The Sokoine University of Agriculture Laboratory for Interdisciplinary Statistical Analysis (SUALISA) project is an initiative led by the Department of Statistics at Virginia Tech University as part of the Innovative Agricultural Research Initiative (iAGRI). SUALISA is one of 20 statistical laboratories created as part of LISA 2020, a program established by LISA (Virginia Tech’s Laboratory for Interdisciplinary Statistical Analysis) to train statisticians from developing countries to become collaborative statisticians who will help enable and accelerate research to solve real-world problems and build a network of 20 statistical collaboration laboratories by 2020. SUALISA was the second of five statistical labs established in Nigeria, Ethiopia, and Brazil under the leadership of Professor Eric Vance.
SUALISA began operations in December 2014 and since then, has linked researchers across disciplines for 58 projects grounded utilizing advanced statistical methods. These have included projects on Gendered Decision-Making at the Household Level among Smallholder Farmers in Arumeru District and Agriculture Services Support Programmes and Socio-Economic Empowerment of Rural Women in Zanzibar. In addition, SUALISA provide consultations for students interested in incorporating advanced statistical analysis into their research.
One such student was Winfrida Mayilla. Currently a PhD candidate at Sokoine University of Agriculture and the University of Copenhagen, Denmark in area of Food Safety and Public Health. During her fieldwork in Morogoro, she came across some difficulties in analyzing her data. Her research centers on “Farmers, vegetable traders, and consumers perception of using irrigation low-quality water for vegetable production in Morogoro, Tanzania.”
“Currently there is limited information on what people think about the use of low quality water and what their perception is and even less is known about the benefits they gain when using this water for vegetable production. I wanted to know what people think of this water first and the perceived health hazards when using this water for vegetable production. Talking with farmers, vegetable traders, and consumers, we were able to determine what the perceived benefits and the health hazards are and whether or not consumers are willing to use the vegetables produced with low-quality water.”
“When I started writing my first paper and sent to reviewers, they came back with a number of issues. One of them was that I should analyze my data in a different way. One of my supervisors Dr. Magayane told me about the stats lab. I went to meet Dr. Kazazuri and made an appointment. He helped me to solve a number of issues. They suggested that I analyze the data using principal component analysis (PCA). I was using a 5-point Likert scale and I didn’t know how to combine the scores. I read about using mean total likert score for data involving perception scales. I didn’t know how to do it but when I consulted the Stats lab, they told me what to do. Once I incorporated that analysis, my paper was accepted and was published.”
Winfrida Mayilla, Bernard Keraita, Helena Ngowi, Flemming Konradsen, Flavianus Magayane Perception of Using Low-Quality Water for Vegetable Production in Morogoro, Tanzania. Journal of Environment, Development and Sustainability, 2015.
“I had learned about testing perception scores before because my supervisor, Dr. Flavianus Magayane. He suggested that I use the likert score and so we constructed the study together. After we received the reviewers comments, we realized we needed to add something more. We used a uni-dimensionality test through PCA to test for validity of the perception scale items.” The stats lab helped me to address this in the paper. It didn’t change the results, but the reviewers wanted to be sure that the results were valid and reliable for measuring the perception scale. After applying the test, the quality of the paper improved. Previously I only performed a reliability analysis. But when we incorporated a validity analysis combined with PCA, the paper was much stronger and suitable for publishing. After the paper was published, I did not use the stats lab again; however, I used the same test for my second paper. The reviewers didn’t have as many questions and it was accepted more quickly.”
Winfrida Mayilla, Flavianus Magayane, Bernard Keraita, Helena Ngowi, Perceived Health Hazard of using Low-Quality Water for Vegetable Production in Morogoro, Tanzania, January 2016, Journal for Environment and Pollution.
As for her advice to students, Winfrida stated the following, “I took research methodology previously during MSc and BSc. I finished my Master study in 2006. It was a long time ago. It was not easy to remember everything about statistics. I had to start re-learning again, reading, to remember the statistics I had learned. I didn’t practice much. Unless you have something to do with it, then you remember. I think SUALISA provides a good service and is a very good program. It really helps with analysis.”
The SUALISA is located on the second floor of iAGRI and is available to all students and researchers. For more information about SUALISA, read the 2014-2015 Annual Report and visit their website here.