Formalize research partnership and establish roles and expectations
This resource outlines steps to establish and build a strong working relationship with an implementing partner at the beginning of a randomized evaluation. Topics include questions to consider when developing a project scope, timeline, communications strategy, and formal agreements between researchers and implementing partners. This information may be most useful for researchers who have identified an implementing partner, research questions, and experimental design.
Once a research team and an implementing partner have been identified, the work of establishing and building the relationship between them begins (Glennerster 2017, 177).1 At the beginning of a new research partnership, explore questions around the project’s goals and activities to create working agreements between the implementer and research team. The following outlines questions to consider as researchers develop relationships with implementing partners:
- Working agreements. What topics and activities will this evaluation include? When will they take place? What falls outside the project scope? What kinds of support do all stakeholders need to provide to make the project successful?
- Communications strategy. How will institutions communicate project updates? How and with whom will the research team share results?
- Data sharing agreements and formal protocols. Are any legal agreements necessary for research activities to proceed? Under what circumstances can the researcher collect or receive data from the implementing partner?
Establish working agreements with partner organizations
When beginning a research partnership, consider the activities and timeline of the research project, along with who will carry out which aspects of the plan. To facilitate this, researchers might draft a memo that includes mutually decided-upon working agreements for the course of the study. This can serve as a tool for onboarding new staff members and maintaining shared goals of the evaluation throughout the duration of the project. The following are topics to explore when drafting a working agreement.
Defining the project scope sets expectations for the working relationship. The process should be collaborative and draw heavily from the implementing partner’s priorities and program strategy (Glennerster 2017, 182).2 The project scope may include elements such as:
- Description of the randomized evaluation, including research questions, random assignment methods, outcomes of interest, research activities, and intervention activities
- Objectives of the research partnership
- Approximate timeline of the intervention and research activities
- Deliverables or periodic updates to implementing partner
- Tasks to be undertaken by research team staff and implementing partner staff
The following are examples of questions to guide the conversation on defining the project scope with an implementing partner.
- What is the purpose of this partnership? What questions are we seeking to answer and how will we answer them? Describe the goals of the project, including a brief description of the intervention, research questions, and how the evaluation will address the proposed research questions.
- Who is responsible for what? What kinds of support and analyses will the research team be able to offer the implementing partner? Who should be contacted for which type of question? Describe the activities associated with the policy or intervention, along with who at the implementing organization is responsible. Define the extent to which researchers will be involved in monitoring the intervention. Establish points of contact on the research team for categories of work, such as analysis, contract negotiation, data sharing, research implementation, and intervention design to facilitate communication. Clarify the extent to which research staff is available to support activities not related to the evaluation, such as analytical and technical support for implementing partner staff.
- How can we secure access to data that will inform the key outcomes of interest? Develop a plan to secure data from the implementing partner and from other data providers. Establish secure data transfer protocols and a plan to communicate protocols with the implementing organization and data providers.3
- What data will the implementing partner need to provide? Extracting data for the evaluation will likely require significant time from a data specialist at the implementing organization. The research team will need details on what every variable means, how the program was defined, and a guide for how to link across different data systems.
- What kinds of analytic support can the research team provide? Implementing partners might find analysis of data from implementation monitoring and information about their participants to be useful (Glennerster 2017, 184).4 Depending on staff capacity, consider whether and how to share this information. Be cautious about potentially misdirecting resources and staff time when providing analytical support unrelated to research activities. Principal Investigators (PIs) and project leadership should take the lead on managing expectations about the capacity for providing additional analyses at the beginning of the project when defining activities, roles, and responsibilities.
- How will we manage fundraising responsibilities? What will funds cover? Determine how funding will be allocated to support intervention and research activities. Decide who will take responsibility for which fundraising activities.
Define when program and research activities will take place, along with a timeline for results sharing. Include training needs in the timeline for researchers and implementing partners. Below is an example of a simple Gantt chart outlining a project timeline. In addition to showing milestones, this chart highlights the relationship between the intervention, data extraction, and analysis.
