The publishers of the Encyclopaedia Britannica have released an extraordinary rebuttal of a report that said the famous reference work had been matched in terms of accuracy by Wikipedia, its upstart online rival.

The 20-page defence, published on Britannica’s corporate website, tears apart a study published by Nature in December, dismissing the report as "without value".

That Britannica was stirred to publish such a rebuke is bound to prompt speculation that traditional encyclopaedias are battling to remain viable – and avoid being branded vulnerable – in the fast-moving media landscape.

Nature claimed that its research comparing Britannica and Wikipedia, which is written by an army of online volunteers and is open to anybody to edit, found "that Wikipedia comes close to Britannica in terms of the accuracy of its science entries" and that "the difference in accuracy was not particularly great".

However, Britannica responded that "almost everything about the journal’s investigation, from the criteria for identifying inaccuracies, to the discrepancy between the article text and its headline, was wrong and misleading".

The American-based company behind the publication said that several inaccuracies cited by Nature were in fact correct, and that Nature’s figures actually showed that Wikipedia had one-third more errors than Britannica.

It added that "Britannica undergoes continuous revision and fact checking" and in a barbed attack on Wikipedia’s reliance on volunteers, "regardless of knowledge or qualifications", highlighted Britannica’s ties to "thousands of contributors and advisers around the world – scholars and experts all".

A Nature spokesman told Times Online: "We reject all of their accusations and are confident that our comparison was fair."

The publication said in a statement: "We note that Britannica has taken issue with less than half the points our reviewers raised. Both encyclopaedias have made corrections to some of the relevant entries since our article was published. We do not intend to retract our article."

Jimmy Wales, the founder of Wikipedia, admitted to Times Online last year that the Nature study came at a crucial time for the website.

It was published shortly after John Seigenthaler, the founding editorial director of USA Today, who was linked with the assassination of President John F Kennedy by a libellous Wikipedia article, described the site as a haven for "volunteer vandals with poison-pen intellects".

Mr Wales said: "It was good to have this thing out there that said it’s not just this crazy place on the internet where people just post nonsense - that Wikipedia’s actually pretty good in parts."

Wikipedia has admitted that its model is vulnerable to vandals and editors who harbour their own agendas.

It recently emerged that political staff working for in Washington had been "polishing" Senators’ biographical entries on the site and even smearing rivals. To counter such problems, the site is considering ring-fencing some "completed" articles so they can not be freely edited.

Meanwhile, the non-profit site remains a hugely popular resource and has recorded more than 2.5 billion page impressions since its foundation in 2001. The English-language version recently reached 1 million articles and traffic is doubling every four months.

Nevertheless, Mr Wales still counsels caution for users. "If what you’re after is 'who won the World Cup in 1984', it’s going to be fine – no problem," he said. "If you want to know something a little more esoteric, or something that’s going to be controversial, you should probably use a second reference – at least."

And Wikipedia can also point to a brisk revision programme of its own. Following the Nature article, it published a list of the errors found during the review and showed how, by late January of 2006, they had all been corrected.

Which is the more valuable resource: Britannica or Wikipedia? Have your say here.

Times Recommends

• Darling under pressure on record public debt
• Mortgage lending falls 32% as drought deepens
• Oil prices firm after 10% drop to below $130