CApStONE is a programme to develop research capacity in applied statistics, in partnership with research institutes and universities in southern Africa.  It is a collaboration between the University of Nottingham, UK, where it is led by Prof. Murray Lark,  and colleagues at the University of Zimbabwe with plans to expand to include Lilongwe University of Agriculture and Natural Resources (Malawi), and the University of Zambia.

The CApStONE programme builds on work that these partner institutions have done together.  Some has been specifically focused on capacity strengthening, including the development of statistical training on the R platform, "training the trainer" development of skills, reflection on the statistical training process, and the delivery of statistical consultation to support the design of surveys and experiments and the analysis of data.

One such project is the CEPHaS project, to develop research capacity for studies on conservation agriculture.  More information about CEPHaS can be found here   A review of the capacity strengthening impact of the project by the Liverpool School of Tropical Medicine's Centre for Capacity Research can be found here.

At present we are collaborating on the development of curricula for statistical training of post-graduate researchers, mainly in agricultural and environmental science.

The first principle of CApStONE is that scientific work should be done collaboratively between researchers and statisticians.  Statistical training for those in specialisms like soil science, agronomy or health sciences should build awareness and understanding of statistical issues, but cannot equip non-specialists to deal with all the challenges that data sets present.  Research capacity strengthening requires that we develop strong relationships between scientists and statisticians, and experience of how the disciplines can work together.

Following from this, we consider that the most important principles for scientists to understand are those of sound experimental or sampling design.  Design is foundational,  a poor design will, at best, be a waste of resources and, at worst, will not provide any meaningful results.

We believe in the value of reproducible science, facilitated by open-source platforms for data analysis, in particular the R platform.

A principle of the CApStONE programme is Equitable partnership.  In all work to date partners from the global north and south have worked together to develop joint capacity to address problems.  Our view is that all research in countries of the global south should be in service of priorities identified by citizens of those countries, and not by funding bodies or scientific establishments from the global north. 

There is often a gap between the statistical methods which students learn as part of agriculture, environment and allied degrees and the complexity of real surveys, experiments and data sets. 

Statistics is therefore often a “bottleneck” for the production of good research. In recent projects, including CEPHaS , GeoNutrition and a project to strengthen  Capacity in Geochemistry in partner institutions, University of Nottingham, University of Zimbabwe, University of Zambia, Addis Ababa University and Lilongwe University for Agriculture and Natural Resources (Malawi) have worked together to develop a statistical training programme to address this problem.

The aim is to produce capable researchers, enabled to collaborate with statisticians, and aware of the fundamental importance of sound design, good data management and statistical methods which are appropriate for the design and the objectives of the research.

This statistical training has been developed into a postgraduate module taught at University of Nottingham, and is in its second year of delivery as an expanded module for postgraduates in the University of Zimbabwe.

Within the programme we have experience in developing experimental and survey designs for use in the field.  (Underlined text below links to papers and reports).

Surveys have been designed for the examination of shallow groundwater pumps across Malawi, Uganda and Ethiopia .  This design required engagement with a range of stakeholders to address logistical issues, requirements of NGOs and the priorities of funders and governments.  We also undertook the eventual data analysis.

We have studied how various stakeholders interpret uncertain spatial information and use it to make decisions, and evaluated different strategies to communicate the uncertainty more effectively.

Surveys were also designed to support the investigation of spatial variation in soil and other environmental constraints on the micronutrient content of staple crops across Malawi and the Amhara, Oromia and Tigray regions of Ethiopia.  This resulted in new insight into the regionalization of deficiency risks and implications for efficient interventions.

We have supported the analysis of legacy data sets at continental scale across Africa to investigate the resilience of groundwater supplies, and at national and regional scale in Zambia to compare sources of soil information and to evaluate their accessibility and quality. 

We have supported experimental design for studies on agrofortification of crops with micronutrients, both on station and on farm, using power analysis and developing tools to facilitate correct randomization.  We have developed resources to support experimenters.

We have facilitated the use of preregistered designs for experimentation and for studies on sampling methodology.

We have developed and applied methodology for expert elicitation, including the assessment of geohazards, assessment of data quality  and interpretation of soil information by experts.

We are collaborating with scientists at CGIAR centres to develop robust statistical models for yield gap analysis

We are interested in collaboration with researchers, NGOs, research institutes and others who share the principles outlined above.  In particular we are interested in opportunities:

• To develop and deliver training programmes 

• To provide statistical consultation to researchers, both remotely and in person

• To deliver support in experimental and survey design

• To support the collection and evaluation of spatial information through surveys

• To collaborate in methodological research, and challenge-led studies

If you are interested in collaboration with the CApStONE programme, then contact Murray Lark on:

murray.lark@nottingham.ac.uk