Part of a series looking at the spin behind a Daily Express article – see this post for details.
Now that we have established that Turkey is not about to join the EU, and even with the visa-free travel scheme which might be implemented for the Schengen Area, Turkish citizens are not allowed to move to the UK, it might seem a little pointless examining the Daily Express’s claims and survey findings. However, comparing the newspaper article to the actual survey that was commissioned is a useful exercise in showing how statistics can be misleading and how, very often, journalists don’t understand the numbers they are reporting on.
The full survey can be found here. It has been undertaken by Konda, a Turkish market research and survey company, who should be ideally placed to ensure that the sample was balanced and fair. The full survey methodology is included in the report, and shows that 2685 respondents were used to collect data, spread over 153 villages or neighbourhoods. In each village or neighbourhood quotas were used to ensure the results were not swung by outliers – but this quota number was only three respondents for each age group and sex. It is possible that some of the more disaggregated analysis may be adversely affected by this, but for the sake of this post it is acceptable.
Size of sample
Is a sample of 2685 too few to draw conclusions from? When compared to the population size of Turkey – 74.9 million, it is a very small figure, meaning that only 0.0036% of the population has actually been surveyed. Importantly, Konda’s report does not at any point give any indication what the overall figures would be for the country as a whole. While it is likely that the sample was chosen to try to represent the Turkish population, there are indications that it was selected so that disaggregated results could be obtained. For example:
- “The sample was selected through stratification of the data on population and educational attainment level of neighborhoods and villages based on the Address Based Population Registration System (ADNKS), and the results of the 2011 General Elections in neighborhoods and villages” (p6) – this implies that when choosing the sample, people from all political parties (and potentially religions) were included in each neighbourhood, even if they comprised a tiny proportion of the total.
- “Among the 18 surveys conducted in each neighborhood, quotas on age and gender were enforced” (p6) – it is unclear from the phrasing if these were strict limits or simply minimum numbers. If the former, it is imposing rigid constraints on the make-up of Turkish society. If the latter, it may have impacts in some areas where one of the groups is in reality significantly underrepresented.
In the report, Konda state:
The survey is designed and conducted with the purpose to determine and to monitor trends and changes in the preferences of respondents who represent the adult population above the age of 18 in Turkey. The margin of error of the survey is +/- 1.7 at 95 percent confidence level and +/- 2.3 at 99 percent confidence level.
Without seeing the full data, it is difficult to verify these claims, particularly since the unit for the margin of error is not defined. The normal way of calculating standard errors for surveys gives a different answer of 1.38% at 95% confidence – this can be verified using the second tool on this page using values of 2685, 74,900,000, and 15.8% for sample size, population and percentage respectively. In any case, this indicates that the size of the sample, assuming it was randomly chosen, should give an idea about intentions of the population at large.
The Express’s claims
While the survey itself appears to have been carried out professionally, this cannot be said about the Express’s interpretation. It is worth looking at each of the statements in the Express’s article, and comparing it to the evidence presented in Konda’s report.
The Express says:
“12 MILLION Turks say they’ll come to the UK once EU deal is signed” … “more than 12 million Turkish citizens are planning to move to Britain when the country joins the European Union”.
Konda’s report shows that the question asked was in fact “… would you, or any member of your family, consider relocating to Britain?” This is nobody saying they will relocate, this is nobody saying they have plans to do so. This is a group of people saying they may consider it. If Turkey becomes a full member of the EU. If the UK remains in the EU. This is willfully dishonest phrasing by the Express.
Indeed, in the next sentence the Express corrects themselves:
“almost 16 per cent of Turkish adults said they would consider re-locating to the UK once their country becomes a full member of the EU”
Their correction doesn’t go very far, and again contains the determination that Turkey will become a member of the EU, which is incredibly unlikely as shown in my previous post. This is a lie that they continue to push to their readership, hoping that repetition will create certainty.
“According to the poll, which saw more than 2,600 adults interviewed across all 27 provinces of Turkey, most of those keen to come to Britain are either unemployed or students”
They have the number of respondents correct, but after that the statistics fall down. The statistics included in the Konda report do not show that the majority of people who are considering relocating (again, not “keen to come to Britain”) are unemployed or students. The table on page 2 of Konda’s report instead shows what percentage of each demographic group are considering relocating, and the two groups who have the highest propensity to consider it are students (with 34%) and the unemployed (with 36%). This makes sense, of course, since seeking work or education is a key reason to relocate.
This does not mean that 70% of people considering relating are students or unemployed – this can easily be verified by looking at the total percentage over all employment statuses, which comes to 176% – this is a meaningless number.
Instead, we can get a better idea of the makeup of those considering relocating by looking at the makeup of the survey panel contained on page 9 of Konda’s report. Of those surveyed, 9% were students and 4.1% were unemployed. Combining the tables on page 9 and page 2, we can see the relative makeup of the 15.8% of respondents who said ‘yes’.
