What’s in a Survey?

Since my time here in Tanzania is almost at an end, I figure it’s about time I write a post about what exactly I’ve been up to this whole time. A large chunk of my work here boils down to conducting a survey on the social implications of a human-wildlife conflict intervention. But there’s a lot more complexity behind the curtain.

Many of you may know that I work on human-wildlife conflict, and feel passionate about interdisciplinary approaches as being the most effective for alleviating conservation issues. Yet, as someone with a background in ecology, conservation biology, and animal behavior, I decided to take a detour from my dissertation research to participate in a summer practicum with the Northern Tanzania Rangelands Initiative (NTRI), and gain more field experience with social science. In particular, I have been working with Tanzania People & Wildlife (TPW) to try and understand the social impacts of their successful (99.9% of reduction in wildlife attacks on livestock) human-wildlife conflict intervention: the Living Walls. The Living Walls consist of planted Commiphora trees surrounded by chain link fence. They act as a much more effective alternative to traditional livestock enclosures, or bomas, and are a community-co-created and environmentally-friendly solution to livestock being attacked by wild carnivores.

A Living Wall in Loibor Siret

Although there are many effective interventions for resolving human-wildlife conflict, very rarely do we attempt to measure the impacts of human-wildlife conflict resolution on human health, economic factors, and gender dynamics. TPW’s Living Walls intervention provides us with an ideal and unique opportunity to assess such social impacts. However, particularly if there is no baseline pre-intervention data on social factors, this is much easier said than done…

Challenge 1: Creating the Survey

In March, my colleague, Steven (from the University of Dar es Salaam), and I set out to figure out how we might develop a survey that would be able to get at the social impacts (if any!) of the Living Walls, as well as the effects of the Living Walls on pastoralist tolerance for wildlife. For about 3.5 months, I took the lead on creating the survey, and it subsequently underwent countless revisions from multiple TPW and other NTRI staff. In order to design a survey that would be effective, incorporate social/health/gender factors, and not waste many peoples’ time, we were tasked with philosophizing about and answering difficult (and sometimes unexpected) questions such as: How can we measure impact without having baseline data? What is the difference between tolerance and attitudes toward wildlife, and which of these is more feasible to study? What counts as a “household” in Maasai-land (where families and friends may live together in the same boma)? … And countless other conceptual and logistical questions. Even before I arrived in Tanzania, devising the survey was teaching me more about the deep and complex levels of thinking needed for social science research, particularly social science research among different cultures and landscapes.

Challenge 2: Translation and Finalizing

Once we arrived at TPW’s headquarters – Noloholo Environmental Center on the border of Tarangire National Park -, and completed multiple survey conceptualization meetings with the co-directors of TPW, Laly Lichtenfeld and Charles Trout, our next step was to finalize the survey and translate it into Kiswahili and Kimaasai. Our plan was to translate the survey from English to Kiswahili to Kimaasai, and then back-translate all the way back to English, to ensure the meanings of the questions were not lost. For translation and back-translation, it is crucial to have different people perform each step, and ideally those will be people who are not at all involved in the survey design. Finding bi- and trilingual people at Noloholo wasn’t difficult, but they all had existing busy schedules and our survey was pushing an hour in length. Nevertheless, we somehow managed to translate, back-translate, and pre-test the survey at camp within the span of a few days. After “trimming the fat” from the survey to bring it under 1 hour in length, we were finally ready to create our sampling plan, find field assistants, and do some field pre-testing.

Challenge 3: Creating a Sampling Plan and Understanding Gender

Because there was no baseline for health, social, and gender factors at Living Wall bomas before they had Living Walls, it was necessary for us to sample bomas without Living Walls as well – which we termed “Non-Living Wall (NLW) bomas”. To attempt to get at the differences, we decided on a paired sampling plan consisting of a Living Wall boma and the closest NLW boma.

We also felt it was important to be able to discern gender differences in survey responses. For instance, women have completely different household roles than men; women may gather and transport wood to fix the boma, take care of the household chores, sell excess cow’s milk at the market and thus have their own money, etc., not to mention the underlying gender power dynamics within each household. So having a Living Wall (due to boma maintenance time saved, fewer livestock lost to wildlife attacks, etc.), is likely to affect women differently than men. But how to quantify these differences in a largely patriarchal culture? The men of the household are more likely to volunteer for the survey, and even if a woman happens to volunteer, it is likely she would have her husband standing over her as she answers the questions – which, of course, would affect her answers. Thus, we came up with an excellent solution: At each boma, interview a man and a woman at the same time, separately from each other, and use a female field assistant for the women’s interviews. Although our boma sample size would be cut down (from 200 bomas to 140 due to time constraints), we would have true gender representation from all of the bomas. Now, we just had to find a male field assistant and a female field assistant.

