer is] older,
Standard

they need an easier job too.” Similarly, another su-

(0.021)
(0.004)
(0.001)
(0.023)
(0.016)
(0.082)
error

pervisor noted, “Depending on whether it [the dis-

ability] is a physical or mental challenge, we go by
Model 1

9,277.4
0.377
0.46

the information provided to us during intake. It is

important for us to look at possible triggers they
*p < 0.10; **p < 0.05; ***p < 0.01.
Coefficient

0.280***

−0.005***
0.080***
estimate

−1.43***

may have. Are they triggered by loud noises? Give

0.000

0.000

them earplugs or put them in a line of business with

a quieter environment. Do they require a low-stress
environment? Put them in a line of business with
Variable description

Disability dispersion2
Disability categories2
Disability dispersion

less people. For physical [restrictions], we look at

Disability categories

Wald chi-squared
Number of workers
Number of leaders

Average skill level

Ratio nondisabled

something like lift restrictions or movement re-

strictions. Communication [of their needs to us]
Site dummy

Constant

is paramount.”
Similarly, a third supervisor noted, “Trial and
R2

error was the best way to deal with the different

ρ
 Downloaded from informs.org by [185.107.56.83] on 13 March 2024, at 10:00 . For personal use only, all rights reserved.

Table 4. Robustness Checks on the Impact of Disability Categories and Disability Dispersion on Productivity Performance

Drop all teams less than 60 days Drop all teams less than 90 days Fixed effects LSDV regression

Dependent variable Labor hours per part Target achievement Labor hours per part Target achievement Labor hours per part Target achievement

Coefficient Standard Coefficient Standard Coefficient Standard Coefficient Standard Coefficient Standard Coefficient Standard
Variable description estimate error estimate error estimate error estimate error estimate error estimate error

Disability categories −0.046*** (0.010) 0.037*** (0.010) −0.047*** (0.010) 0.038*** (0.010) −0.055*** (0.009) 0.044*** (0.009)
Disability categories2 0.007*** (0.001) −0.005*** (0.001) 0.007*** (0.001) −0.005*** (0.001) 0.008*** (0.001) −0.006*** (0.001)
Disability dispersion 0.400*** (0.075) −0.330*** (0.075) 0.410*** (0.075) −0.330*** (0.075) 0.480*** (0.082) −0.360*** (0.082)
Narayanan and Terris: Inclusion, Disability Diversity, and Productivity

Disability dispersion2 −0.450*** (0.083) 0.370*** (0.082) −0.450*** (0.083) 0.370*** (0.082) −0.580*** (0.089) 0.440*** (0.088)
Site dummy −0.300*** (0.021) 0.290*** (0.022) −0.300*** (0.021) 0.280*** (0.022) −0.260*** (0.017) 0.270*** (0.017)
Number of leader −0.004 (0.004) −0.002 (0.004) −0.005 (0.004) −0.001 (0.004) −0.005 (0.005) −0.001 (0.004)
Number of workers −0.000 (0.001) −0.002*** (0.001) −0.000 (0.001) −0.003*** (0.001) −0.001 (0.001) −0.002*** (0.001)
Ratio nondisabled 0.027 (0.029) 0.040 (0.029) 0.013 (0.030) 0.055* (0.029) 0.010 (0.028) 0.072*** (0.028)
Average skill level 0.017 (0.017) 0.008 (0.017) 0.009 (0.017) 0.015 (0.017) 0.031* (0.016) 0.004 (0.016)
Time trend n/a n/a n/a n/a −0.000*** (0.000) 0.000*** (0.000)
Constant 1.60*** (0.086) −1.45*** (0.086) 1.64*** (0.087) −1.49*** (0.086) 1.53*** (0.076) −1.45*** (0.076)
Manufacturing & Service Operations Management, 2020, vol. 22, no. 6, pp. 1112–1130, © 2020 INFORMS

N 12,891 12,850 12,657 12,618 13,266 13,224

R2 0.91 0.45 0.91 0.44 0.943 0.545
ρ 0.396 0.380 0.397 0.379 n/a n/a
Chi-squared (C)/(AIC) 129,319.6 (C) 8,637.6 (C) 125,990.4 (C) 8,235.7 (C) 9,579.945 (AIC) 8,932.714 (AIC)

