Well, what happened, I mean, we first worked with MORI, what is now called MORI, and then Populace on polling. There was a slight gap in between when we didn't do polling for budgetary reasons, but effectively those two firms. What happened is I would discuss with a senior executive of the two firms what questions had been asked -- some of the standard questions -- you would always ask about voting intention and that involves about six different questions to get a proper voting intention, and some others were kind of regular ones. Then we would add on questions and it would a dialogue between myself and the person.
Would we report all answers? Not necessarily, mainly for space reasons. However, what we did do, certainly in the latter period, is put all the answers on the internet. What would is after the results were published -- sometimes they would be published over two or three days, mainly so we got maximum bang for our buck on that. The person concerned, say when I was working with Populace, would say, "Can we put it all on the website?" So even if there were two or three questions which weren't published, they would all be there available to see. And the reason they weren't published -- it was purely for space reasons. I mean, I took a judgment on what the story was. There would often be a dialogue with the news desk but also, crucially, Populace and MORI were members of the British Polling Council, which has very high standards about how polls are represented, partly to do with phone-in polls, phone-in things and things like that, but it has very high standards for polling.
I was actually, for a period, on their kind of ethics panel. We had to rule on what -- the use of one poll. But in general, there were high standards about how they could be presented, if you presented a sample size, when it was done, all that. But they wouldn't all be published, entirely for space reasons, but they were accessible to readers.
And the question -- what was very clear at that time, for example -- and this isn't always true -- is that people could see what the actual question being asked was, because often that could be distorted quite easily.