Justin Kernoghan / TROCAIRE

IAs of under-evaluated areas of CGIAR research - inception workshop at IFPRI

An inception workshop was held on March 10, 2016 at IFPRI for four studies funded under SIAC activity 3.3 (under-evaluated areas of CGIAR research): (1) alternate wetting and drying (AWD), Philippines; (2) agroforestry, Kenya; (3) brachiaria grass, Colombia; and (4) forest co-management, Guinea (more information on the project page and the events page). As is the norm with SPIA inception workshops, the March workshop was structured to provide detailed feedback on each of the studies, and also convey to the proponents the kind of information SPIA is looking for

The methodological challenge underlying all these studies is that a substantive amount of time has passed since these innovations were disseminated or they have been adopted at a large scale. Without going into the details of each study, below is a snapshot of some interesting discussion points – these are relevant to IAs of such research outputs.

Both the agroforestry and forest co-management studies will rely on remote sensing images to construct a baseline and examine environmental outcomes (forest cover, land use etc.). In the absence of baseline data at the household level, research teams intend to construct a baseline ex post using time invariant variables as well as those that can be easily recalled. A question that remains open and will need to be examined through survey piloting as well as literature reviews is what these variables might be. One can think of household structure and occupation as something that may be easy to recall, subject to a clear definition of what a ‘household’ is and what constitutes an ‘occupation’. Researchers have also used community participatory processes to construct movement of households into/out of poverty (see, for instance, Anirudh Krishna’s work).


An issue for both the AWD and brachiaria grass studies is the challenge in documenting spatial impacts: in case of AWD, one can anticipate that there will be differential impacts on upstream and downstream farmers – they share a water source for irrigating rice fields. Downstream farmers are more likely to experience benefits from AWD – anecdotal evidence suggests that they are more willing to pay irrigation service fee in areas with AWD and this is considered an indirect indicator of anticipated benefits. However, in a randomized experiment where the unit of random assignment (the number of TSAGs) is a limiting factor, one can only stratify up to a certain point and infer something about differential impacts with confidence. Another question the AWD researchers will put substantial thought into is defining the boundaries of adoption and monitoring compliance, even in a case where an irrigation association or its subsidiary (a TSAG) will be randomly assigned to AWD. For instance, what is the risk of a farmer or a group of farmers with access to irrigation valves at the TSAG level violating the wetting/drying approach?

In the case of brachiaria grass, the impact on land use could be in either direction: does improved productivity of pastures reduce the demand for land and therefore deforestation rates, or does it increase the demand for land (that maybe currently forested)? Even in a case where reduced deforestation is observed, it may be a small, indirect effect: for e.g., if land frontiers are already closed or if the grass is only one component of the feed system. In both case studies, there is also the question of soil health: AWD requires a period of drying between irrigation cycles but without yield penalties, and the claim for brachiaria grasses is that it is more productive (higher amount of crude protein and increased availability of green matter for livestock). Then, are the crops taking more out of the soil in terms of nutrients, with implications for resource sustainability in the long run? These are questions worth considering.

As these studies progress we will post updates, particularly on any innovative methodological and data collection approaches. However, if you have examples of studies that respond to the measurement and methodological issues we have raised, please do let us know!