Notes. Coefficients and standard errors rounded to three decimal places. N = 13,266 for Ln(Labor hours per part). N = 13,224 for Target achievement. All models estimated with fixed effects
corresponding to garments, year, month, and day of operation. (C), chi-squared; AIC, Akaike information criterion.
*p < 0.10; **p < 0.05; ***p < 0.01.
1125

1126 Manufacturing & Service Operations Management, 2020, vol. 22, no. 6, pp. 1112–1130, © 2020 INFORMS
disabilities I encountered on the floor. As supervisors, we
are not privy to disability detail; much of it was trial and
error as well as observation. If I noticed an individual
struggling to get through [product A] because the ma-
terial was too heavy/scratchy feeling, I would move
them to something like [product B] or [product C] and
see how they reacted.” In another task reallocation
Downloaded from informs.org by [185.107.56.83] on 13 March 2024, at 10:00 . For personal use only, all rights reserved.
example, a supervisor stated the following: “Know-
ing that I am in the same boat helps. One example is on
[product C]. . . . I knew that he had a stent in his leg
and some mental health issues, so I tried my best to be
as patient as possible. Soon, we had the mechanic
move the machine pedal over, and it worked out
great for him.” These quotes illustrate the challenge
of addressing issues that require the thoughtfulness
of actions in the line and, sometimes, substantial trial
and error in allocating employees. Although these
examples show that the managerial approach toward
resource allocation on the line is likely to help im-
prove team productivity, it was not empirically
possible to capture this within the data collected and
is also a limitation of the empirical analysis.
Also, informal discussions were pursued concerning
the need to support individuals with a diversity of
disabilities in two other not-for-profit organizations:
one based in Alabama and another based in Michigan.
Both are AbilityOne-affiliated firms that specialize in
employing individuals with disabilities and in pur-
suing different forms of apparel production. Both
organizations are substantially smaller than Peckham,
Inc. The executive from the Alabama-based not-for-
profit organization noted that providing support services
to a broad spectrum of disabilities at the organization
enhances their ability to carve jobs in comparison with
when individuals with only specific disabilities are
employed. Additionally, this executive noted that indi-
viduals with learning challenges (although not good at
multitasking) are adept at handling specialized tasks.
However, the executive also noted that there are limits to
their ability to carve out jobs within their setting.
Similarly, the CEO of the Michigan-based firm
noted that, if everybody in the line has physical
challenges, it might be difficult to manage the entire
line from an output standpoint. Task-level execution
challenges require them to augment the labor force
with other, more skilled employees with different (or
no) disabilities, depending on the contract specifi-
cations. The inclusion of individuals with no dis-
abilities was also referred to by the CEO as “reverse
inclusion.” On the behavioral side, the executive
also noted that specializing in serving individuals
with a specific disability could also result in indi-
viduals within the organization developing assump-
tions about accommodations, which may not be ef-
fective. In general, these observations corroborate the
Narayanan and Terris: Inclusion, Disability Diversity, and Productivity
argument of diversifying individuals across disability
categories within operations.
From an empirical standpoint, both disability cat-
egories and disability dispersion are linked. To better
understand this, another analysis was conducted to
separate disability categories and disability disper-
sion into tertiles. Specifically, interactions were run
between the three tertiles of the disability categories
and the three tertiles of disability dispersion for ease
of interpretation. These results are presented in Ta-
ble 5. The marginal plot of the interactions is shown in
Figure 3. As can be seen in these marginal plots, along
with an increasing number of disability categories,
increasing disability dispersion positively impacts
productivity or reduces labor hours per part. Spe-
cifically, the marginal plots suggest that productivity
is greater when a higher number of disability cate-
gories within the line is accompanied by greater
disability dispersion. Furthermore, the highest av-
erage productivity occurs at the moderate level of
disability categories and a high level of disability
dispersion. The interactions are significant and rep-
licable for continuous variable operationalizations as
well. These analyses provide evidence of the im-
portance of managing disability diversity within a
production line to improve productivity.
More generally, this study contributes to the area of
WISEs (Ebrahim et al. 2014, Battilana et al. 2015). The
importance of organizations that have the social
mission of providing opportunities to “the disabled
with finding long-term employment or transitioning
into the open labor market” (Hoffman 2013, p. 151)
cannot be underemphasized. Such organizations exist
all over the world, and Peckham, Inc., is an example.
The primary challenge for such organizations is to
bridge the efficiency gap to better leverage the existing
workforce and manage its internal processes while
achieving its core mission. Studying internal processes
and constraints within these WISE organizations may
suggest ways in which firms can reach their efficiency
goals given social mission constraints.
This study sheds light on how other WISE orga-
nizations can improve performance by potentially
diversifying employment across a range of disabil-
ities to achieve productivity goals. In this sense, the
study also raises additional questions for advancing
the current understanding around the boundaries of
research on people-centric operations (e.g., KC et al.
2020). Beyond WISEs, some commercial organiza-
tions may find ideas about disability diversity em-
phasized in this study useful. However, it may be
difficult to compare WISEs to traditional organiza-
tions from a productivity growth standpoint or from
the standpoint of organization’s mission, that is, the
mission of serving individuals with disabilities.
 Narayanan and Terris: Inclusion, Disability Diversity, and Productivity
Manufacturing & Service Operations Management, 2020, vol. 22, no. 6, pp. 1112–1130, © 2020 INFORMS 1127