Define communications strategy
A conversation at the beginning of the partnership will help researchers and implementing partners define institutional roles and manage expectations. Questions to consider discussing with partners include:
- When and how should the implementing partner notify researchers about changes or aspects of the intervention that may have an impact on the analysis and results? While the program implementation information relevant to the project may be obvious to the research team, it may not be obvious to the implementing partner. Establish regular check-ins with the implementing partner to create space for regular updates and to help build the relationship. This will ensure the research team has access to small but potentially important updates about the intervention and help the partner understand what kind of information is useful to share. Use these check-ins as an opportunity to ask questions about updates to the intervention, ensure partners understand the importance of providing regular updates and information about changes to the program or study, and create the expectation that implementing partners will share updates with the research team. If possible, make a plan to proactively gather updates about randomization, recruitment, or changes to the intervention. Explore the feasibility of directly accessing study monitoring data, including fidelity to treatment assignments, enrollment, take-up, and other program monitoring data.
- Will the researcher share results with implementing partners before they are finalized? Carefully consider trade-offs associated with sharing results before they are finalized. Doing so may support the implementing partner’s decision-making and maintain the relationship. Refraining from sharing until the conclusion of the analysis may prevent the sharing of incorrect information, in case results change over time and prevent the implementing partner from making changes to the intervention based on interim result.
- Who can the research team and implementing partner share information with, and who should be removed from all communication? How can the research team and implementing partner coordinate? Clarify who should be able to access certain pieces of information about study implementation, including treatment assignments and interim outcomes. Consider embargos on communication with media to prevent public discussion of results before they have been published.5
When necessary, establish a legal research agreement
After establishing working agreements in a memo, the project may require a legal agreement to move forward. If this is the case, expect to dedicate significant time and effort to discussing and iterating agreements to suit the needs of participating institutions.6 Home institution research administrators will be able to provide more detailed information about data sharing and legal agreement procedures, including preferences, agreement templates, and key institutional constraints. Institutional and data provider policy on IRB approval and protocol will likely factor into this process.7
The following includes general background on different types of agreements commonly used for research projects. Legal experts at the researcher’s home institution will be able to provide more guidance about which agreement is appropriate to pursue.
To secure confidential data:
- The project may require a Data Use Agreement (DUA) or Data Sharing Agreement (DSA) 8 to outline data requested and process for transferring, protecting, and using data. This document will lay out a number of provisions that may impact the research project, including guidelines for publication, review periods, liability, data destruction protocols, data description and timeframe, research subject privacy, and personnel access restrictions.9 A DUA or DSA may also include a description of the project and research activities. While each of these topics will be documented in the DUA or DSA, researchers should plan to discuss and clarify each of these provisions separately with implementing partners. Data security and transfer protocols are especially important to clarify at the outset of a research project.
- Depending on the specifics of the project, researchers and implementing partners might decide to explicitly restrict sharing information with external entities through a Non-Disclosure Agreement (NDA) or Confidentiality Agreement (CDA). These are useful when data or other confidential information related to the intervention or program being studied will be used for the development of the research project. Typically, an NDA/CDA covers proprietary information that may be exchanged in the process of research, but does not cover the exchange of a data set or disaggregated data that would be used in data analysis. Expect to include content similar to a DUA or DSA, including a description of the project, data requested from the data provider, personnel access restrictions, provisions for maintaining subject privacy, and protocols for destroying data. Again, it is useful for information on each of these provisions to live separately from the formal agreement to make sure that all research and implementing partner staff understand data security and confidentiality protocols guiding research activities.
To formally lay out the project scope and plan for moving the work forward:
- Start with a working agreement in a memo or conversation with the implementing partner that includes a scope of work, timeline, and communications strategy. This might include the project’s goals, a description of the evaluation, approximate timeline, research deliverables, description of the intervention and research activities, and tasks to be undertaken by research team staff and implementing partner staff. In some instances, a grant award may require both parties to determine the scope of work, deliverables, and timeline. A working agreement will allow researchers and partners to work directly with one another throughout the course of a negotiation, while a legal agreement process will involve the organizations’ legal teams. Researchers will likely negotiate many agreements in the course of the partnership. Starting with a comprehensive working memo at the outset of a project will allow researchers and partners to draw from this document and ensure that important details are not lost in subsequent negotiations.
- After establishing working relationships in a memo, one or both parties may prefer to create a Memorandum of Understanding (MOU), which formally expresses goals, structure, and actions that all parties will take within the partnership. Depending on institutional preferences, an MOU may not be feasible to establish for a single research project. Since it is not a legally binding document, it may be preferable to simply use an informal memo to establish roles and responsibilities, using a structure similar to the working agreement section described in the previous section.
