Scotland's higher education institutions are failing to be sufficiently self-critical when assessing their teaching quality, the Scottish Higher Education Funding Council has warned.
In its second annual report on quality assessment, the council says there is a "major mismatch" between institutions' own view of the quality on offer, and the outcome of the assessments.
But Jim Donaldson, director of quality and learning, said that despite the rapid expansion in student numbers, the 1993/94 assessment round did not indicate any major reduction in educational quality.
Last session, the council carried out 80 quality assessments in nine homogeneous subject areas, giving 58 per cent of institutions worse gradings than those in the self assessments. Only 39 per cent of internal gradings were confirmed, and just 3 per cent won better results than their self assessments.
Mr Donaldson said the overwhelming majority of insitutions had claimed excellence.
"For example, in civil and mechanical engineering, the council had 14 claims for excellent, but only one was awarded," he said.
"A major disparity obviously exists between the institution's view of its quality of provision and the view of external peers."
Mr Donaldson urged institutions to analyse their claims more rigorously before submitting them, and said there was a clear need for staff training in preparing self-assessments.
The council has already trained some 160 academics as assessors. They had looked at the quality of teaching and learning in their discipline in other institutions. By the end of the five-year assessment cycle, around 450 academics would have been trained.
Assessment teams, which visit all institutions in each subject area, generally have five members, four academics and one industrial assessor who gives an employer's point of view.
Mr Donaldson said the quality assessment committee would pay more attention to ensuring that the four assessment categories of excellent, highly satisfactory, satisfactory and unsatisfactory were used consistently across teams and subject areas.
"For example, the 1993/94 exercise resulted in no excellent rating in civil engineering while 50 per cent of the institutions funded to offer chemistry were rated excellent," he said.
"Again, in the nine cognate areas assessed, the full range of the four point scale was used in computer studies only."