Table 5. Robustness: Regression of Tertiles of Disability Category and Disability Dispersion and Their Impact on
Productivity Performance

Ln(Labor hours per part) Target achievement

Variable description Coefficient estimate Standard error Coefficient estimate Standard error

Disability categories (T2) −0.043** (0.022) 0.047*** (0.018)

Disability categories (T3) −0.190*** (0.058) 0.110** (0.051)
Downloaded from informs.org by [185.107.56.83] on 13 March 2024, at 10:00 . For personal use only, all rights reserved.

Disability dispersion (T2) −0.034*** (0.012) 0.026** (0.012)

Disability dispersion (T3) −0.058* (0.030) 0.060** (0.030)
Disability categories (T2) × Disability dispersion (T2) 0.056** (0.024) −0.055*** (0.021)
Disability categories (T2) × Disability dispersion (T3) 0.100*** (0.037) −0.110*** (0.035)
Disability categories (T3) × Disability dispersion (T2) 0.170*** (0.059) −0.087* (0.052)
Disability categories (T3) × Disability dispersion (T3) 0.250*** (0.065) −0.160*** (0.059)
Site dummy −0.300*** (0.021) 0.290*** (0.021)
Number of leaders −0.005 (0.004) −0.001 (0.004)
Number of workers 0.003*** (0.001) −0.005*** (0.001)
Ratio nondisabled −0.011 (0.027) 0.066** (0.027)
Average skill level 0.018 (0.017) 0.008 (0.017)
Constant 1.64*** (0.084) −1.48*** (0.083)
R2 0.92 0.46
ρ 0.390 0.372
Wald chi-squared 137,805.1 9,442.1

Notes. Coefficients and standard errors rounded to three decimal places. T1, first tertile (baseline); T2, second tertile; T3, third tertile. N = 13,266
for Ln(Labor hours per part). N = 13,224 for Target achievement. All models have fixed effects corresponding to garments, year, month, and day of
operation.
*p < 0.10; **p < 0.05; ***p < 0.01.

6.1. Discussion of the Key Limitations
This research focuses on understanding the impli-
cations of disability diversity on productivity using
microdata in an apparel-manufacturing context. A
key finding of this study is that having moderate
levels of disability categories on a team with a higher
level of evenness in the disability category dispersion
is associated with better productivity rather than having
a concentrated team focused on specific disabilities.
Although this research offers an important finding, it
also offers multiple limitations. The most important
being generalizability. These are detailed as follows.
Figure 3. (Color online) Predictive Plot of Interaction
Between Tertiles of Disability Categories and Tertiles of
Disability Dispersion
First, this study is restricted by the context of a
single firm in a repetitive task context. Although some
of the results may apply to other team settings, these
may not necessarily carry forward to other settings
in which individuals with disabilities work on in-
dependent tasks. Examples of such contexts could be
service settings, such as call centers, and manufacturing
settings other than apparel in which individuals work
on their own with minimal coordination with others.
These need separate investigations. However, nar-
rowing the context also provides us with the benefit of
collecting objective data that are contextual. More
contextual studies may build collective knowledge in
this domain. Prior studies in the context of disabilities
and productivity have not controlled or examined
such a context (Lindsay et al. 2018).
Second, the study context is a not-for-profit WISE
setting. This may limit the generalizability of the
findings. In WISEs, such as Peckham, Inc., the pri-
mary priority is on achieving a social mission. Thus,
profitability is only important in as much as sus-
taining the organization to accomplish the goal of
work integration. For-profit firms may or may not
share such a mindset.
Third, the costs of supporting employees with
disabilities also need to be kept in mind. These costs
may be felt at both the organizational and process
levels. One element is the cost of accommodations
(Stamps 1995). Studies note that these fears have
largely subsided in recent times with some re-
search finding that accommodations are cost-efficient