To get the total number, we must work out the sum of products for each category. This is made difficult due to the fact that for some categories there are too few observations, so the percentage saying ‘yes’ is not reported. If we simply ignore these, the total percentage is 13.94%, which is close enough to the 15.8% reported to show this is a valid calculation.
Calculating the sum of products for unemployed and students gives a figure of 3.64%. This is the percentage of the total population that said yes and is either a student or unemployed. Dividing one figure by the other shows that only 23% of those saying they would consider relocating were unemployed or students. This is not, by any definition of the word, ‘most’.
“Today’s poll reveals for the first time the intention of those living in Turkey if it joined the EU, with 15.8 per cent of the population – the equivalent of 12.6m people – in favour of making a new home for themselves and their families in Britain.”
The population of Turkey is 74.9 million. If the survey was scaled up to the entire population, this would give a number of 11.83 million people who would consider relocating to the UK. This figure would include the families, whereas the Express implies that the families would be coming as well.
“Foreign Office Minister James Duddridge said: ‘These figures reveal the huge pressure our public services will come under if we stay in the EU. We will be powerless to stop millions of Turks accessing our schools, hospitals and housing, who will be attracted to the UK by our higher wages and living standards.'”
This quote makes the several leaps of logic that are discussed in this series of posts in one go. It assumes that Turkey will join the EU, it assumes that the UK will have no power to stop migration, it assumes that this will put pressure on services. It is a classic case of scaremongering.
“Prominent Tory eurosceptic Peter Bone added: ‘These figures demonstrate the true danger of remaining in the EU. Our public services are at breaking point and our housing stock depleted with the current situation of uncontrolled migration. Our country cannot accommodate another 12 million people. Unless we leave the EU, there is absolutely nothing we can do to stop this influx of Turks.'”
12 million people would consider relocating. Only a small fraction would actually do so.
“lurking amongst the results were 10 per cent of religious hardliners who would also consider emigrating and nine per cent of women who wear head coverings such as the burka.”
The way piety is defined does not match up to the Express’s description of “religious hardliners” – those in the ‘pious’ category are those “who completely fulfils the requirements of the religion”. The Express’s language is designed here to create links in the reader’s minds to extremists and hate preachers, who have been demonised over the past few years by the papers.
There are many questions in the Konda report which are not discussed by the Express, likely because they do not satisfy their agenda.
- The group most likely to consider relocating are those with university degrees, indicating the UK could gain a skilled workforce.
- Those with higher incomes are more likely to consider relocating, again indicating that successful businesses would be built up by migrants (and they would be likely to pay more tax).
- Younger people are more likely to consider migrating, which would be good for the UK’s pension issues.
Building on shaky foundations
Having established that the UK is shortly to be filled with unemployed students from Turkey, the Express moves on to other partially related fears.
“Analysis by Vote Leave suggests Turkey’s membership in the EU would result in far higher numbers of criminals coming to the UK. Figures released by the campaign group show the crime rate in Turkey far exceeds that of the UK, with the murder rate at least four times that of Britain.”
There is no attempt at correlation between these two points, or with the survey. It is left to the reader to make the logical connection that unemployed people coming to the UK would mean higher crime.
“There are also nine million privately-owned firearms registered to its citizens.”
There are millions of guns registered across the UK and the current EU, which has no bearing on anything. Anyone travelling from Turkey to the UK would not be able to bring a gun, as they would need to go through UK immigration and customs where it would be seized without the proper paperwork.
“Responding to the poll, the former shadow home secretary said: ‘You cannot blame young people who want a better life in Britain, but it will overwhelm our public services; it will have an extraordinarily depressing effect on wages and as a result it will cause real pain and penalties for the poorest in Britain. It will be the least well off who will be competing for the same jobs; competing for the same public services.'”
This is the only place the Express acknowledges that it is younger people who were more in favour of considering relocating. However, the other statistics contradict David Davis’s concerns. Those thinking of relocating are not going to be competing for the same jobs as the current less-well-off in the UK – instead, they are more likely to be well-educated on higher salaries. The idea of putting pressure on services will be covered elsewhere.
“Turkey, which has an 80 million strong population, is pressing its case to become a full EU member after wrestling the right to visa-free travel for its citizens across the Schengen Zone. Full-membership would give the country’s predominantly Muslim population the right to free-movement across Europe, with un-fettered access to Britain.”
Full membership would – but we’ve covered that elsewhere. This isn’t going to happen, but the Express is wording it as if it will.
A final word
“Last night Malcolm Rifkind, former Foreign Secretary and Chair of the Intelligence and Security Committee and member of Britain Stronger in Europe, said: ‘This is a totally meaningless poll as Turkish membership of the EU is simply not on the cards. In nearly 30 years of negotiation, they have completed just one of the 35 tests they need to fulfil to apply to join. At this rate it would take them over 1,000 years to meet these criteria. And even then, the British Government has a cast-iron veto on any new country joining, including Turkey. If in the very distant future, they did become a member, our Government has the power to introduce rules to restrict immigration from Turkey.'”
The poll is not meaningless. The conclusions the Express reaches are.