Findings, implications and lessons learned

• Courses in biometrics and research methods rarely deal with climatic data. Such data are useful in their own right and can be an important component in research. Climatic data are an example of routinely collected data that need to be carefully handled and require special methods of data management.

• Courses in applied biometrics tend to emphasise formal methods of analysis and go straight to the fitting statistical models without any preliminary exploration of the data. A simple descriptive analysis is sometimes all that is needed to answer a question, as is shown in this case study. Case Study 11 makes the same point.

• It is necessary for a biometrician to have a continual dialogue with a researcher during any consultation if he/she is to have a full understanding of the background of a study and the questions that are required of the data. In this way suitable approaches proposed for analysing the data in a way that best answers these questions can be discussed and agreed. This was important in this case study for agreeing on a definition for the ´Start of Rains´ and a suitable length for a dry period following the ´Start of Rains´.