Build a greater understanding of who you’re working with and how they experience the situation you’re addressing. Research is vital to a successfully project and a mutually-helpful collaboration. 

Practice inclusive recruitment 

Once the goals and objectives of a research initiative are established, it’s important to carefully define the audience (see 18F’s guide for more info) and then to take the time needed to recruit for that audience.   

To quote 18F, you must do research with people who will actually use your services. It’s counterproductive and potentially harmful to draw conclusions from research conducted with people who are not in the right demographic or target group, or who don’t share the experience or situation that the research aims to understand.

Consider the unintended consequences of your recruitment approach. For example, recruiting via personal networks can lead to hearing only from people who share similar characteristics, values, and beliefs as yourself. Ask yourself whether these people are authentic representatives of the people who will ultimately use your product. Similarly, recruiting via social media can omit people who don’t have internet access at all. These issues are known as sampling bias.     

Even when your study targets the “general population”, take steps to recruit for a diversity of participants, rather than recruiting for convenience (which could potentially introduce sampling bias), or trusting to randomized recruiting. This might include taking steps to recruit and screen for specific groups—here is an incomplete list:    

• People with disabilities, who may or may not use assistive technologies   
• People with limited English proficiency, including newcomers to the United States  
• Members of the BiPOC (Black, Indigenous, and people of color) community  
• People in communities who may have limited trust in the government  
• People with limited or no internet access, or who are smartphone-only  
• Rural community members   
• Veterans  
• Older adults   
• People who are formerly incarcerated   

For long-term success, invest in relationships with communities and community organizations to establish a collaborative partnership, thereby shifting from transactional to relational recruitment.  

Compensate fairly 

It’s important to pay research participants and community collaborators for the value they contribute to your work. At the Bureau, members of the public who participate in research are almost always paid for their time. Exceptions include employees of institutions either regulated by the CFPB or who work with the CFPB, and federal employees working on government time.   

Compensation models should consider the additional costs some might incur to participate in research. This might include time off from work, child care, and transportation costs.  

If you are working with or for a federal agency, depending on the structure of your research, compensation may have implications for the Paperwork Reduction Act (PRA). Be sure to consult your agency guidelines.   

Consider other forms of compensation that have value to the people or community you are researching if you can’t pay them or they decline payment. Some examples of what you could offer:

• Curated package of agency resources tailored to the participant’s needs  
• Technical assistance in the form of trainings or webinars on topics relevant to the participants  

Take care not to harm participants through compensation.  

Note at least two ways in which compensation can be problematic: 

• Undue influence: An excessive or inappropriate offer of compensation is made in order to obtain participant compliance. The amount or type of compensation should never influence people to participate against their better judgment.  
• Coercion: An overt or implicit threat of harm or negative consequences is intentionally presented by one person to another in order to obtain participant compliance. Compensation for research is not coercive in and of itself, since it does not involve a threat of harm. However, compensation can create potentially coercive situations, as when a third party is paid for another subject’s participation, and that third party can exert coercion over the subject in order to obtain payment.  

Obtain informed consent  

Informed consent is a person’s or community’s permission to participate in a study or design activity with their full knowledge of possible consequences—both good and bad. People must be able to grant permission free of coercion or undue influence and comprehend what they are agreeing to participate in. They should also be aware of what information will be collected as part of the activity and how their information may be used or shared downstream. 

In practice, this means the participant should be provided the following information before they agree to participate:    

• An overview of the nature of the research, the topics covered, and how the research will be conducted   
• The types of information you may record ( including audio or video) or collect, and how you will use the information 
• How you will share the information collected  
• That the participant has control over their participation and can stop research at any time  

There may be additional privacy notice requirements if personally identifiable information (PII) is collected from people. Be sure to consult your agency guidelines and the Privacy team. 

Make consent forms clear.  

An important underlying principle is that participants can understand what they’re agreeing to. Make sure that consent forms are written using plain language guidelines. Avoid unnecessarily complex terms or legalese.  

Collect only the bare minimum of information. 

