Partnering for project success in global health
An MIT student team describes their collaborative process and shares tips on conducting qualitative surveys on the field

Following the completion of their Global Health Delivery Lab (ghdLAB) projects in Africa and India, we asked our MIT student teams to reflect on their ghdLAB project experience and share what they learned – insights, a how-to guide, best practice tips  – in ways that can be used by others undertaking similar work or facing similar issues in global health. We share their thoughts under the ‘management for the world’ (mftw) collection in the hopes that others will build on it in keeping with the creative commons license.

By S. Atkins, A. Geertz, K. Kung and N. Patel

Often, with an MBA background, we arrive on-site at a host organization already with a slide deck of cure-all recommendations to be presented. After all, efficiency is highly valued, and why should time be spent listening to the host on-site, if we already have all the information, either through background research or prior long-distance communications with the host?

The question is, do we?

To start off, many organizations may have certain pieces of essential but sensitive information that they are unwilling to openly share with us, not at least until they establish a certain level of trust with us, the level of trust that can only be engendered by face-to-face contact. This challenge can especially be exacerbated when we deal with a host from another country, where cultural differences abound.

Our approach is based on two assumptions. Firstly, we cannot fully appreciate the nuanced context of the organization unless we spend some time with them. Secondly, the host is frequently as well-equipped as we are to solve their challenges, though often it does not realize this. Therefore, it is often more valuable not only to simply deliver recommendations to the host, but also to build problem-solving capacity within the host, such that it can address other issues in the future by itself, without our intervention.

As a group of four MIT students working to improve health care utilization in Kenya, we found that process and approach are as critical to our success as our recommendations. We hope to provide a high level outline of our strategy and demonstrate how the process yielded positive results. We feel that the process can be broken out into 10 steps, outlined below:

In this article, we will briefly describe these different phases, and outline the approach of the qualitative survey, a tool that we found extremely useful in our project.

Phase I: Prepare

Research to prepare your team, NOT to become an expert!

We found preparation to be important in ways different than we initially expected. The benefits of background research can be key to getting your team up to speed, but recognizing that this research is more for you than for your host! Without experience on the ground, the benefits of research become more about getting the team prepared for the time on-site, knowing which questions to ask, and getting ideas for where future opportunities lie.

Our Advice: Doing research DOES NOT make you an expert! Beware of presenting initial findings before you have listened to the real issues on the ground!

Phase II: Listen, Identify Opportunity, Collaborate, and Scope

The value of the information you gather is entirely dependent on how well you listen! THIS IS WHERE YOUR OPPORTUNITY LIES!

Listening to people on the ground is the secret sauce to your project’s success! There are distinct benefits here: First, it breaks the ice and makes people feel included. Setting the dynamic of being open to suggestions is critical to developing relationships with people most important in the community. Second, nobody knows the problem like they do! People on the ground are living in the very adversity that you are trying to fix. Their insight will be critical to finding solutions that bring change. Lastly, by letting your host drive the conversation initially, you will really hone in on the most pressing needs, and what they need from you.

Then your team can better scope the project needs and match the feasibility of what you have to offer.

Our Advice: LISTEN, LISTEN, LISTEN…. Opportunities can become evident if you approach conversation with a keen ear and open mind. Lastly, poor scoping leads to project failure every time. Find where the most critical needs are and carefully assess the impact you can make. Often, it is necessary to systematically take to people—i.e. carry out a survey. For some advice on how to design a qualitative survey, please see the “suggested tool” section later in this document.

Phase III: Drive at Results and Make Something Tangible

Don’t sit around and make it perfect! PITCH SOMETHING and see how it’s received… then your time can go towards making it better!

As MBA students, we are focused on perfection. Take an entrepreneurial approach. Generate lots of ideas, even if they are crazy. Don’t waste time trying to find the perfect recommendation initially. Throw something out and see how changes can be made. This will often help you drive at results that are feasible and the process can frequently help your process evolve in unpredictable ways!

Lastly, MAKE SOMETHING! There is nothing better that a tangible, model, a poster, or a pamphlet. It can always be polished and amended, but get something in hand. It will bring up moral and inspire more than the promise of something tangible. Getting to this phase earlier than you think can ensure that you have enough time to get the high quality final product as an end result.

Our advice: Get something in the hands of your team and your host. No power point slide or concept idea gets people as excited as the real thing! Don’t underestimate the time it takes to make this stuff (printing, design, building etc!) It may feel too early or not the best use of your time, but it will yield better end results!!!

