Yes still ahead, but gap narrows in new Turkish referendum poll

Yes still ahead, but gap narrows in new Turkish referendum poll

The results of Qriously’s second poll for this website, as others suggest a narrow advantage for Yes

Over the last week more undecided voters in Turkey have opted to vote No rather than Yes in next weekend’s referendum, the second Qriously poll released to this website suggests.

The national poll of smartphone users produced another strong lead for the Yes campaign, which wants to grant sweeping executive powers to the country’s president, Recep Tayyip Erdoğan.

But the gap between the two camps was narrowing, the poll suggested, with Yes falling below the 60 percent threshold and No pushing above 40 percent.

The number of undecided voters remains a significant factor, however: a quarter of those polled said they had not made up their mind.

The topline figures – with changes from Qriously’s first poll last week – were YES 43.5 (-0.1) NO 31.1 (+3.7) UNDECIDED 25.4 (-3.6).

That produced an expressed vote of YES 58.3 / NO 41.7, a swing of 3.1 percentage points from Yes to No.

Mixed signals

As ever before a Turkish election, the polling picture remains mixed. Qriously’s extremely high Yes result last week – coming in above even the predictions from pro-government pollsters – caused a bit of a stir.

Gezici, a Turkish pollster that is certainly not pro-government, found a lead for the Yes campaign for the first time (although it was Gezici’s first survey since early February).

There have been no other major polls this week, but the message from so many polling company directors appears to be the same: it’s close, very close.

It’s a view echoed by politicians. Mr Erdoğan, for example, declined to discuss specifics during a Kanal 7 interview this week, saying only that Yes was building on its lead.

Ahead, but not in the bag

But the signs are that Yes is creeping ahead. 

The likeliest position right now seems to be that the Yes campaign has a statistically significant lead, but it is in the low-to-mid 50s, meaning it is vulnerable to change in the last week of campaigning.

There will be a flurry of polls coming over the next few days, including from Gezici – which will help establish whether its Yes lead was a statistical anomaly or a true change in voter intentions.

Qriously will be doing a final survey this time next week, two days before the referendum.

Plenty more from this site before then.

Qriously worked with journalist Michael Daventry to develop the survey questions and gave his website James in Turkey exclusive access to the weighted results.

The poll surveyed 2,593 Turkish-language smartphone users based in Turkey who identified themselves as likely voters between Tuesday 4 – Thursday 6 April 2017. Results were weighted on gender, age and region to be nationally representative of the Turkish adult population based on the latest available census data.

The cost of the poll was met entirely by Qriously.

21 comments

  • What about the voters outside of Turkey? There are alot of religious and gray-wolf Turks in Europe. I think +- 70% of those voters will vote for a YES which will have a positive impact of 2% to 3% for the YES camp.

    Reply
    • As I walk the streets and ask people in Istanbul, if they are alone and are voting NO they will tell me. If they are in a group with some who are voting YES they will say they are still undecided. This entire referendum is, as many have pointed out, a sort of farce. The AK party changed the laws that penalize stations who don’t present balanced coverage. So with 90 % of the media effectively in the AK party’s pocket you can guess who s getting the overwhelming media coverage. Also the AK party has painted NO voters as in the same boat with terrorists. My hope is that polls are way off and people are hiding from speaking because they don’t know who they are speaking to or if their cell phones are being tapped. Of course the president constitutionally is still a figure head position so E has been breaking the law for the last few years now. If the vote is YES E will get his way. If the vote is NO then some calamity will be constructed or early elections called or something and E will get his way. Of course he will not say “the people have spoken” I will stand down. In his book if NO prevails, the wrong people have spoken. As others have pointed out, a referendum to change the fundamental way Turkey is run should not be determined in a vote where 50.1 wins against 49.9.

      Reply
      • Perhaps what happened in the USA is happening in Turkey.
        Every poll was wrong.
        People did not want to admit they were voting for Trump, because of the video that had just come out and because of his personal character.
        However, they agreed with his policies, and actually lied to the pollsters.
        Everyone was stunned.

        Reply
  • Aside from the multiple self-selection biases that make this poll unreliable, your explanation states;

    “Results were weighted on gender, age and region to be nationally representative of the Turkish adult population based on the latest available census data.”

    The problem with that is gender,age and region (region is more valid than the other two I’ll gice you that, though not definitive) are not the main variables that determine voter preferences. Leaving education level (perhaps the most relevant indicator according to dozens of detailed polls) out of the equation distorts any normalization efforts substantially. If they did include education level, I would state that since it would make the poll a lot more believable.

