Field reflections: an experience of RHoMIS enumerator training in Uganda

Sam Adams
​Sam works on the RHoMIS team in Bristol, UK, building surveys and looking after user experience. He grew up in England, Zimbabwe and South Africa and has a background in agroecology, training, and project management.

​I’ve worked with the RHoMIS survey for over a year now, built close to thirty applications, and worked with research teams in over twenty countries. Yet my trip to Uganda last week was my inaugural experience with applying the tool in situ. The trip gave me a first-hand experience of some of the challenges and nuances of local application of a survey.

The research in Uganda is part of a two-country nutrition study delivered by ILRI. Earlier in 2019, the data collection kicked off with 200 households interviewed in the Thai Nguyen province in Vietnam. My colleague, Nils Teufel, leads the project and joined me in Uganda. 

After various flights, hotels and road travel, including a ferry crossing of the Nile (an unexpected treat), the research team finally congregated in the small rural town of Kamuli. The point of our gathering, at the Kyembe Sande Garden Hotel, was to localise RHoMIS for the specific Ugandan context, and train the two enumerators who will be conducting the research.

As with all RHoMIS training, the three days were an introduction to digital data collection, to Android tablets, to ODK, and to the RHoMIS survey. During the training, we comb through each question and every possible answer available to the farmers, checking for local relevance. 

It is the local relevance that struck me afresh this trip. Part of the customisation of each unique RHoMIS is a localisation form, which asks for information on local geography, units, crop types, livestock, and so on. Yet even with this initial information gathering, the local nuances show more complexity.

​In Uganda, it appeared that this complexity is particularly acute. Uganda is a country with a very high population density. The landscape is made up of smallholdings of one or two acres and each house is surrounded by crops. There is no house-farm divide, or a ‘kitchen garden’ to speak of, as is asked in one of the standard RHoMIS questions. 

Adding to Uganda's complexity, there is no clear seasonal difference: the equatorial climate means that crops are cultivated and harvested consistently throughout the year. The rich soils mean that crop diversity is very high, making it hard to differentiate ‘most important’ crops. 

Ms Namuganza answering the RHoMIS survey
​on her farm near Kamuli, Uganda. 

On one of the days of our trip, we visited Ms Ruth Namuganza’s farm. Illustrative of the local crop diversity, Ms Namuganza is growing banana, pawpaw, maize, coffee, mango and numerous types of leafy vegetable. These are all important to her household. A brief walk through the smallholdings along her road revealed small patches of many other crops, including lemongrass, jack fruit, napier grass, sorghum, and herbs such as perennial basil. 

A challenge we have when collecting data of such diverse farms is knowing what crops to focus on. With RHoMIS, we ask farmers to isolate a maximum of four or five. But for Ms Namuganza and her neighbours, this was not a straightforward question. The family had one primary crop - matoke, or the staple plantain banana - but the rest of the crops were of equal supplementary importance.

Compared to other RHoMIS applications, the climatic and farming-system nuances we found in Uganda are uncommon. In most other instances, there is a single growing season and a lower species diversity. The adaptability of RHoMIS allowed us to make changes to the survey during the training to improve local relevance. 

Another local nuance that we observed in Uganda was the practice of giving ‘tokens’ to thank interview respondents. Some soap, or a bag of sugar, was typical. This is counter to my training in research methodology, with concerns that the giving of these items will add bias and influence to the data capturing. 

This issue is part of a larger debate, but for me it was a reminder of the local cultural practices, sensitivities and variables. The reality is that RHoMIS was taking an hour of the farmer’s time and this was a small but much-appreciated gift to thank the farmer for her time. 

There were other questions that arose during the RHoMIS training. Responding to these issues showed how essential it is for enumerator training. One example concerns the timeframes used in the survey questions.

The default timeframe for RHoMIS questions is 12months, for example "How many kg of carrots have you harvested in the last 12 months?" This has proven to be clearer than ‘in the last year’, which has the ambiguity of calendar year, farming year, or another calendar system such as Islamic or Chinese.

This data is important for us to gather, however, it is difficult for farmers to recall with accuracy, especially for household consumption. For example, in Kamuli we asked the weight of cassava harvest over 12 months. However, we were told that the entire harvest was for home consumption and was not weighed (compared to the weighing that happens when taken to market).

We also saw that crop weights in general are hard to measure as some crops are dried first and there needs to be clarity on whether the survey is asking for fresh or dried weight.

Species rich intercropping farming system in Uganda (Nils Teufel)

These nuances showed how essential it is to take time to train enumerators and provide the clarity on what they were asking of the farmers. I appreciated how this is a two-way process, as the enumerators and farmers could inform the survey with new questions, for example, that specify 'dry' or 'fresh' crop. 

​This process of seeing the research first-hand gave me a deeper understanding of the variabilities of farming systems across both time and space. This in turn has helped me to better design the RHoMIS surveys including the dependency logic and flow of questions.

Even with my background in social anthropology and having visited Uganda for 25 years, I was struck by the diversity of these farming systems and the cultural nuances I have described above. I saw first-hand how essential the training of enumerators is. Finally, the real-time editing of the survey, informed by local feedback, showed just how valuable and adaptable the RHoMIS system is.