Deeply flawed poll from Migrationwatch is not questioned by the papers

The latest press release from Migrationwatch is swallowed hook, line and sinker by the Express and the Mail - as usual. This time, it's a poll that YouGov have done on behalf of Migrationwatch, which apparently shows the British public don't like immigrants very much. Imagine that.

But, as ever with Migrationwatch, the detail needs examination, because it doesn't really stand up to scrutiny. For a start, a finger needs to be pointed at YouGov for asking two questions which seem solely designed to produce the answers Migrationwatch want.

One was: In general, how concerned are you about the issue of immigration? Are you…? Very concerned, concerned, not concerned, not at all concerned or don't know.

The question is flawed. You can be 'very concerned' about immigration because you are 'very concerned' about the shabby treatment many immigrants receive from both the immigration agencies and the media. But if you say 'very concerned' in answer to that question, it is only interpreted one way by Migrationwatch.

But this gives them the answer that as 36% are very concerned and 43% are concerned, then 79% of the public are 'concerned' about immigration. The question of context is missing, as it is with the other, arguably more important question that was asked.

Which was this: According to official statistics, foreign immigration has been running at a net figure of about 300,000 a year over the last five years (that is, there are 300,000 more immigrants each year arriving in Britain than emigrants leaving Britain). What do you think would be best for Britain?

And the answers to choose from were these (with percentage of responses in brackets):
Net immigration of more than 300,000 a year - 2%
Net immigration of 300,000 a year - 3%
Net immigration of 200,000 a year - 4%
Net immigration of 100,000 a year - 8%
Net immigration of 50,000 a year - 17%
No net immigration (e.g. "one in, one out") - 39%
There should be more emigrants than immigrants - 16%
Don't know - 12%

Based on the fact that the final three answers (ignoring the don't knows) got 72% of responses, and 50,000 would represent a cut of about 84% on 300,000, Migrationwatch and the papers got their 'seven out of 10 voters want immigration cut by more than 80 per cent' line.

Migrationwatch's press release starts: 'A new poll has found that more than 7 out of 10 adults want immigration cut by over 80%.'

The Mail story starts: 'Seven out of ten adults want a massive cut in immigration, a poll has revealed.'

The Express story starts: 'Seven out of 10 voters want immigration cut by more than 80 per cent, a survey has revealed.'

But the survey doesn't say that. The survey asked about net migration being cut, not overall levels of immigration. Can they really not tell the difference?

(For the record, on this survey, the 250,000 cut is around 84%; on overall immigration figures - using 2007's 577,000 total, a 84% cut would be 484,680 people - nearly twice as many.)

It is also worth pointing out that based on this graph from the Office of National Statistics, net migration has not been running at 'about 300,000 a year over the last five years'. In fact the highest year for net migration was 2004, when it hit 244,000. In 2006, it didn't even hit 200,000. So where is there the slightest evidence that net migration is 'about 300,000 a year over the last five years' - a figure repeated in both the Mail and Express? Using a figure 100,000 above what it should be will inevitably distort the reactions, and therefore the answers you get.

And asking the question in terms of using just numbers is meaningless too. Ask those people who want the '80% cut' if they want the fewer nursing staff and plumbers and see if they answer the same way.