Can we ever be sure how many copies of newspapers really find their way to readers? Perhaps not.

“I would just ignore circulation figures. They are rubbish.” So said the guru of newspaper readership measurement, Gary Morgan yesterday when Crikey asked him about the increasing anomalies between Roy Morgan readership survey figures and audited circulation.

Courtesy of the media consultants Fusion Strategy, Crikey has an analysis of circulation and readership figures over the last five years. The figures show startling anomalies between readership and circulation, particularly among national daily newspapers and particularly over the last few years.

The figures are really strange – which raises several questions, including whether Morgan’s figures are reliable, or whether he is right and circulation figures are the ones that are suspect. Or perhaps they are both bodgy.

In 2007 the Weekend Australian Financial Review lifted circulation by 3.33%, but according to Morgan’s figures, readership dropped by 13.87%.

The weekday Australian, meanwhile, was basically steady in circulation with a lift of 0.25%, but readership went up 11.49%.

There are also anomalies among the national dailies. The Sydney Morning Herald dropped .04 in circulation, but lifted readership by 7.66%. The Daily Telegraph dropped 3.59 in circulation, but lifted readership by 2.79%. And so on. There are unexplained variations all over the place.

Here is the full chart, with the more startling variations circled.

Newspaper Circulation – Year average & movement analysis 2004 – 2007

 Newspaper Readership – Year average & movement analysis 2004 – 2007  

Steve Allen of Fusion says the overall trend is circulation down by about 1%, and readership up. Circulation and readership moved more or less together in 2005, but since then the anomalies have been increasing. Allen says that on these figures it is no longer possible to say that increased circulation leads to increased readership, nor that increased readership leads to increased circulation.

So what’s going on? Are the readership figures bodgy, the circulation figures doctored, or is there something new happening in the market?

Morgan referred us to this 2003 paper in which he gives reasons why readership and circulation figures do not always track each other.

In conversation yesterday, he blamed suspect circulation figures. Deep discounting and giveaways tend to mean readers get a paper but do not necessarily read it.

At the same time he said some of the anomalies might be because circulation figures are “more honest now” and the publishers have been forced to admit declines – but says “they still get around it” by counting copies as sold that are effectively free in the hands of the consumers.

Morgan also acknowledges that online readership may lead to some misreporting in his surveys. A reader who has seen the paper only online may answer that they have “read or looked in to” it when answering one of Morgan’s surveys. He says he will be amending the methodology to address this.

Meanwhile a bit more analysis on the figures by Crikey leads to this graph, tracking the number of readers per copy of circulation for some of our major newspapers.

As can be seen, they hop around a fair bit. Are people really sharing newspapers more widely since 2005? Are there more multiple reader copies circulating in offices and cafes? And do readers of the Sydney Morning Herald really share the paper more than readers of The Age? It is hard to see why, though Morgan assures us that it is so..

As for Allen, he agrees that the anomalies are something to watch, though quite what it all means, nobody can be sure.