On public opinion surveys

Last week on July 28 the results of a public opinion survey were discussed on social media platforms. For the moment, I will leave out the question of social media as a reflection of general public opinion. The survey, conducted by MetroPOLL, a social and strategic research company, is not yet on its website. I called the MetroPOLL agency but there was no answer. Then I emailed them to ask some precise questions but I received no response. This is puzzling because Iand’ll have to base my criticisms on the partial results published by several media outlets. (See, for example, CNNTanduRK: http:www.cnnturk.comfotogaleriturkiyemetropolden-cok-carpici-suriye-arastirmasi?page=1) As a side note, I cannot understand how these media outlets obtained these partial results or why they would publish them without any discussion. Regardless, the survey is about Syria and the perception and self-perception of Turkish public opinion. Letand’s first look at the results as they have been published by some websites. One of the questions is, and”Who would you like to control the north of Syria?and”
Letand’s start with the question itself. Well, what to say? What kind of question is that? What is it measuring? As my friend and colleague, Professor Emre Erdogan, who is an expert on public opinion surveys and teaches at Boiaziandci University, says, and”Garbage in, garbage out.and” It reminds me of the answer French gastronomy expert Jean Pierre Coffe gave when asked if frozen meals were good (Please forgive me, but these are his words), and”If you freeze s–t, you will unfreeze s–t!and” These kinds of perception and self-perception questions are simply useless for two reasons. First of all, and”public opinionand” is an abstract concept and very dynamic. and”Public opinionand” is fluid tomorrow it will change. Second, perception and self-perception questions have no solid basis. There is a story I tell to my students to illustrate this. In a village, a and”polland” with two questions is conducted. The first question is, and”Are you idiot?and” In response, 100 percent of the villagers answer and”no.and” The result of the poll is therefore that there are no idiots in the village. The second question is, and”Are there idiots in your village?and” In response, 100 percent of the villagers answer and”yesand” — 98 percent even give names covering 100 percent of the inhabitants. The result of the poll is therefore that all of the villagers are idiots. My second concern is about the sampling. The websites who have referenced the partial results say the sample included 2,074 people from all around Turkey without specifying that it is a and”representativeand” sample. However, as MetroPOLL is a serious institution, and I do believe it is, but the sample remains problematic. Letand’s take the example of Felicity Party (SP) and Grand Unity Party (BBP) voters. If the sample is representative, it means there were 42 people who were questioned by the surveyors — though we do not know how was it face-to-face, e-mail, by phone? — because the SP and BBP coalition obtained 2.06 percent of the votes during the June 7 election. Given that according to the survey 5.6 percent of the SP and BBP electors prefer that the Islamic State in Iraq and the Levant (ISIL)ontrol northern Syria, it means that two people gave this answer! How can a conclusion be drawn from such a small sample? I think it is unreasonable. My third concern is about the percentages. I am assuming that all of the respondents know precisely who ISIL and the Democratic Union Party (PYD) are. But what about the remaining answers? Were they open questions or closed questions? And what were the other possible answers as none of the results total 100 percent? For example, do 44 percent of the Justice and Development Party (AKP) voters or 70 percent of the SP and BBP voters believe that Turkey must control this area? We donand’t know. Therefore, these results give no information at all. Last, but not least, the survey was conducted — again according to the websites who somehow got access to the results — between July 7 and 12. In other words, one to two weeks before the Suruandc massacre and before the Kurdistan Workers Party (PKK) attacks. But the and”resultsand” have been partially published after this breaking point. In relation to this, please see my previous column published here on July 29. Letand’s go back to the beginning of this article: public opinion is very flexible and dependent on political and media propaganda. Therefore, to ask this question before and after such a traumatic event could have quite different results. Furthermore, all surveys are subject to manipulation and in all surveys leading questions result in a biased answer. In political science to understand and”public opinionand” we use three agendas: the political agenda, media agenda and public agenda. Issues and problems that are discussed by political parties and leaders fall under the political agenda. This includes things like laws, decrees and policies in general terms. The media agenda is of course the front pages — what popular and mainstream media underline at a given moment. The public agenda covers the preoccupations of the citizens. These are at the same time micro-issues (everyday life problems) and macro-issues (nationalism, foreign policy, ideologies, religions, etc.). The hierarchy among these three agendas is unclear. Who influences whom? In democratic regimes, the public agenda is at the top and media is a real force on the political agenda. But in authoritarian regimes, as in Turkey, each time I see news on the front pages, the result of a public opinion poll or simply a topic that starts to be discussed suddenly, I cannot help but see unprofessional and, sometimes, unethical behavior.