1128 Manufacturing & Service Operations Management, 2020, vol. 22, no. 6, pp. 1112–1130, © 2020 INFORMS
(Cimera 2000). Hernandez and McDonald (2010, p.
15) conclude that “costs for accommodations for the
most part are minimal.” Yet, when employing indi-
viduals with severe disabilities, studies have not been
conclusive, as the problem is challenging. Nonethe-
less, it is important to remember that the cost of ad-
ministering operations is just one part of the overall
Downloaded from informs.org by [185.107.56.83] on 13 March 2024, at 10:00 . For personal use only, all rights reserved.
cost structure in a social enterprise of this kind. The
costs in these settings need to be evaluated from a
holistic (societal, governmental, and consumer) per-
spective (Clark et al. 1998).
Studies that have taken a collective perspective on
cost-efficacy of support employment programs for
specific disabilities find both positive and negative
sides of the overall costs for such programs (Cook and
Burke 2002). Thus, the interpretation of some of the
insights needs to be tempered with higher-level social
considerations. This point was not lost among indi-
viduals within the not-for-profit firms with whom we
interacted as well as within Peckham, Inc. Consid-
ering this, lower productivity, in and of itself, is not a
substantial issue; rather, it requires a different eval-
uation lens that considers societal inclusion. Overall,
this issue is beyond the scope of the current work and
borders on social and ethical considerations regard-
ing employing individuals with disabilities. Yet the
empirical results of this study suggest that higher
disability diversity can result in higher productivity.
This finding can be useful for managers within not-
for-profit firms that specialize in employing indi-
viduals with disabilities.
In addition to generalizability, another key limi-
tation of this study is the lack of visibility into indi-
vidual accommodations over a larger time because of
several practical considerations. Within the rehabil-
itation literature, there is a substantial focus on the
role of accommodations (Zuckerman 1993, Stamps
1995). From an operational perspective, it may be
important to focus on task redesign and approaches
to make the processes dynamic in terms of matching
the individual’s environment. The CEO of one of the
organizations we talked to noted that “the concept is
to get the input from the person, match the input to the
person’s [specific] needs [with reference to the job].”
In practice, it is quite challenging. This is attributed to
the varying range of disability acuity levels as well as
the changing nature of acuity levels for disabilities.
These call for an organization to build additional
processes, training plans, and the construction of pro-
grammed instruction approaches to help employees
accomplish their work effectively.
Furthermore, when individuals work as a team,
there are additional considerations of task handoffs.
Such considerations also call for a better understanding
of the underlying parameters of the person–job fit and
its connection to task analysis in both individual and
Narayanan and Terris: Inclusion, Disability Diversity, and Productivity
team contexts. Specifically, dimensions of person–job
fit articulated in the seminal approaches of Hackman
and Oldham (1975) can be augmented with opera-
tional considerations that consider individual ability.
Models of task analysis at a microlevel, a key tenet of
time and motion studies (Mithaug 1979), in the do-
mains of operations management can, thus, com-
plement the traditional job–fit approaches.
Finally, this study does not focus on the mecha-
nisms of team “formation” involving workers with
disabilities despite documenting the impact of di-
versity. Studying team formation would require con-
sideration of both demand and supply factors given
the unique model of firms such as Peckham, Inc. The
mandate in this setting was to have at least 75% of
hours logged by individuals with disabilities. Such
mandates are also applied to other similar not-for-
profit organizations. Thus, employee availability and
exit and the need to maintain a mandated ratio may
also drive team formation in addition to the capacity
available to train and support individuals with specific
disabilities. However, the results of this study do
suggest that organizations are better off supporting
individuals across a range of disabilities from a pro-
ductivity standpoint, likely providing a social case
for inclusion.
Acknowledgments