    But even then there is the self-selection bias of which voter would be more likely to be gullible enough to click on a random ad saying “are you happy” in an iphone app (which was stated to be the methodnof entry into the poll per your initial Qriouslt article if I’m not mistaken) 🙂

    Reply
    • Also you state that this company correctly predicted Brexit but you leave out the fact that it failed to predict the election of Trump (it predicted Clinton)

      Reply
      • Trump-Clinton was a very hard one, it was like 0,0001 difference. But here it states a difference of 17% which is a HUGE difference.

        This London based poll company also made a correct election prediction on the Italian referendum and Dutch elections. (+Brexit) that is a nice acomplishment.

        Reply
        • Well for this election their results are a total outlier compared to traditional polling methods .. this merits some discusssion .. probably they don’t have the right weight yet for the sampling by smartphone approach. …

          Reply
  • Trump-Clinton wasnt really that close in the states that Trump won outside of Wisconsin and there were multiple other polls that indicated a Trump majority in the Rust Belt. The Mainstream Media refused to focus on the state by state vote and instead focused on the national vote (which is actually irrelevant in the US electoral system due to the Electoral College).

    The Dutch vote predicted by most other mainstream polling company so not an extraordinary accomplishment.

    The Italian referendum and Brexit werent as big suprises either actually for anyone that read beyond the usual BBC, Guardian, Nytimes echo chamber. But I’ll grant that it was misstated by a number of normal polling companies (though again not all).

    The point remains, that the population in Turkey is significantly different structurally, mainly by smartphone penetration rate but also the sharp divergence of peoples voting habits by education, among other issues.

    There was mention of this company polling successfully with this method in India. I would need to see that example before I believe it. Something tells me its highly overstated but I havent seen it so I dont want to insinuate anything.

    Reply
  • I live in the Netherlands and i know alot of MHP voters are voting for YES. AKP + MHP , and a few other small political parties together will lead to a +-60% YES outcome of the referendum on 16 April. I do not believe that there are chances that the NO camp will win the referendum.

    HDP voters will boycot the referendum becuase their possessions have been destroyed. A part of the MHP voters will vote for no and eventuely we will see a maximum 40% NO and around the 60% YES as a end result.

    The secularists never won a election in Turkey. They are doomed to lose.

    Reply
  • A lot of people are voting YES, not because they believe in the amendments but because of the people who are campaigning for “NO”

    These include

    PKK
    FETO
    CHP (din dusmanlari, communists, alevis and these days also pro-PKK)
    Anti-Turkish groups, lobbies and movements across the world.

    Reply
    • Just a point to be observed my friend. FETO is an accronym created by the AK party. There has not been a single crime ever attached to this quasi group. The coup attempt may be proven to have been carried out by Gulenists but not as a terrorist organization but rather as people afraid to lose their influence. Besides the AKParty was friends with these socalled terrorists for a decade or more. Lastly, there are many AK party politicians who were Gulenists right up to the day of the coup. While journalists have been arrested not a single AK party gulenist has been arrested. This is a simple blood feud between Gulen and Erdogan. Nothing more

      Reply
    • hahahaha your website is a joke and is personally biased. dont you ever say that alevis are pro pkk you fuck. you know erodogan is a piece of shit dictator. turkish people want democracy and european and western traits. not that bullshit ottoman islamic executive idealogy that erodogan has brian washed most of the turkey with. lies and more lies

      you should quit writing and being a journalist and find a new day job. becuase you know this is strictly ignorant, biased and in favor of the yes campaign that will truly destroy the futures of Turkey and eliminate democracy not strengthen it. not credible i hope everyone reads this,

      Reply
  • This site is bombarded with AKP trolls who are paid soldiers to promote propaganda perpetrated by the government. Just a warning. Many comments show an extreme bias without rational thinking.

    Reply
  • Pingback: Will Turkey's referendum mark the end of democracy and the birth of 'Erdoğanistan'? | Wikipedia Editors

  • The poll result is utter crap!

    Let’s take the Southeast of Turkey. One stated above already that smart phone penetration in Southeast and east of Turkey is negligible almost non existent. And he who thinks Kurdish voters whose towns are decimated by the governing “yes” front will vote “yes” is dillusional.

    The poll methodology only works on a fully developed society, on big data, not on populations where state propaganda is oppressive, labelling “No” people as terrorists is considered legit and not punishable by law. I tend to believe the big city rates of this poll somewhat reliable but nowhere near close on east and Southeast of Turkey. The poll did not even bother to check the traditional election results on certain provinces. As young poll providers who sit in New York, London and haphazardly figured Brexit and Italian Referandum, you don’t stick your nose to countries where you have virtually no clue about the dynamics of society as in this case, Turkey.

    Reply
  • Pingback: Yes still ahead but watch out for Istanbul - James in Turkey

Leave a Reply

Your email address will not be published. Required fields are marked *