Tony Blair’s memoir, A Journey, has triggered much discussion and a moderate amount of shoe-throwing.

We decided to skip past the bits about Iraq and Gordon Brown, and have a look at what Blair has to say about polling and focus groups.

“Polls are an absolute nightmare,” he writes, as he describes how Labour found itself eight points behind the Tories in the aftermath of the fuel protests of 2000. “All leaders will tell you they don’t pay attention to them,” he says, “but all leaders do.” We can believe that – he manages to use the phrase “the polls showed that…” (or variations of it) dozens of times throughout his book.

The problem with polls, Blair complains, is that you never know whether you’re seeing a superficial reaction that will soon be forgotten, or a deeper, longer-term trend – or whether it might be possible to persuade people of a different view. Unfortunately though, you can’t dismiss the polls because “your supporters and the media dwell on them”.

“They help to create a mood, which itself often then reinforces the polling,” he writes. “You watch any US election and it’s amazing the degree to which the polls create the weather.”

Blair goes on to say that be became less concerned with the polling over time, “which may have coincided with the fact that it became less amicable”. That would seem to chime with the conclusions of Gordon Brown’s former pollster Deborah Mattinson, who claims in her new book that Labour stopped listening to voters when they didn’t like what they heard.

Although Blair praises pollster Philip Gould, who he describes as providing the “divining rod” in his inner circle, he also reveals a lack of trust in focus groups, saying: “I used to laugh at how extraordinary the confluence was between [Gould’s] own thoughts and what the groups seemed to say. Also, so much depended on the individual people. Though pollsters always swore blind these groups were selected on a very ‘scientific’ basis, the truth about any group of people chosen like this is that they are utterly in thrall to their own mood on the day, any recent experience, what they think they should think, and above all to the voice in the group which speaks most definitively and so influences the dynamics that will occur within any collection of strangers sitting in a room together for the first time.”

His concerns about bias and representivity did not, however, stop him turning to his mother-in-law to keep him “informed of the views of pensioners”.

Blair’s belief is that while polls and focus groups should be treated with the utmost caution, “they never are”. From observing the fuss around opinion polls, particularly at election time, we can sympathise with that view, although it’s worth pointing out that it’s generally the media, rather than the pollsters, who are at fault. You could argue that they could do more to prevent their work being misinterpreted, but they are generally more upfront than they are given credit for about the limitations of what they do.

On the subject of focus groups, Blair throws in a tantalising ‘what if’: “I always wanted to attend one secretly and then at the end jump out and confront them with all the vicious calumnies they had just been uttering against me!”

Here, research clients will know how Blair felt. But with hindsight, he was probably wise to stay quiet.