The authors acknowledge feedback from Ann Gilmore,
corporate vice president, Peckham, Inc.; Sarah George, di-
rector of mission initiatives; and Amy Seabold and Alex
Casamatta for their help at various stages of data collection in
addition to Han Wang and Purva Shanker for helpful work in
the early stages of data collection. The authors also thank the
Peckham senior leadership for their encouragement and
support of the project. This study would not be possible
without their valuable help and input. The authors also ac-
knowledge the questions and input from Participants at the
Source America Annual Conference and the Manufacturing
and Service Operations Management Annual Conference at
the University of North Carolina at Chapel Hill. They also
thank Shawnee Vickery, Dustin Cole, Kalyanmoy Deb, and
Tharo Soun for their input in prior versions of this document.
References
AbilityOne (2019) AbilityOne program fact sheet. Accessed June 15,
2019, https://www.abilityone.gov/media_room/documents/2018
_AbilityOne_Facts.pdf.
Akşin Z, Deo S, Jónasson JO, Ramdas K (2020) Learning from many:
Partner exposure and team familiarity in fluid teams. Manage-
ment Sci., ePub ahead of print June 25, https://doi.org/10.1287/
mnsc.2019.3576.
Allen J, Cohen N (2010) The Road to Inclusion: Integrating People with
Disabilities into the Workplace (Deloitte, Toronto).
Allison R, Hyatt J, Owens L, Clark K, Test D (2017) Competitive
integrated employment (CIE) toolkit. Accessed March 18, 2020,
https://files.eric.ed.gov/fulltext/ED601255.pdf.
Barnes RM (1949) Motion and Time Study (Wiley, Hoboken, NJ).
 Narayanan and Terris: Inclusion, Disability Diversity, and Productivity
Manufacturing & Service Operations Management, 2020, vol. 22, no. 6, pp. 1112–1130, © 2020 INFORMS
Bartholdi JJ, Eisenstein DD (1996) A production line that balances
itself. Oper. Res. 44(1):21–34.
Battilana J, Sengul M, Pache A-C, Model J (2015) Harnessing
productive tensions in hybrid organizations: The case of work
integration social enterprises. Acad. Management J. 58(6):
1658–1685.
Beck N (2001) Time-series–cross-section data: What have we learned
in the past few years? Annual Rev. Political Sci. 4:271–293.
Downloaded from informs.org by [185.107.56.83] on 13 March 2024, at 10:00 . For personal use only, all rights reserved.
Berg P, Appelbaum E, Bailey T, Kalleberg AL (1996) The performance
effects of modular production in the apparel industry. Indust.
Relations 35(3):356–373.
Black JT, Kohser RA (2011) DeGarmo’s Materials and Processes in
Manufacturing (Wiley, Hoboken, NJ).
Boehm SA, Dwertmann DJ (2015) Forging a single-edged sword: Fa-
cilitating positive age and disability diversity effects in the
workplace through leadership, positive climates, and HR prac-
tices. Work Aging Retirement 1(1):41–63.
Bureau of Labor Statistics (2017) Employment status of the civilian
noninstitutional population by disability status and selected
characteristics. Accessed August 11, 2018, https://www.bls.gov/
news.release/disabl.t01.htm.
Cherns A (1976) The principles of sociotechnical design. Human
Relations 29(8):783–792.
Cimera RE (2000) The cost-efficiency of supported employment pro-
grams: A literature review. J. Vocational Rehabilitation 14(1):51–61.
Clark RE, Xie H, Becker DR, Drake RE (1998) Benefits and costs of
supported employment from three perspectives. J. Behav. Health
Services Res. 25(1):22–34.
Colella A (2001) Coworker distributive fairness judgments of the
workplace accommodation of employees with disabilities. Acad.
Management Rev. 26(1):100–116.
Colella A, Paetzold R, Belliveau MA (2004) Factors affecting coworkers
procedural justice inferences of the workplace accommodations of
employees with disabilities. Personnel Psych. 57(1):1–23.
Cook JA, Burke J (2002) Public policy and employment of people
with disabilities: Exploring new paradigms. Behav. Sci. Law
20(6):541–557.
Cooney K (2016) Work integration social enterprises in the United
States: Operating at the nexus of public policy, markets, and
community. Nonprofit Policy Forum 7(4):435–460.
Dunn DS, Burcaw S (2013) Disability identity: Exploring narrative
accounts of disability. Rehabilitation Psych. 58(2):148–157.
Ebrahim A, Battilana J, Mair J (2014) The governance of social en-
terprises: Mission drift and accountability challenges in hybrid
organizations. Res. Organ. Behav. 34:81–100.
Economist, The (2012) The new green: Disability and business.
The Economist (September 8), https://www.economist.com/
international/2012/09/08/the-new-green.
