Exam officials believe the provisional figure, understood to be 1.8 percentage points, is low enough to enable them to ward off criticism that A-level standards are falling.
A similar improvement was achieved in both 1992 and 1994. The biggest improvement was in 1994, when the figure was 1.9; last year's increase was only 1.1 percentage points.There has been an increase of between 1 and 2 percentage points in the A-level pass rate in each of the past 15 years. More than four out of five candidates now achieve a pass. However, the proportion securing a grade A is rising more slowly.
This is the first year in which significant numbers of candidates have taken modular A-levels, in which they sit a series of tests over two years plus a final exam rather than one big exam at the end of their course.
Traditionalists had predicted that the pass rate would rise because candidates are allowed to resit modules, but first figures suggest that the effect of the change has been limited.
Although modular exams may have improved the pass rate because weak candidates who fail modules can drop out or resit them, they may make it harder for candidates to secure A grades, for which a consistently high performance is required over two years.
Only about a quarter of all candidates took the modular exams, mainly in English, science and maths. Both the School Curriculum and Assessment Authority and the examination boards point to other reasons for candidates' improving performance besides changes in the format. One is the improved performance of girls at A-level. Since 1989 the percentage of girls gaining three or more passes has gone up from 10.4 to 16.4 per cent. The equivalent figures for boys are 11.3 and 15 per cent.
In a recent paper the authority suggests that another explanation may be an increase of 20 per cent in the past 30 years in the proportion of candidates in the top two social classes. "Since performance at A-level and university entry correlates with socio-economic grouping, the growth rates may well be related," it says.
Alan Smithers, Brunel University's professor of public policy, said: "One reason for improving A-level results may be changes in content to include subjects such as sports studies and media studies. Another may be different types of exam, including coursework and modules.
"Perhaps the key factor may be what examiners regard as acceptable. Each year they look back to what happened last year. Candidates on the borderline are rarely put down."
Tony Higgins, chief executive of the Universities and Colleges Admissions Service, said: "If the results are better than expected then overall more people will make the grade and there will be fewer places in clearing."
Improved results would cause no difficulties in most university departments. The only problem might be over places for medicine and dentistry, where there are strict targets to be met.
The official Ucas listings of places available this year through clearing will appear exclusively in the Independent, starting on Monday.