There were large variations in how the "esteem" of academics was rated across different subjects in the 2008 research assessment exercise, according to an analysis by Times Higher Education.
The analysis shows that departments in the arts and humanities were more likely to earn a perfect rating from RAE judging panels for the esteem of their academics than those in the sciences.
Similar variations in the panels' ratings of departments' "research environment" were also revealed - with both indicators affecting a department's overall RAE score.
The analysis is based on the fine detail of the 2008 RAE results, which has now been published by the Higher Education Funding Council for England. A detailed "quality profile" for each department has been published, breaking down its performance in "research outputs", "esteem indicators" (which gauge the standing of a department and its academics in their field) and "research environment".
On each aspect, the proportion judged to fall into one of five categories, from "world-leading" (4*), through "nationally recognised" (1*), down to unclassified, is given.
The three measures were used by the RAE judging panels to calculate an overall quality profile for every department, the results of which determined the allocation of more than £1.9 billion in research funding across the UK.
The Times Higher Education analysis shows that about one in five departments in some arts subjects was judged to have 100 per cent of its academics in the "world-leading" category when it came to esteem. In some science fields, no perfect scores were awarded for esteem.
Departments judged under the RAE's main panel N - covering history, Classics, philosophy and religion - were among the most likely to receive a perfect score for esteem.
Of 179 submissions, 22 per cent received 100 per cent for esteem, and 5 per cent scored full marks for research environment.
Of 241 submissions to main panel O - covering art and design, history of art, drama and performance, media studies and music - 18 per cent received full marks for esteem and 6 per cent received full marks for research environment.
Main panels A and B - covering various aspects of medicine - also score highly for esteem, with about 17 per cent of submissions gaining full marks.
In contrast, no perfect scores were awarded in the submissions to eight main panels covering physical sciences (panel E), biological sciences (D), subjects allied to medicine (C), mathematics (F), engineering (G), geography and architecture (H), economics and business (I) and education and psychology (K).
The 15 "super panels" set up to oversee the 2008 RAE have already been accused of failing to deliver comparability between the subjects.
Malcolm Grant, who is provost of University College London and chairman of the Russell Group of large research universities, implied last month that some arts-based subjects were given an easy ride.
A Hefce spokesman said that no expectations had been set across the main panels about the proportion of submissions that would receive certain scores. Although the council was "generally expecting" the results to be broadly comparable, it was also expecting some variation.
"The results are broadly comparable," the spokesman said, pointing out that the average grades across all the main panels were roughly the same.
Peter Golding, chair of the media studies panel (66), said that any analysis of grade-point averages would show that accusations that arts-based subjects had inflated their grades were a "myth".
International comparisons had shown the UK to be world leading in many arts and humanities areas, he said. "We should be celebrating (media studies), not knocking it."