Elkind J (1990) The incidence of disabilities in the United States.
Human Factors 32(4):397–405.
Erickson W, Lee C, von Schrader S (2017) Disability Statistics from
the American Community Survey (ACS) (Cornell University,
Ithaca, NY).
Fesko S, Varney E, DiBiase C, Hippenstiel M (2008) Effective part-
nerships: Collaborative efforts that support customized em-
ployment. J. Vocational Rehabilitation 28(3):159–168.
Ford H, Crowther S (1922) My Life and Work (Doubleday, Garden
City, NY).
Garnero A, Kampelmann S, Rycx F (2014) The heterogeneous effects
of workforce diversity on productivity, wages, and profits.
Indust. Relations 53(3):430–477.
Gaudiano P, Hunt E (2017) 10 tips for unlocking the talent and
potential of employees with disabilities. Forbes (June 12), https://
www.forbes.com/sites/gaudianohunt/2017/06/12/workplace
-disability-tips/#4743e784df47.
1129
Green Economy Coalition (2019) Inclusive workplaces in the garment
industry. Accessed March 18, 2020, https://goodmenproject.com/
business-ethics-2/inclusive-workplaces-in-the-garment-industry/.
Griffin C, Hammis D, Geary T (2007) The Job Developer’s Handbook:
Practical Tactics for Customized Employment (Brookes Publish-
ing, Baltimore).
Hackman JR, Oldham GR (1975) Development of the job diagnostic
survey. J. Appl. Psych. 60(2):159–170.
Hamilton BH, Nickerson JA, Owan H (2003) Team incentives and
worker heterogeneity: An empirical analysis of the impact of
teams on productivity and participation. J. Political Econom.
111(3):465–497.
Hamilton BH, Nickerson JA, Owan H (2012) Diversity and Pro-
ductivity in Production Teams, Advances in the Economic Analysis
of Participatory and Labor-Managed Firms (Emerald Group
Publishing Limited, Bingley, UK).
Harrison DA, Klein KJ (2007) What’s the difference? Diversity con-
structs as separation, variety, or disparity in organizations. Acad.
Management Rev. 32(4):1199–1228.
Hartmann WR, Manchanda P, Nair H, Bothner M, Dodds P, Godes D,
Hosanagar K, Tucker C (2008) Modeling social interactions:
Identification, empirical methods and policy implications. Mar-
keting Lett. 19(2–3):287–304.
Hernandez B, McDonald K (2010) Exploring the costs and benefits of
workers with disabilities. J. Rehabilitation 76(3):15–23.
Ho J (2018) Exploring the job carving process for individuals
with disabilities: A thematic analysis of the facilitators and
hindrances. Unpublished master’s thesis, University of Ottawa,
Ottawa.
Hoffman LC (2013) An employment opportunity or a discrimination
dilemma: Sheltered workshops and the employment of the
disabled. Univ. Pennsylvania J. Law Soc. Change 16(2):151–179.
Huckman RS, Staats BR (2011) Fluid tasks and fluid teams: The impact
of diversity in experience and team familiarity on team perfor-
mance. Manufacturing Service Oper. Management 13(3):310–328.
InWork (2020) Work integration social enterprise. Accessed January 1,
2020, http://www.inworkproject.eu/toolbox/index.php/glossary
-resources/glossary/work-integration-social-enterprise.
KC D, Staats BR (2012) Accumulating a portfolio of experience: The
effect of focal and related experience on surgeon performance.
Manufacturing Service Oper. Management 14(4):618–633.
KC DS, Staats BR, Kouchaki M, Gino F (2020) Task selection and
workload: A focus on completing easy tasks hurts performance.
Management Sci., ePub ahead of print May 20, https://doi.org/
10.1287/mnsc.2019.3419.
Kennedy T, Gaul-Stigge K (2019) Recognizing the value of disabled
workers. Wall Street Journal (February 4), https://www.wsj.com/
articles/recognizing-the-value-of-disabled-workers-11549321435.
Kerlin JA (2006) Social enterprise in the United States and Europe:
Understanding and learning from the differences. Voluntas
17(3):246.
Kolenc V (2017) El Paso garment factory hiring 500 people. El
Paso Times (October 11), https://www.elpasotimes.com/story/
money/business/2017/10/11/el-paso-garment-factory-jobs
-disabled-workers/751946001/.
Lindsay S, Cagliostro E, Albarico M, Mortaji N, Karon L (2018) A
systematic review of the benefits of hiring people with disabil-
ities. J. Occupational Rehabilitation 28(4):634–655.
Loch C, Sting F, Bauer N, Mauermann H (2010) How BMW is defusing
the demographic time bomb. Harvard Bus. Rev. 88(3):99–102.
Mallender J, Liger Q, Tierney R, Beresford D, Eager J, Speckesser S,
Nafilyan V (2015) Reasonable accommodation and sheltered
workshops for people with disabilities: costs and returns of
investments. Accessed May 21, 2020, https://www.europarl
.europa.eu/RegData/etudes/STUD/2015/536295/IPOL_STU(2015)
536295_EN.pdf.

