A "health check" on the pound;2 million applied educational research scheme, run by a consortium of Edinburgh, Stirling and Strathclyde universities, has suggested it spread its funding too widely on too many small projects and lacked focus.
A report evaluating the five-year project suggests its strategic leadership should have been more centralised and limited funding concentrated on a handful of projects.
But the team leading AERS insist they have been right to devolve as much power as possible to the networks pioneered by the scheme to involve policy-makers and teachers with academic researchers.
The scheme's funding runs out in December and Pamela Munn, chair of the AERS management and executive committees, said discussions were being held with the Government about what, if anything, might replace the scheme.
Professor Munn described the work as "pioneering and innovative" but conceded: "As we have worked through it, we have been conscious of trying to do a number of different things and perhaps we have been over-ambitious - but there is nothing wrong with trying things out."
One unnamed contributor to the evaluation said: "I felt the work that was being described, while being interesting, wasn't asking the questions that needed to be asked.
"It was set up with the needs of researchers in mind rather than the teachers ... it hasn't focused on learning in the classroom."
The scheme, set up in January 2004 with funding from the Scottish Executive and the Scottish Funding Council, aimed to improve the infrastructure of educational research in higher education institutions and to enhance long-term educational needs. It set up four networks. Three focused on substantive research themes: learners, learning and teaching; school management and governance; and schools and social capital. The fourth dealt with research capacity.
"While the overall resource allocation of pound;2 million appears substantial, the number of activities is such that at the level of individual projects, resources are very stretched," state the evaluation team from Cardiff University.
Midway through the project, there were relatively few "hard" indicators of research, such as peer-reviewed journal publications, and the work under AERS was unlikely to contribute to universities' ratings under the 2008 research assessment exercise - the five-yearly judgment of quality, which has a significant bearing on how much research funding universities receive.
The report states: "There is some concern about the general quality of the research being undertaken. But in general, our overall finding is that it is probably too early to predict the number and quality of research outputs at this stage."