Saturday, May 08, 2004

Is moving the goal posts a good thing?

I have just been reviewing the changes made in the Logical Framework used in the DFID-funded PETRRA project (rice research with and for poor farmers). Since the project started in 1999 there have been quite a few changes, at least once a year, if not more often. I decided to go back and compare the contents of the first Logical Framework developed in 1999, with those of the most recent version, last changed in mid-2003.

Somewhat to my surprise I found that many of the changes that had been made meant that the current Logical Framework was now more demanding, in terms of the expectations it set, than it was in 1999. There were now three, rather than one Purpose statements (heresy in some quarters, but viewed approvingly by the last OPR). There are now six Outputs instead of five (Communications work was given much more importance as the project developed). Of the original five Outputs, three had clearly developed a more demanding set of indicators. The other two were neutral, if not a bit more demanding. And the total number of indicators for all the Outputs and Purposes had grown from 18 to 30, roughly counted. Now an average of 3.3 per Purpose or Output.

And the project, which is due to finish by August 2004, looks like it will score above average on the achievement of most of the Outputs and Purposes, when assessed by the OPR team in July.

All this leads me to speculate about to what extent we could read changes in the contents of Logical Frameworks, as indicators of achievement (or lack thereof), even before we look at the evidence on the ground, or elsewhere. This might be half true at least. Where the Logical Framework has been scaled down, to be less demanding, that might relect a movement from unreality to realism, or from realism to failure to achieve. But even that difference should be possible to identify by reading the original Logical Framework. Another concern is sampling. How many changes were made to the Logical Framework, and how evenly over a period of the project? Many, spread well over a long period, would suggest the project managers had some ability/right to make changes. Few might suggest the latter. But that could be investigated. And the more changes there are made, the less likely they might be seen as "random pertubations" versus real trends.

Anyway, for thought.