1130 Manufacturing & Service Operations Management, 2020, vol. 22, no. 6, pp. 1112–1130, © 2020 INFORMS
Martz E, Xu YJ (2008) Person-related and service-related factors pre-
dicting employment of individuals with disabilities. J. Vocational
Rehabilitation 28(2):97–104.
McClain JO, Schultz KL, Thomas LJ (2000) Management of
worksharing systems. Manufacturing Service Oper. Management
2(1):49–67.
Migliore A, Grossi T, Mank D, Rogan P (2008) Why do adults with
intellectual disabilities work in sheltered workshops? J. Voca-
Downloaded from informs.org by [185.107.56.83] on 13 March 2024, at 10:00 . For personal use only, all rights reserved.
tional Rehabilitation 26(1):29–40.
Miralles C, Garcia-Sabater JP, Andres C, Cardos M (2007) Advantages
of assembly lines in sheltered work centres for disabled. A case
study. Internat. J. Production Econom. 110(1):187–197.
Mithaug DE (1979) The relation between programmed instruction
and task analysis in the prevocational training of severely and
profoundly handicapped persons. AAESPH Rev. 4(2):162–178.
Mohammed A (2019) Opinion: Enabling the SDGs through disability
inclusion. Accessed November 10, 2019, https://www.devex
.com/news/opinion-enabling-the-sdgs-through-disability
-inclusion-94288.
Nam S, Manchanda P, Chintagunta PK (2010) The effect of signal
quality and contiguous word of mouth on customer acquisition
for a video-on-demand service. Marketing Sci. 29(4):690–700.
Narayanan S, Terris E (2020) Conversation with a WISE executive on
improving opportunities for individuals with disabilities, via
telephone, January 31.
Narayanan S, Balasubramanian S, Swaminathan JM (2009) A matter
of balance: Specialization, task variety, and individual learning
in a software maintenance environment. Management Sci. 55(11):
1861–1876.
Narayanan S, Swaminathan JM, Talluri S (2014) Knowledge diver-
sity, turnover, and organizational-unit productivity: An em-
pirical analysis in a knowledge-intensive context. Production
Oper. Management 23(8):1332–1351.
Narayanan S, Terris E, Sharma A (2019) Abilities-first: Steps to
create a human-centric, inclusive supply chain. Supply Chain
Management Rev. 2019(November/December):34–41.
Neumann E (2019) Sheltered workshops for people with disabil-
ities: A reliable opportunity or an outdated system? Accessed
December 1, 2019, https://www.kuer.org/health-care/2019-04
-08/sheltered-workshops-for-people-with-disabilities-a-reliable
-opportunity-or-an-outdated-system#stream/0.
Nicolae M, Arıkan M, Deshpande V, Ferguson M (2017) Do bags fly
free? An empirical analysis of the operational implications of
airline baggage fees. Management Sci. 63(10):3187–3206.
Nietupski JA, Hamre-Nietupski S (2000) A systematic process for
carving supported employment positions for people with
severe disabilities. J. Developmental Physical Disabilities 12(2):
103–119.
Novak J (2015) Raising expectations for US youth with disabilities.
Federal disability policy advances integrated employment. Center
Ed. Policy Stud. J. 51(1):91–110.
Parrotta P, Pozzoli D, Pytlikova M (2014) Does labor diversity affect
