There were clearly differences in the development of the REFLECT methodology in each of the pilot programmes. Since REFLECT was not a "prepackaged method" and since it was applied in such different settings this is unsurprising. It is, in effect, through the creative adaptations of the basic ideas in these pilots that what is now known as REFLECT (and which is consolidated in the "Mother Manual") has evolved.

In Bangladesh and Uganda the methodology was adapted to the local situation with considerable success and a high degree of understanding of the basic principles. However, it is sometimes from the mistakes that one learns the most and aspects of the El Salvador pilot proved particularly revealing. In some of the literacy circles in Usulutan it became clear how the REFLECT methodology could be distorted with the graphics being regarded as ends in themselves rather than as a means to an end -and with the literacy and numeracy work not being closely interwoven with the graphics. It is important to note that a similar problem arose to some extent in Uganda with the numeracy work which was not well designed in the original manual and which was therefore separated off and became more abstract than would be desirable. This was not a problem in Bangladesh where the numeracy work was a central focus of the pilot.

In contrast to the above, some of the adaptations and innovations of the pilot programmes do not undermine the basic principles of the REFLECT approach. For example, it is not inherently contradictory for the maps and matrices to be constructed on tables (as was done in most circles in El Salvador) instead of on the ground (as was the norm in Uganda and Bangladesh), so long as movable objects are used. Likewise, the materials on which the graphics are drawn (whether large sheets of manila paper as in Bangladesh or harder cardboard as in Uganda) and even the visual quality of the resulting maps and matrices (meticulous and colourful detail in Bangladesh compared to simpler graphics in Uganda and a broad range of styles in El Salvador), though significant differences to an outside eye, are not in themselves key to the methodology. It is the process of constructing these materials and how they are then used which are the essential things to consider.

It is important to note the different rate of progress in the three pilots. In Uganda and El Salvador, most circles had covered just ten Units (though some had done 15). In Bangladesh, which started later (though circles met 6 days a week compared to 2/3 days a week in Uganda and El Salvador) most circles covered 20 Units. This is not in itself a sign of greater achievement. Programme managers in all three pilots were keen to emphasise that no pressure was put on circles to "get through" the manual quickly. The key is the process. A single Unit could take two days, two weeks or a month depending on the intensity of discussion that the graphic produced and the amount of literacy and numeracy work that flowed from it.

The observations on the problems encountered within the pilot literacy programmes are a very important introduction to the results of the evaluations. They help the reader to see that these were not "specially preserved" or "uniquely attended" pilots or models. Each programme was operating in a very specific moment in time with very real problems. Every literacy programme is unique and so we cannot generalise too much from any one programme. But when we compare the three programmes some wider conclusions can be drawn.