Working Draft Document Written by:
Kara L. Hall (NCI) , Kevin Crowston (NSF), and Amanda L. Vogel (Leidos Biomed)
November 19, 2014 Page 1
How to Write a Collaboration Plan
Why Plan for Collaborations?
Science is becoming increasingly collaborative, and frequently involves multiple
investigators, institutions, disciplines, and fields. Such collaborations often are able to
address more complex and sophisticated research problems, by integrating the expertise
and resources of multiple collaborators. But they also involve a number of costs, most
particularly management complexities, including additional attention to planning for and
facilitating effective team functioning, and preventing or addressing challenges specific
to teamwork that can threaten the success of the initiative. Poorly managed collaboration
may negatively impact the quality of the science, whereas well managed collaborations
have the potential to foster innovation, creativity, and productivity.
To help enhance the success in scientific collaborations, funding agencies may ask
investigators to submit “collaboration plans” as part of their funding applications, just as
they ask investigators to submit research plans. Collaboration plans may benefit any
scientific endeavor that includes two or more investigators working together. Though as a
proposed scientific collaboration grows in scope and size, such plans become
increasingly important.
Collaboration plans should address a range of issues relevant to laying the foundation for
the collaboration, as well as implementing and managing the collaboration, and engaging
in quality improvement activities specific to collaborative interactions. These plans
identify existing supports and challenges relevant to the collaboration, and describe a
program of action that will be implemented to help support smooth collaboration.
Submitted collaboration plans can be used by investigators as “roadmaps” for their future
initiatives, and by reviewers to help assess the capacity of a proposed team to
collaboratively execute its scientific objectives.
Collaboration plans may vary substantially in size and complexity, as reflective of team
science and complexity. Science teams can vary on a host of multi-level factors, such as
team members’ past experience with collaboration, geographic distribution of team
members, and scientific problem space, among many others. In addition, each
collaboration occurs in a unique context, shaped by the participating departments,
institutions, and disciplines. As a result, each collaboration plan will be unique. For
example, three co-PIs from the same department who have worked together successfully
in the past might need only a modest collaboration plan. But a large multi-disciplinary
multi-institution collaboration will likely require a detailed collaboration plan that, for
instance, discusses how disciplinary differences will be bridged, and how the
participating institutions will work together.
How to Use this Document
Funding agencies will provide specific requirements for collaboration plans in funding
announcements and policy documents. The goal of this document is to provide general
Working Draft Document Written by:
Kara L. Hall (NCI) , Kevin Crowston (NSF), and Amanda L. Vogel (Leidos Biomed)
November 19, 2014 Page 2
guidance for investigators on key factors to consider in preparation for a collaborative
initiative, and how factors may be described in a collaboration planning document. The
document offers ten key factors important to planning for effective team functioning.
Systematic consideration of these factors will help investigators to identify strengths and
weaknesses of their proposed collaborations, and to plan strategies to support smooth
team functioning given this unique set of conditions, with the ultimate goal of optimizing
the team’s scientific work together.
1. Rationale for Team Approach and Team Configuration
Team Approach: The collaboration plan should begin by justifying why the particular
scientific questions and goals require a team approach. For example, to answer a question
beyond the scope of an individual discipline, experts from other disciplines may be
needed on the research team. Alternatively, a research question may require that a team
rely on equipment or infrastructure located at multiple institutions, within multiple
centers or labs.
Not all research questions are necessarily best addressed using a team-based approach.
Scientific considerations are paramount when determining whether an individual or team-
based approach is best. For instance, Hays notes, “if fields of science have not
sufficiently evolved toward one another or their underlying support structures are
incongruous, it may be difficult or impossible to initiate and maintain cross-disciplinary
research even though the participants are eager and other readiness challenges have been
successfully met” (Scientific Readiness: Hays, 2008).
Team Configuration: The collaboration plan also should identify how the proposed team
composition is necessary to pursue the research goals. The team should include sufficient
breadth to ensure the expertise and manpower necessary to fulfill research objectives.
However, as the number of collaborators increases so do the potential challenges to
collaboration (Cummings et al., 2012; Hall et al., 2012). The collaboration plan should
describe the team size in light of this knowledge.
2. Collaboration Readiness
The plan should provide evidence for the collaboration readiness of the individual
participating investigators, the team as a unit, and of the institutions involved. Not all
individuals, teams, and institutions may be optimally ready for complex collaborations.
Therefore collaboration plans should describe the existing collaboration readiness and
describe strategies to address limitations.
