When worlds collide: lessons learned from the intersection of behavioral and human-centered design in humanitarian contexts
“Practical things are entrenched in the memory. If there is no planning, there is no personal fulfillment of goals and no ability to defeat personal challenges to motivate ourselves to work,” said Khalid, one of the job-seekers who participated in prototyping a job search planning tool aimed at helping Syrian and Jordanian job-seekers find work in Jordan. The tool Khalid tested was a combination of a calendar and a series of reminder SMS messages designed to help address the gap between people’s intentions to apply for work and their failure to complete job applications. The impetus for this tool derived from behavioral science research, which suggested that planning can help people take action on explicit decisions they have made to change their behavior. Khalid’s feedback helped researchers understand what worked and what didn’t, leading to critical modifications in the final version of the tool.
This example demonstrates how the IRC’s Airbel Impact Lab integrates behavioral science and human-centered design to develop scalable solutions to humanitarian problems. On their own, these approaches have been leveraged in a variety of contexts across the world — what is unique about the Airbel approach is bringing them together.
What is behavioral science and human-centered design?
Behavioral science explains how an individual’s social, psychological, and economic reality influence her behavior. In the humanitarian sector, behavioral science can help us identify interventions that encourage behaviors aligned with an individual’s desires and goals (such as searching for work) and reduce the challenges of acting on these behaviors (such as difficulty understanding how to apply for jobs in a new country).
Unlike other fields that study human action, such as traditional economics, behavioral science recognizes that the majority of people experience intention-action gaps. For example, even if we have a strong intention of spending three hours a day applying for a job until we are hired, the reality of our lives like the demands of our families, daily chores, perceiving a lack of jobs on the market, self-doubt, and sometimes just exhaustion can sometimes prevent us from following through. Often, there is not much humanitarians can do to change the overall context — but we may be able to design light-weight interventions that remove some of these barriers or otherwise help.
Airbel also uses human-centered design to improve our ability to generate breakthrough solutions to humanitarian problems. As the term suggests, it focuses on understanding the needs and values of the end-user because “you cannot meet people’s needs without consulting with them.” Those perspectives then feed the design of products, services, and experiences that address people’s daily challenges.
An iterative process
But how does this work in practice and why is it important to combine these approaches? Let’s look at Project Match, a two-year research and innovation project that aims at building tools and evidence to improve employment outcomes for low-skilled Syrian refugees and vulnerable Jordanians in Jordan. The first year of the project was dedicated to start up qualitative research, prototyping, and designing solutions to the employment problem. The second year was exclusively dedicated to testing out the impact of the solutions developed from the first year within the context of a Randomized Controlled Trial (RCT).
Using human-centered design for problem identification
Project Match began with field visits in which we talked to job-seekers to understand the constraints that they faced in finding formal work in Jordan. Those visits, along with an initial round of desk research, helped us understand initial constraints. These insights — information asymmetries between job-seekers and firms, lack of knowledge on formal work environment, financial inability to get transportation to work, childcare needs, skills mismatch, and more — then informed the tools that we developed to gather information on user experiences with the existing job search process. The outcomes included a set of user journeys and profiles that demonstrated the day-to-day process of searching for work and that presented job-seekers with different backgrounds in Jordan.
Behavioral science for problem break-down
Having used HCD to broadly identify the constraints different types of job-seekers experience with finding work in Jordan, we went through and tagged the pain points on the user journey and then broke them down into digestible human behaviors that could be altered through behavioral interventions. The table below illustrates the outcomes of this process, and highlights the challenges that prevent job-seekers from matching with great employment opportunities:
Start with divergence. Test many different potential solutions
After boiling down the problems and resulting behaviors, Project Match designed several prototypes to test whether we could (1) dispel biases and misperceptions that lead to job-seekers not seeking employment, and (2) encourage target behaviors like applying regularly for formal wage-paying work, attending job interviews, and attending the first day of work. Among the tested prototypes were interventions providing information about labor law, more informed and active job seeking practices, and interview preparation guidance. We tested these interventions for overall language, usability and potential for impact and then also studied how changing the delivery mechanisms between SMS, WhatsApp, printed flyers, personal interviews, and group sessions affected potential for impact and scalability.
When testing out the language used to deliver intervention prototypes, design, and behavioral science came into play together. We used HCD to observe and record how users responded to different ways of communicating about the same message and adopted behavioral science to identify the language that would optimize engagement with the tools and participation in the program. For instance, we compared the term “assistance,” which often implies cash in the Jordanian humanitarian context, to “services” to describe interventions. This process helped inform the future adaptations of the interventions.
Converge on interventions to move forward.
By the end of the design phase, it was time to converge on interventions that would move forward into the RCT phase of Project Match. Within the RCT phase, interventions developed and agreed upon could no longer be changed as this would disrupt the controlled environment required to generate concrete findings from an RCT. In doing so, we worked with designers to develop final versions of our interventions, which included video, paper, cash, and SMS tools and messaging. Our data collection tools included questions intended to capture changes in job-seeker behaviors, such as:
- In the last 30 days, how many job applications have you completed?
- In the last 7 days, how many hours did you search for work?
- How did you use the cash that you received?
Inclusion of these questions in the baseline, monthly, and endline surveys means we’ll be able to measure behavior change and assess how particular interventions impact particular behaviors.
The whole is greater than the sum of all parts
Our successes and failures with Project Match have shown us that integrating these two approaches is not easy — it requires explicit plans on how and why both approaches will be leveraged in different project phases and it requires staff who have specific expertise both in design and behavioral science. In addition, these processes take time — assumptions need to be validated, solutions that don’t work need to redesigned — and require flexible funding. Both of these are not always readily available in humanitarian contexts, but integrating these methodologies deliberately and with rigor at the very early stages of a project will allow us to maximize impact.