The “Referandum”

Tomorrow Easter Sunday April 16 2017 is the date of the Turkish Constitutional Referendum. Turkish voters are called to vote for (evet) or against (hayir) 18 amendments to the Turkish constitution. The amendments would give Recep Tayyip Erdoğan unprecedented power in effect turning him into a democratically elected dictator. The outcome of the referendum is unknown.

Diaspora Turks were polled once about two weeks ago: 62% for yes! In Turkey the polls of the last week show an average of 52.9% for yes among those who are considering a yes/no vote only. Among those who are considering an undecided vote  on the average across polls we have 48%, 43% and 9% for yes, no and undecided respectively.

All polls show a slight lead for the yes vote but strategic answering is highly likely in a country were people are persecuted and imprisoned for no good reason after the coup of July 15 2016.

Erdogan played every political trick in the book, including the dirty kind, in order to cause his people to rally around him. Diplomatic crisis with Germany and the Netherlands, imprisonments, threats, rejection of international treaties, rhetoric of Turkish expansionism and increased violations of Greek airspace by Turkish air force after the coup are only some of the pages from his playbook.

The Turkish referendum is important beyond Turkey and very important for Europe because it could create a radicalized neighbor for a Europe who is struggling  to manage anything beyond a Sunday picnic let alone a refuge crisis, terror or military threats or even a new coup in Turkey should Recep Tayyip Erdoğan win.

What can we learn from Google searches about the referendum? The Turkish word for it is “referandum” which allows us to re-discover where Turkish diaspora has the largest footprints (Germany, UK, US).

There is interesting variation within Turkey, across its provinces, with more interest in the deep east.

The top such searches include the word “anket” i.e. “survey” which means that most searches are about information on the latest projections. Searches for yes or no do not help to find a balance of power of the two camps as the no is overwhelming and this seems at odds with polls.


The geographic distributions does give us some clues though. The No searches are more prominent in Ankara and minority rich provinces. The reason searches are that lopsided might be due to fear of revealing preferences due to a climate of terror by the current regime.



One thing we learn here is that while in democratic regimes people feel relatively safe in the privacy of their internet sessions this might not be the case in less democratic climates.

This entry was posted in Big Data, nowcasting, Social Media and tagged , , , , . Bookmark the permalink.