Collect, record, and share only the bare minimum of information needed to accomplish the goals of your research, and carefully safeguard that information. See the section on data handling for more information.  

Consent is never binding. 

Once you’ve obtained a participant’s agreement at the beginning of a research session, continue to ask for consent as needed during the research itself. Participants may not have fully understood the implications of what topics might be covered, and specific questions or issues might hit a nerve.  

Anticipate sensitive questions or topics as best you can, and look for signs of trauma. Be prepared to ask your participant if they’d be willing to talk about a topic, and if they’re not, move on, or end the research activity as needed. Note that even if the participant chooses to end research sessions early, they should still be paid for their time.   

FURTHER READING

18F’s guide to informed consent  

Check your biases 

As part of beginning the project for which you’re conducting research, your team ideally spent time uncovering biases and walking through an exercise such as Frames of reference. If you have not taken those steps yet, considering doing them before entering into research work.  activity

Be aware of how bias can show up during research.  

Research is the primary time when your team will interact directly with the people you are seeking to help. In order for the research to be as useful as possible, it’s important to look at how bias can show up during the research process. Whether in choices around who participates, what information is collected, or how information is interpreted, there’s always a potential for bias to limit or skew the work. You will not be able to completely remove the potential for bias, but recognizing and proactively working to mitigate it will improve the credibility of your research and design.    

Some specific biases worth noting:  

• Sampling bias, where participants are chosen in a way that is not representative of the intended population. For example, recruiting research participants via personal networks can lead to hearing only from those who share similar values and beliefs to yourself. Recruiting via social media can omit people who don’t have internet access.   
• Social desirability bias, where participants answer questions or give feedback to conform to how they want to be viewed by others. For example, a research participant may say they like a prototype they don’t actually understand, for fear of being perceived negatively if they admit otherwise.   
• Interview bias and confirmation bias, where we seek or favor information confirming our own prior beliefs and assumptions. If you’re a sighted person, you may omit thinking about the needs of people with low vision when you conduct research. If you design a product and then go on to research it, you may consciously or subconsciously shape the results to support your initial design decisions.   

FURTHER READING

18F has put together a list of common biases in research. 

Handle data and stories with care 

Safeguard the data and information gathered from participants in a sensitive manner. Many underserved and underrepresented communities have had historically difficult experiences with the federal government. Safeguarding their PII and maintaining compliance with your organizational data management, privacy, and security standards are important to ensure participants are not vulnerable to further trauma from mishandling of their data, such as a privacy breach.  

18F defines PII as “information that can be used to distinguish or trace an individual’s identity, either alone or when combined with other information that is linked or linkable to a specific individual”; see their research guidelines for a more detailed explanation. 

When it comes to collecting PII from participants, best practices are as follows:   

• Collect only the bare minimum needed to accomplish your research. In practice, most PII, such as names or faces, are not needed and should not be collected. Other PII, such as participants’ voices, may be recorded for internal use only.   
• Anonymize and aggregate information collected from participants before it’s stored. For example, give each participant a code number (e.g., P01, P02) before saving and storing their data. Refer only to those code numbers, even in your notes, unless you have explicit consent to collect and store PII otherwise. Even if you do have such consent, take all steps necessary to safeguard the privacy of data stored.   
• Limit access to anonymized data to those conducting the research.   
• Delete data containing PII when it’s no longer needed.  
• Consult with your Privacy team before engaging in research to ensure compliance with your organizational and federal standards. 

Recognize and respond to trauma in participants 

Trauma is an event, a series of events, or a set of circumstances experienced by an individual or group as emotionally or physically harmful, and that has lasting effects on their sense of well-being. Research sessions may touch on difficult or sensitive topics. Familiarize yourself with the 4 R’s framework and apply it if needed to avoid re-traumatizing a person.  

FURTHER READING

Much of this guidance is based on well-established research and resources from experts. If you are interested in learning more about trauma and trauma-informed/responsive practices, read this comprehensive report on trauma-informed approaches created by the Substance Abuse and Mental Health Service Administration (SAMHSA).

Related pages

• Overview
• Setting the foundation
• Design and development