Phase IV. Set Goals, Have a Follow-Up Plan, and Circle Back

Good ideas DO NOT bring change. Change comes from implementing good ideas and building a team to carry out this vision.

Setting goals will be critical. Without milestones, grandiose ideas can seem daunting and morale can suffer. Set realistic goals and maintain vision toward the ultimate destination. Having clear distinctions of responsibilities are critical to motivation as well as turning ideas into action. Articulate who will own which portions of implementation and be sure to routinely follow up to make sure the recommendations are achieving the intended results. Better to make change early on, than implement a large-scale project that doesn’t work the way you had hoped.

Our advice: Good ideas are worthless unless implemented. Leave time to initiate follow-up while you are still on the ground and develop a clear plan with milestones and plans to check in for when you leave the site. This is where you can really make a difference!

Suggested Tool: The Qualitative Survey

After listening to key issues, it is critical to ask the right questions at various stages of your project. The qualitative survey can be a great tool that can help hone in on trends and press to gain more information in identified opportunity areas.

The philosophy of the qualitative survey, in contrast to the quantitative survey, is that instead of systematically going through a list of questions, it seeks to engage the interviewee in a conversation. While it is desired to have some loose structure of the conversation, it is fine to omit some parts if they are not context-appropriate or to explore certain topics further if the interviewee happens to yield a very interesting view that, for example, has not been anticipated before. Furthermore, if you or your host organization plan to work with the community in the future, a quantitative survey is essentially an intimate conversation that can establish the requisite rapport between you and the community to increase the project’s success.

We think that the qualitative survey is also superior to the traditional town-hall-style community meeting in several aspects. Firstly, it ensures that everyone’s voice is heard, not just the loudest or the most outgoing. Secondly, when coming unannounced to a random sampling of houses, you are much less likely to be targeted by specific stakeholders who may have some vested interest that they may try to promote to you in order to influence your project to better serve their personal benefits. Thirdly and most importantly, if executed correctly, it can establish rapport between you and the community, and may well be the gateway to a successful project that involves the community.

When language is barrier, it is critical to have the host screen questions for cultural appropriateness and potential linguistic hiccups. Translation is key and the fundamental success will depend on a mutual understanding of the desired results. The value is in the conversation, which is easily interrupted if you are constantly switching back to English. The translator must know what information you are trying to get from each of the questions, when to press further, and most importantly, must share all of the “little details” with you in their answer. Opportunities can be revealed in the subtleties and there is a huge potential for loss of information if the translator screens out what they perceive to be irrelevant details.

This tool can help generate key trend results that can often be more impactful than data from quantitative surveys of much larger scales.

Our Advice:

• The translator can make or break your survey. Interact with the translator for a few days to understand his/her personality, strengths, and weaknesses, as well as to build mutual trust. It is often much better to find a volunteer who takes pride in his/her work with the community, rather than to pay and hire a disinterested translator.
• Be thoughtful about your questions and be sure to openly communicate your goals for each question with the translator, such that he/she understands the motivation behind each question. Be prepared to listen to and implement the feedback from your translator, as he/she is usually more familiar with the community and context than you are.
• Avoid asking quantitative questions such as the daily amount of money spent on food, or the liters of water collected daily. If the number is not put in terms of familiar units, the interviewee may not know the answer, or worse, give you a random guess. If you nonetheless need to find out some quantities anyways, try to familiarize yourself with the local unit. For example, in rural Uganda, water is measured in 20-L jerry cans, while in the Kibera slum in Nairobi, water is measured in 40-L jerry cans, not in liters. In rural Ghana, the amount of money is frequently given in the old currency (second Cedi), not the new currency (third Cedi), even though the old currency is already out of circulation. . Avoid hypothetical questions such as the willingness to pay for an item. Instead, involve the interviewee in a scenario where he/she has to make realistic trade-offs (such as between health and education). Alternatively, ask the interviewee to tell you a story that involves his/her spending some money on an item or service, and proceed to ask if that amount is reasonable.
• It is potentially possible to ask sensitive questions (such as whom they trust) if you absolutely need that information, but approach them gingerly with lead-in questions. If possible, place these questions towards the end of the survey, after the interviewee builds sufficient trust with you.
• Be prepared to deal with statistics. The law of statistics says that a low average income (such as $1/day) tends to have a much relative variance than a high average income (such as ours). Because income is the main input for the interviewee’s daily activities, these activities may also have a much larger range of variation than we are accustomed to. There are several approaches to tackle this statistical challenge. You can ask the interviewee outright what these ranges may be, though be mindful that these ranges will at best be estimates, and at worse be guesses. Another potent approach is to ask for the most recent data point (e.g. “Tell me how much you spent the last time you visited a clinic”). Even though this specific data point may not at all reflect the usual mean, the hope is that if you gather enough data points (by visiting enough households in an area), the mean computed by averaging these data points will well reflect the true mean.
• Spend time upfront sensitizing the survey and mastering translation is critical to success. If several translators are present, it is often helpful to leave them an hour or two so that they can work out and discuss the questions in their own language amongst themselves. This ensures consistency for surveys being carried out in parallel.
• Even though the translator will be doing the most talking in the local language, by all means learn some phrases yourself—at the minimum the greeting, “my name is…”, “thank you”, and the farewell. Speaking the language makes you that much more accessible to the interviewee.
• Know that carrying out a good interview takes practice and coordination between you and the translator. A lot of this practice is only learned by doing actual surveys, so your first few house visits will inevitably be your pilot run, with suboptimal results. Over time, ideally, you should achieve sufficient coordination and mutual understanding with your translator that he/she can sometimes anticipate the impromptu questions you are about to ask or the alterations you are about to make to the current conversation.
• As long as the translator knows the sequence of questions well, it is not always necessary for you to ask every question in English. Feel free to remain in the background, and let the translator handle most of the interaction. A conversation flows more naturally this way. The translator can occasionally update you in English on the responses, or can record the responses themselves.
• Remember: quality, not quantity. To have a quality conversation with a household, be prepared to spend at least 20 minutes with them. From our experience, if a conversation goes really well, it can last up to 90 minutes.
• A conversation is a two-way exchange of information. Try to offer some information about yourself as well, and not just to extract information from the interviewee unilaterally.
• The conversations can be emotionally draining. As an interviewer-translator pair, if you interview for a full day, you may be able to complete 10-15 surveys. But at this rate, the likelihood that the translator will be burnt out is also high. It is often more sensible to carry out half-day interviews (which will already be very tiring). If your stay on-site is also short in duration, you will not be able to have a large sample size. For qualitative survey, a large sample size is not always needed as long as you can identify salient trends in the responses.
• DO NOT be afraid to make impromptu changes. The benefit of this type of survey is to see trends and NOT to run data! Don’t be afraid to adapt your survey based on the results you are seeing, but be wary of subconsciously tuning in to the identified trends and tuning out everything else that contradicts the trend.
• Consistency is key, but phrasing and interpretations will never be perfect on the first draft! It is better to compromise consistency in your survey questions than to consistently ask inappropriately worded questions throughout your study.
• When interacting with the interviewee, eye contact is essential. Even if you are asking a question in English, and the person does not understand English, it is often better to maintain eye contact with the interviewee instead of the translator (but be sure that the translator understands why you are doing so). It makes the interviewee feel actively engaged with you.
• Be on the look-out for unexpected responses. Often when we design a question, we already have certain possible answers in mind, and during the surveys will specifically tune in to these answers, while filtering out answers that may not fit within our paradigm. This subconscious filtering may also take place within the translator. So when, for example, the interviewee is telling you a story, it is better for the translator to translate the story in its entirety rather than to give you a synopsis, even if it takes more time.
• During the interview, also monitor the translator for signs of burn-out. If the translator appears to be slacking off in order to finish the interview more quickly, it is more desirable to terminate the interview prematurely rather than to risk obtaining answers whose accuracy or truthfulness may be dubious.
• After each interview, check in with the translator as to what worked well and what did not work as well. Praise the good points and discuss points that need improvement. This includes both the translator and you.
• If you need to document your work using photographs or videos, remember to ask for permission first. If you use a digital camera, it makes the experience that much nicer if you can instantaneously show the photograph to the interviewee.
• Sometimes the interviewee will ask you for financial or other kinds of help beyond your capability. It may make you feel sad, but there is very little you can do. Be prepared to encounter this kind of situation and have a response ready beforehand. Remember why you are there for the survey in the first place.
• Understand that many interviewees will display “interview fatigue”: that is, they have been interviewed so many times by other NGOs but have not yet seen any concrete improvements to their lives. If this is the case, then they may be leery of your approach or be resigned that this is just “yet another interview”. Sometimes they may express this frustration. It is a good idea to have a tactful answer ready beforehand for situations like this.

Download 10 steps to delivery (pdf).

Was this article useful to you? Please give us feedback on how to improve sharing our work by leaving us a comment or e-mailing us at ghd.projects.lab@mit.edu.

 This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 3.0 Unported License.