Challenge 4: Politics and Field Assistants

Steven coaching Edward on the survey app

With the help of the staff and community members, it wasn’t too hard to find field assistants who were bi- or trilingual, literate, and knew how to use a smartphone (for the survey app). We invited our two new field assistants, Edward and Rebecca, to Noloholo for training on survey implementation before heading out for a day of pre-testing in the field. However, before doing any work in our two villages, Loibor Siret and Narakauwo, it was crucial to meet with the mwenyekiti (chairperson) for each village, to get approval and make them aware of what we intended to do. Especially as a foreigner (called mzungu here), I’ve always found that you never know what to expect when talking to local political forces, so, as usual, I was somewhat wary of these meetings. Yet, even with tumultuous politics happening in Narakauwo, this process went surprisingly smoothly.  We were then in the clear to field test the survey with Edward and Rebecca.

DSC05794 copy
Mzee Leposo Noongululu (our first pre-test interviewee), and his youngest of 40 children.

The field testing turned out to be very necessary and very productive. Steven and I had a chance to see how the field assistants performed in a real survey situation, and we were able to give them feedback on various aspects of their technique: how to avoid probing (and instead ask the question exactly as it is written), how to work through logistical issues with the phones and with Open Data Kit app, and (specifically for Rebecca) to make sure that women interviewees were answering the questions with information about themselves, rather than about their husbands. After some more practice, and several revisions to the survey resulting from the pre-test findings, we were ready to begin the survey in earnest.



Challenge 5: Getting Around

So, we were ready to go, but there were a few issues. For one, how should we find the Living Wall bomas that are in our random sample? We had GPS points, yes, but we luckily also had the names of the boma owners, which proved to be much more useful. By enlisting the chairman of each sub-village (or their designee) within Narakauwo and Loibor Siret, we were able to easily contact boma owners ahead of time, and find their respective bomas, since the sub-village chairmen know everybody in their sub-village, and know where each person lives. Rather than enlisting more field assistants, we used our budget to pay these chairmen and their designees for each day that they rode with the team and helped find people for the survey. Additionally, although we had originally chosen our paired NLW bomas using high-resolution satellite imagery, we found that it was easier to simply go to the closest NLW boma for each Living Wall boma that we surveyed, and give them a heads up that the team would be coming by the next day to conduct the survey. These two elegant solutions helped immensely, given our small time frame for getting the survey done.

Additionally, we got extremely lucky that we had access to one of TPW’s vehicles for every day we had scheduled for the survey. We had a decent chunk of funds designated for paying for vehicles (normally $150-200/day to rent a vehicle (!), yet TPW gave us a nice discount since we were working on a project relevant to them), but the organization has many projects and we weren’t guaranteed vehicle availability each day. Simply traveling around the bush, over rough roads with plenty of opportunity for flat tires and other vehicle malfunctions, is taxing for both the drivers and the cars, and our survey involved traveling from 7:30am-6:30pm, 6 days a week. Transportation isn’t something to take lightly here in the bush, and I tried to show some extra gratitude by purchasing small gifts for our helpful drivers.

Loibor Siret village center (and one of the flattest/nicest roads around)

This past Saturday was the last day of the survey, after about one rapidly-paced month. We met our sampling goal, employed some great folks as field assistants, and collected a ton of data. It felt amazing to see the survey to completion after months of preparation, and now we get to start our analyses before presenting preliminary results to TPW and to the rest of NTRI at The Nature Conservancy headquarters in Arusha on Monday! We’ve done many more things here, which can be saved for a later blog post, but one thing I’ve learned for sure: There is a lot more complexity than meets the eye behind implementing a survey- especially a survey that tries to discern the connections between human-wildlife conflict resolution and social factors, and especially a survey conducted in a different culture, in multiple languages, in the bush.

A boma in Loibor Siret, on a late afternoon.






One thought on “What’s in a Survey?”

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s