Individual collaboration readiness refers to individual characteristics related to
collaborative initiatives, e.g., willingness to and interest in engaging in collaboration, and
past experience with team collaborations (Stokols, et al., 2008; Hall et al, 2008,
Stipelman, 2010). Team members may have different personalities, work styles, and
experience that influence their engagement in team-based work, therefore recognition of
those differences and strategies for managing such differences should be considered (e.g.,
See Bennett, Gadlin, Levine-Finley, 2010 at
Working Draft Document Written by:
Kara L. Hall (NCI) , Kevin Crowston (NSF), and Amanda L. Vogel (Leidos Biomed)
November 19, 2014 Page 3
https://www.teamsciencetoolkit.cancer.gov/public/TSResourceTool.aspx?tid=1&rid=267
)
Team collaboration readiness includes factors such as the mix of collaboration histories
of proposed team members. For example, teams with a mix of prior collaborators and
new collaborators may be most successful because they can benefit from the good
working relationships of prior collaborators as well as the new perspectives injected by
new team members(Uzzi & Spiro, 2005). Online team diagnostic surveys can be used to
assess a team’s readiness for collaboration (e.g., Wageman R, Hackman R, and Lehman
E, 2005; Bietz MJ, Abrams S, Cooper DM, Stevens SR, Puga F, Patel DI, et al., 2012).
Institutional collaboration readiness refers to the resources, infrastructure, and policies
that each of the participating institutions have in place to support the collaboration.
Institutions may provide support relevant to each of the key planning factors described in
this document. Examples include: technologies to support communication and
coordination; training opportunities relevant to collaboration; promotion and tenure
policies that recognize (or at least do not penalize) participation in science teams;
research development professionals (e.g., http://www.nordp.org/) whose job it is to
support team formation and nurture partnerships; and consultants who can enhance
leadership and management skills, and facilitate quality improvement activities. This
section should address each of the institutions involved in the proposed collaboration.
The goal of this portion of the plan is to demonstrate individual, team, and institutional
readiness for the proposed collaboration. It is not expected that conditions will be ideal at
any of these levels, but challenges should be identified before entering upon a
collaboration, and whenever possible, the plan should identify steps that will be taken to
address the challenges.
3. Technological Readiness
The plan should document the availability and planned use of technological resources to
support the scientific collaboration. These may include mechanisms to support the
scientific process, including technologies for data sharing and analysis (e.g., data sharing
agreements, common databases, online collaborative data analysis platforms). These may
also include mechanisms to support collaborative processes, such as communication
technologies (e.g., videoconferencing, teleconferencing) and coordination technologies
(e.g., calendaring, task management platforms, and work flow or project management
tools).
An important consideration is whether members of the team are ready to use the available
technologies, including having both the willingness and skills to use the technologies
necessary for the collaboration. Another important consideration is the interoperability of
systems, as members of the group may have trouble working together if they are using
different systems (e.g., different videoconferencing systems, different database systems,
or different data analysis packages).
Working Draft Document Written by:
Kara L. Hall (NCI) , Kevin Crowston (NSF), and Amanda L. Vogel (Leidos Biomed)
November 19, 2014 Page 4
4. Team Functioning
The collaboration plan should describe plans for carrying out key processes that underlie
effective team functioning. These might include plans for generating a shared vision,
mission and goals; creating shared mental models; and externalizing group cognition
throughout the collaboration (e.g., generating visual schematics that capture the group's
understanding of a variety of factors relevant to their collaboration, including the
scientific problem space and collaborative research work flow). Additional key team
processes include developing team-level understanding of each team member’s areas of
expertise, roles on the team, and contributions to the science. This is also known as
developing shared understanding of who knows what (compilational memory) and does
what (compositional memory), and how things get done (taskwork memory) (Hall et al.,
2012).
Collaboration plans for interdisciplinary teams may also include plans for fostering key
team processes necessary for interdisciplinary work, such as critical awareness of the
strengths and weaknesses of contributing disciplines, and shared language that bridges
disciplinary differences (Hall et al., 2012).
Examples of strategies that can be used to support these team processes include: kick-off
retreats focused on developing the shared vision, mission, and goals; development of
cooperative agreements and operating manuals for collaborations; and regularly
scheduled “touch” meetings for team members to maintain and update their shared
understanding of the science and work process. These strategies should take into account
unique or complex factors for the proposed collaboration, such as the phase of the
research process (Hall et al., 2012), interpersonal relationships and collaborative history
of team members (Stokols et al., 2008), and the complexity of the collaboration (e.g., the
size, diversity, dispersion, task interdependency of team). Professional consultation or
facilitation may be helpful to support some of these processes.