firm productivity? Eur. Econom. Rev. 66:144–179.
Narayanan and Terris: Inclusion, Disability Diversity, and Productivity
Peckham (2020) Our mission. Accessed September 22, 2020, https://
www.peckham.org/about-us/our-mission/.
Santos F, Pache A-C, Birkholz C (2015) Making hybrids work:
Aligning business models and organizational design for social
enterprises. Calif. Management Rev. 57(3):36–58.
Scoppetta A, Davern E, Geyer L (2019) Job carving and job crafting:
A review of practices. Working paper, European Centre for Social
Welfare Policy and Research, Vienna.
Shima I, Zólyomi E, Zaidi A (2008) The labour market situation of
people with disabilities in EU25. Policy Brief 2.1/2008, European
Centre, Vienna.
Shore LM, Chung-Herrera BG, Dean MA, Ehrhart KH, Jung DI,
Randel AE, Singh G (2009) Diversity in organizations: Where are
we now and where are we going? Human Resources Management
Rev. 19(2):117–133.
Staats BR, Gino F (2012) Specialization and variety in repetitive tasks:
Evidence from a Japanese bank. Management Sci. 58(6):1141–1159.
Stamps D (1995) Just how scary is the ADA? Training 32(6):93–98.
Thanem T (2008) Embodying disability in diversity management
research. Equal Opportunities Internat. 27(7):581–595.
Trewin S (2018) AI fairness for people with disabilities: Point of view.
Preprint, submitted November 26, https://arxiv.org/abs/1811
.10670.
Trist E (1981) The evolution of socio-technical systems. Occasional
Paper No. 2, Ontario Quality of Life Centre, Toronto.
UCLG (2018) Culture in the sustainable development goals: A guide
for local action. Accessed November 10, 2019, https://www.uclg
.org/sites/default/files/culture_in_the_sdgs.pdf.
United Nations (2015) Sustainable development goals. Accessed
November 11, 2019, https://sustainabledevelopment.un.org/
?menu= 1300.
United Nations Economic Commission for Europe (2018) UN Alli-
ance aims to put fashion on path to sustainability. Accessed
May 21, 2020, https://www.unece.org/info/media/presscurrent
-press-h/forestry-and-timber/2018/un-alliance-aims-to-put-fashion
-on-path-to-sustainability/doc.html.
von Schrader S, Malzer V, Bruyère S (2014) Perspectives on disability
disclosure: The importance of employer practices and workplace
climate. Employee Responsibilities Rights J. 26:237–255.
World Health Organization (2011) WHO World Report on Disability
(WHO Press, Malta).
Williams J, Mavin S (2012) Disability as constructed difference: A
literature review and research agenda for management and
organization studies. Internat. J. Management Rev. 14(2):159–179.
Williams-Whitt K, Kristman V, Shaw WS, Soklaridis S, Reguly P
(2016) A model of supervisor decision-making in the accom-
modation of workers with low back pain. J. Occupational Reha-
bilitation 26(3):366–381.
Womack JP (2007) Respect for people. Email to the Lean community.
Lean Enterprise Institute (December 20), https://www.lean.org/
womack/displayobject.cfm?o=755.
Zuckerman D (1993) Reasonable accommodations for people with
mental illness under the ADA. Mental Physical Disability Law
Reporter 17(3):311–319.