Study design

It is useful to consider here to what extent this trial comes under the category of participatory research. The majority of the data were collected by the research scientists, and not in a participatory manner. The researchers decided on the questions they wanted to ask and designed the questionnaires to collect the information they required. They analysed the data.

On the other hand, farmers were not prevented from making adjustments to the recommended concentrate feeding guidelines. Thus, the researchers had no control of how the feeding protocol was implemented. So, in this sense, the trial was participatory.

Indeed the level of participation that occurred resulted in essentially only one 'treatment' being applied without a control, when two (a treatment and a control) were planned and included in the design at the outset.

Thus, the idea of being able to analyse a study with the conventional, classical approach of comparing the means of a treatment and control were thwarted. Without the existence of a formal control, analytical tools needed to be devised to provide interpretations that looked at variability and took into account changes over time.

This will be a common feature of experiments where some participation by the farmer is expected. Researchers may need to be flexible in changing their approaches to data analysis from those envisaged when a study is first designed.