5. Communication & Coordination
Team-based scientific initiatives require that effort be invested in supporting effective
communication and coordination of tasks. As team size increases, so does the investment
needed for effective communication and coordination.
Communication: Teams that are particularly diverse in terms of team members’
locations, demographics or disciplinary cultures (e.g., authorship traditions, work styles,
terminology, preferred methods) will face increased communication challenges. The
collaboration plan should describe the plans for communication within the team, such as
meeting frequency and modality (e.g., teleconference, in-person meeting), for
collaborations across time zones (e.g., mutually respectful meeting timing) and for
asynchronous communications (e.g., email use or document sharing).
Coordination: Greater use of coordination mechanisms has been found to be related to
more successful outcomes in large teams (Cummings & Keisler, 2005, 2007). The
collaboration plan should include strategies to coordinate day-to-day operations and
Working Draft Document Written by:
Kara L. Hall (NCI) , Kevin Crowston (NSF), and Amanda L. Vogel (Leidos Biomed)
November 19, 2014 Page 5
approaches for the completion of scholarly products (e.g. work flow, data sharing and
integration). The plan should address how tasks get allocated, how resources get shared,
and how work gets integrated into the collaborative effort. These plans need to be
specific to the particular collaboration, addressing, for example, such factors as the
number and distribution of team members and the design of team tasks.
6. Leadership, Management, and Administration
The plan should describe the planned leadership and management approach that will be
used to address the other components in the specific team context proposed in the
initiative.
Leadership: Providing vision and direction for an initiative is critical to success,
particularly in team-based science. The more complex the initiative, the greater the
demands on leadership and management. There are numerous approaches to leadership
(e.g., hierarchical, heterarchical, transformational, transactional). A leader’s approach
will depend on the particulars of the initiative and the personalities of the leader and other
team members. When more than one formal leader is identified for a collaboration,
specific shared leadership strategies are needed (e.g., see NIH multiple PI leadership plan
examples: http://grants.nih.gov/grants/multi_pi/sample_leadership_plans.pdf).
Management: Ensuring that the vision established by the scientific leader is carried out
requires roles to be established, tasks to be identified, and research plans to be executed.
The collaboration planning documents should outline overall strategies for managing
personnel, processes and procedures within the team and across institutions. For example,
teams should establish how key decisions will be made about scientific direction or
changes in the team.
Administration: As team size increases, administrative tasks become increasingly
important to team coordination. Administrative activities of critical importance to large
teams including recruitment, hiring, and daily administration of the team workforce.
7. Conflict Prevention and Management
Some degree of conflict within a collaboration is inevitable and indeed, may even be
helpful for the team to achieve its goals, for example by leading to new avenues of
thinking for everyone involved. But there is a qualitative difference between generative
debate that focuses on scientific disagreement, which may help to further scientific goals,
and relational conflict that may undermine team functioning, and ultimately negatively
impact the science.
The potential for conflict will depend on the specifics of the team, e.g., member diversity
in terms of demographics (e.g., age, gender, culture) or disciplinary training may lead to
conflict. But the specific areas of conflict, and the ways they play out, will vary with the
unique combination of types of diversity on the team (Eigenbrode et al., 2007). For
example, the existence of subgroups among team members may produce fault lines along
which conflicts emerge (Bezrukova, 2013).
Working Draft Document Written by:
Kara L. Hall (NCI) , Kevin Crowston (NSF), and Amanda L. Vogel (Leidos Biomed)
November 19, 2014 Page 6
Many sources of team conflict can be anticipated (e.g., disciplinary differences). But
conflicts may arise even when not expected. For example, investigators with similar
training may underestimate the potential for conflict due to incorrect assumptions about
areas of agreement.
Conflict Prevention: Considering potential factors that may lead to conflict (e.g.,
ownership of data; intellectual property rights; authorship order) and addressing these
factors before the collaboration begins can reduce conflict later on. The plan should
identify strategies for engaging in this process. For instance, for a small scale
collaboration, an example strategy is the use of a pre-collaboration agreement, also
sometimes called a “prenuptial agreements for scientists” (Gadlin & Jessar, 2002). For a
large scale collaboration, development or use of an operating manual may be warranted
(e.g.,
http://www.teamsciencetoolkit.cancer.gov/public/TSResourceTool.aspx?tid=1&rid=371).
