When is quantitative data collection appropriate in farmer participatory research and development? Who should analyse the data and how?