Be assured that readers will draw their own conclusions, both about your "warning" as well as the data and the methodology (or the purpose of such studies) . I am one of your readers.

I also saw the "raw data" and that is why I claim the methodology is highly questionable. There is no cross-tabulation of responses except for the sample's status (which, in this context, is the least important of all characteristic). We still do not know how much those who are willing to pay would pay. We do not know which form of payment they would prefer. We do not know how much, after all, would pay those who first claimed they would switch to another free news site, and how this affects the validity of the earlier negative responses, etc.

The very fact that you ask about willingness to pay in the context of three other suggestive options may form bias. And why these three options? Why not more, including switching to TV or radio as the main source of news? Why not an open-ended quation? Why don't you simply ask yes or no, and then try to find out what other choice each of the groups would consider, if any?

As you must know, asking wrong (leading) questions can be even bigger a source of bias than reporting only partial results — and partial they are, indeed.

"You said in the introduction to the first part, The results will warn the publishers who consider the paid-content strategy against introducing it. Considering the shown “results,” this reveals a clear agenda."

– It reveals nothing of the sort – that's a *conclusion*, not an *agenda*. Our readers know that we are dispassionate but healthily skeptical about all kinds of trends and issues. Many other readers would assume we're paid-content *advocates* because of the site's name. We are neither "for" nor "against" – we're just here to weight it all up.

"Just a question Robert: How do you code an answer … Harris Interactive owes us an explanation of the methodology and full results."

– As I said, not only have we published the methodology statement, but the full questions asked and data received are available, which is rather rare, via the link that's on every post in this series. Harris Interactive's methodology is available on its website, which we've also linked to every time (but here it is – http://www.harrisinteractive.com/partner/methodology.asp).

Compared to most reports of research, we have been remarkably transparent.

You said in the introduction to the first part, The results will warn the publishers who consider the paid-content strategy against introducing it. Considering the shown "results," this reveals a clear agenda.

Just a question Robert: How do you code an answer I would pay maximum 25p a day? Was it in your survey entered as a "Yes, I am willing to pay," or "No I would switch to another free site." You cannot say that ONLY5% is willing to pay, and then report that 68% [of the same population] will pay less then 10L, 20% between 10and 20L, etc.

Harris Interactive owes us an explanation of the methodology and full results

The *purpose*? Are you kidding? It's kinda the hottest topic in online media right now.

The *methodology* about asking people "who already declared they are not willing to pay"? It's perfectly valid to ask *everyone* who reads a news site at least once a month how *much* they would want to pay – if that is minimal, then so be it. We asked how much people would pay, whether they *chose* to pay or *had* to pay, in a hypothetical situation where all news sites go pay-for. The sample size from those who *chose* naturally becomes quite small.

"It seems that the Poll is not to inform us about readers intentions, but “to warn us against paid content initiatives."

It was *absolutely* "to inform us about readers intentions" – we didn't go in to this with an agenda, and would have reported *whatever* conclusion the results had yielded.

It seems that PCUK/Harris Poll focuses on only one group of readers: those who gave "the LOWEST possible amount in each category," or those who read a free news site at least once a month. So large a group must include, the so-called fly-by audience (those who visit a news site only once a month) mixed with more regular readers and core loyalists (who visit the site many times a month or even daily).

Other studies show that the fly-bys are the largest group (up to 54% of all online visitors); however, they drive only about 2% of all pageviews. They are also the least interested in day-passes or subscriptions, as normally they visit many such sites. Mixing them with the regular and core readers, who deliver up to 89% of pageviews makes no sense — unless you have a hidden agenda to bias the results against paid content.

Equally questionable, and frankly ridiculous, is collecting answers about maximum possible payments from people who already declared they are not willing to pay, because this is what the Poll seem to have done: Are you willing to pay? No! How much are you willing to pay for a daily pass? 20p! What kind of methodology is it? How can one reconcile such answers into a set of data?

The valid procedure would be to ask the 5% who are willing to pay how much they would spend, given the various options. But then — again — it seems that the Poll is not to inform us about readers intentions, but "to warn us against paid content initiatives. "

I would love to see a comparison with how much an article currently would be worth in the breakdown of the Guardian's 80p (+ad revenue). It can't be much more than 1 or 2p as it currently stands, and printing/distribution costs a whole lot more in paper form.

The only stat there that I see as surprising is the £10 per year fee, which seems amazingly low even as an option compared to the other two.
£0.25/day = £91.25/yr – even with a discount for bulk, surely this would be offered somewhere closer to £50/yr?

I would happily pay £50/yr for a complete digital copy of the Guardian every day except for the fact that currently there is no decent alternative for reading it on whilst commuting etc. The current crop of Kindles and similar just aren;t there yet in my opinion.