- Your partner may request a Research Agreement or a Collaborative Research Agreement, which formally defines the research scope and activities to be undertaken by all parties. This should include a project scope, activities, and a timeline for completion, as well as topics covered under both DUAs and informal working agreements. This should draw from prior conversations about the goals of the project, how responsibilities will be divided and staffed, and deliverables. It may also govern intellectual property, liability, taxes, insurance, warranties, and other items necessary for contracts between institutions.
- If the research team might pursue multiple research projects with an implementing partner, consider a Master Research Agreement. This agreement will lay out common terms for data management, use of institutional names in publications, and intellectual property considerations between your home institution and the implementing partner. In this format, a brief description of each research project and the data to be shared will be included in a separate addenda to be approved by the home institution and data provider.
- If the implementing partner for the research project will pay the researcher, consider a Sponsored Research Agreement or services agreement. This should define the scope of work, schedule of activities, deliverables, and data management procedures. This agreement will include content similar to a Research Agreement with separate provisions for budgets, payment obligations, and timeline.
- If an implementer and a researcher receive an award for the same project and from the same funder, one may give a subcontract to the other. This will require negotiating budgets, payments, obligations, and reporting, similar to the terms defined in a Sponsored Research Agreement.
Elements of a communications strategy
Prospective development of a communications plan between researchers and partners ensures a common understanding and agreement, and may prevent tension from misaligned expectations. Teams may decide to formalize this plan through a written memo, in addition to discussing in meetings or phone calls. Documenting the partner’s expectations about outcomes of interest and expected results may facilitate productive conversations about research results by cutting down on scope or mission creep.
A communications and results plan might include:
- Who will have access to:
- Treatment assignments
- Individual or study unit-level outcomes
- Aggregate outcomes to date
- Enrollment data
- Monitoring or process data
- What, when, and with whom will results be shared during the study:10
- Timelines for results sharing
- Outcomes to be shared
- External parties to consider (or embargo): news outlets, data agencies, funders, participants
- Timeline of academic publication process and influence on external communications timelines
- Reasonable effect sizes based on past literature or previous programs
- Rationale for or descriptions of:
- Blinding certain members of the study team or partners from assignments or outcomes
- What is reasonable to infer from “interim results”
- Importance of completing the planned study time period
- Importance of meeting enrollment and recruitment targets
- Importance of communicating about all program changes
- Agreement on study motivations and definitions of “success:”
- Research questions of the partner and researcher. For example, an implementing partner may focus on understanding whether their program works, while an academic researcher may focus on understanding why participants engage or respond in a certain way, or on broader theoretical questions.
- Specifically defining which outcomes the partner cares about the most.
- For example, in an education program this could be the number of college credits earned, degree completion, number of years completed, starting salary, etc.
- Agreement on academic and other publications:
- Whether the partner has a right to preview manuscripts prior to publication in order to check for inadvertent disclosure of confidential information
- Confirm that researchers have the right to publish results and have discretion over accepting any suggestions from the partner with respect to confidential information, description of the program, or interpretation of results
- Whether and under what conditions the partner will remain anonymous or identified, e.g., “a large university” versus “MIT”
- Guidelines for branding and approvals of public materials, like project signage, event agendas, and research and policy briefs
We are grateful to Noreen Giga, Rohit Naimpally, Elisabeth O’Toole, and Kristen Watkins for their insight and advice. Alex Marlowe copyedited this document. This work was made possible by support from the Alfred P. Sloan Foundation and Arnold Ventures. Any errors are our own.
"Using administrative data for randomized evaluations." J-PAL North America.
This resource provides practical guidance on how to access and use nonpublic administrative data for a randomized evaluation.
"Data security procedures for researchers." J-PAL North America.
This resource provides a primer on basic data security themes, provides context on elements of data security that are particularly relevant for randomized evaluations using individual-level administrative and/or survey data, and offers guidance for describing data security procedures to an Institutional Review Board (IRB) or in an application for data use.
Duflo, Esther, Rachel Glennerster, and Michael Kremer. 2006. “Using Randomization in Development Economics Research: A Toolkit.” w333. Cambridge, MA: National Bureau of Economic Research. http://www.nber.org/papers/t0333.
Glennerster, Rachel. 2017. “Chapter 5 - The Practicalities of Running Randomized Evaluations: Partnerships, Measurement, Ethics, and Transparency.” In Handbook of Economic Field Experiments, edited by Abhijit Vinayak Banerjee and Esther Duflo, 1:175–243. North-Holland. https://doi.org/10.1016/bs.hefe.2016.10.002.
“MIT Office of Sponsored Programs.” Accessed July 18, 2018. https://osp.mit.edu/.