Conflict Management: Despite efforts to prevent conflict, conflict may still arise. To be
successful, initiatives must develop systems for managing conflicts, e.g., processes for
encouraging debate and facilitating productive conflict while preventing or managing
negative forms of conflict; as well as processes and procedures for resolving detrimental
conflicts. Institutions might support teams by providing informal and formal channels for
conflict resolution. Plans for managing conflict should be included in the collaboration
planning document. The approach taken should be commensurate with the characteristics
of the proposed collaboration (e.g., size, geographic dispersion of members, cross-
cultural make-up).
8. Training
Training plans should be included to help participating investigators to enhance
collaboration. Training may be included, for investigators for whom collaboration is new
as well as for those with prior collaborative experience, to enhance knowledge and skills
specific to factors related to the proposed collaboration. Training may occur at start of
the initiative and/or periodically throughout the collaboration.
Training Content: Training for scientific collaboration can help to build skills in many
of the key areas identified in this document (e.g., team processes, leadership,
management, communication, coordination and quality improvement activities) (Fiore,
Hall, et al., in progress). For interdisciplinary collaborations, training might also include a
focus particular to cross-disciplinary work, such as critical awareness of the strengths
and weaknesses of all disciplines, and strategies for combining approaches (e.g., theories,
concepts, methods) from two or more disciplines. Trainings may also convey skills
related to using platforms and technologies that will be used in the particular
collaboration, e.g., shared databases and data analysis software.
Training format: Training should be designed to meet a wide variety of investigator
circumstances and needs, including: different career stages, learning styles, training
interests and needs, and practical constraints. For example, web-based and webinar-based
training may be most appropriate for geographically distributed teams. Training can be
Working Draft Document Written by:
Kara L. Hall (NCI) , Kevin Crowston (NSF), and Amanda L. Vogel (Leidos Biomed)
November 19, 2014 Page 7
carried out at the individual level and for team units. Training might be formal (e.g.,
online courses, such as teamscience.net, which provides a completion certificate) or
informal (e.g., seminar series featuring the work of all participating team members, to
build cross-disciplinary awareness and greater mutual understanding).
9. Quality Improvement Activities
Teams that engage in systematic and iterative reflection about team performance and
subsequently adapt their team objectives and processes show better performance,
including higher levels of innovation (West et al., 2011; West & Lyubovikova, 2012).
The collaboration plan should describe activities that will be implemented over the course
of the research initiative to facilitate reflection about team performance (e.g., pre-briefing
and debriefing). It should also describe how the resulting information will be used for
continuous quality improvement, to help address challenges and improve the quality of
the collaboration. For a large and complex initiative, it may be helpful to involve outside
experts to design and implement these feedback and quality improvement oriented
activities (e.g., appreciative inquiry processes, Toolbox Collaborative Communication
Instrument -
http://www.teamsciencetoolkit.cancer.gov/public/TSResourceTool.aspx?tid=1&rid=775.)
Institutions might help by providing support for such activities.
10. Budget/Resource Allocation
Successful collaborations require investments in training, management and quality
improvement, all of which require financial support to ensure their successful
implementation. The collaboration plan should identify the specific budget items that
support the activities included in the plan. Clear but flexible plans for funds can allow
optimal preparation for and facilitation of collaboration. This can be particularly
important in large and complex initiatives where directions can change and additional
collaborations can be formed during the course of the initiative.
Final Considerations
As noted above, collaboration plans offer a structure that can guide investigators in
thinking through how they will plan to support the activities of collaboration involved in
a team-based scientific endeavor. It also helps them consider what challenges the
collaboration will face—for example, related to institutional barriers, cross-disciplinary
collaboration, or distance collaboration—and to plan for whether and how these might be
addressed through management and leadership activities. Accordingly, collaboration
plans should be “living documents,” first developed in the planning stages of a
collaboration, and then periodically updated as the collaboration is underway, to reflect
changes in the circumstances of the collaboration and the growing experience and
knowledge of the team as they continue to work together.
As investigators gain experience in collaboration planning, plans and procedures may
become standardized for a research group, institution or even a community of scholars
(e.g., a discipline or field). Such standardization might be reflected in the development of
Working Draft Document Written by:
Kara L. Hall (NCI) , Kevin Crowston (NSF), and Amanda L. Vogel (Leidos Biomed)
November 19, 2014 Page 8
language that reflects particular institutionalized procedures, as well as available
resources. Having such language to draw upon, in addition to examples of prior
collaboration plans with information about successful and unsuccessful strategies, can
greatly facilitate the development of future collaboration plans. Ultimately, however,
each plan will be tailored to the unique circumstances of the proposed collaborative
initiative. The goal is to effectively collaborate to more rapidly advance science.
