As Turkey’s voters are dragged to ballot box for the second time, how do opinion polls say they will vote?
In a three-and-a-half weeks, people across Turkey will be going to the polls for the first repeat general election in the country’s history.
The early election means an early return for a popular James in Turkey favourite: the rolling average of opinion polls.
Those who were with us in the run-up to June will remember the drill. We take the results of the five most recent voting intention surveys, weigh it down with the previous five results to dampen any extremities, and churn out the result on a regular basis. It’s that simple.
But, I hear you ask, was the self-lauded polling average worth its salt for the June election? Well, have a look for yourself:
|Final JiT call||41.1||27.3||16.2||11.2|
The answer is: so-so. The AK Party and MHP calls were just about spot-on, but our average overestimated CHP support (in fact, the rise was negligible) and underestimated HDP votes (which far exceeded anyone’s expectations).
The JiT average was far more useful in identifying voting trends. Have a look at the line chart on this page and you’ll see that polls in the first half of 2015 showed a significant drop in AK Party support, coupled with equally important increases for the MHP and HDP. That turned out to be the story of the June election.
We cranked out the polling average machine – that is to say, a collection of poorly-maintained Excel spreadsheets – over the last couple of weeks and fed it data to see how the polls have shown Turkish voting intention shift over the summer. Each of main Turkish pollsters have churned out at least one opinion poll since the June election, so the JiT average is operating pretty much like it did last time around.
The overwhelming result is that things have barely shifted at all.
So what are the trends this time?
Some aren’t moving
Turkey’s electoral system is such that the AK Party – the party that wanted an early election the most, and has changed its candidates and manifesto more than any other since June – probably needs to take at least 44 percent of the national vote to win back its parliamentary majority. Our polling average has not yet placed it anywhere near that.
In fact, of the 19 surveys published since the June election, just two have put the AK Party above the 44% mark – and both were from a sympathetic pollster.
The same fate appears to have befallen the nationalist MHP, looking becalmed in the mid-teens, and the pro-Kurdish HDP, although the latter is consistently above the 10% election threshold.
Others move too much
By contrast, the CHP appears to be coasting in fairly stable fashion around the 27 percent mark, such so that if the general election were another poll it would be dismissed as an outlier result.
The final JiT polling average before the June election also showed a higher result than the CHP eventually got. Is it possible that the polls give too much weighting to urban CHP types over rural HDP supporters? Only the polling companies themselves can tell us – but Turkish pollsters are notoriously opaque in their methods.
The overall trend so far is pretty clear, though: all the parties are hovering around where they were at the last election. The upcoming campaign might change that, but if it doesn’t, what was the point of a re-run?
Last modified: Thursday 2 June 2016