“The Walls Have Ears”: Finding Out what the Iranian Public Really Thinks

What is the legacy of the Green Movement? IranWire asked polling group Information and Public Opinion Solutions LLC (iPOS) to talk to ordinary Iranians from across the country about their views on the 2009 disputed election and its aftermath. Pollsters telephoned Iranians of different ages, backgrounds and incomes, and the results of the survey generated widespread debate across social media networks, in both Persian and English.

We asked one of the survey’s creators, sociologist Hossein Ghazian, about the project, its results and public opinion in Iran.

 

There are obvious security risks involved when you telephone Iranians from abroad and discuss potentially sensitive issues. Did you really expect people to give an honest answer?

Absolutely not! Of course, given that we do not necessarily expect an honest answer, you would be right to ask just why we are wasting our time and your money and going through so much trouble.

To be honest, when I follow up on reactions to this survey and opinion polls like it I often come across this question. When raising this issue, some people are sarcastic; others are reproachful. Of course, those who question the validity of the poll usually have preconceived ideas about it — ideas that do not match the actual results. If they did not have these preconceptions, they would assume the results were watertight and might see the survey as a kind of sacred tool.

The problem starts when people try to imagine how we conduct our surveys, including this one about the Green Movement. They imagine  we call people up and say: “Hello! Do you think fraud was committed in the 2009 election?” Then they imagine that the clueless respondent, confronted with an inappropriate question from an unknown caller with a strange phone number, is surprised out of his skin and thinks: “Who are these idiots who don’t know that the walls have ears?”, clears his throat and gives the answer that he thinks the security “brothers” want to hear.

And then, it is assumed, we continue by adding up the number of these frightened people and announcing that 59 percent of Iranians say that the 2009 election was not fraudulent.

In addition to this, there are some who even believe we want to prove that fraud was not involved in determining the election — and who say they have a pile of evidence to prove that it was fraudulent. In other words, they do not distinguish between reality and the people’s perception of reality. Neither this survey, nor any other survey, can say whether fraud was involved or not. The survey can only say how people think about electoral fraud.

But perhaps these people should give a little thought to the fact that if they, like other ordinary people, can easily figure out that nobody gives an honest answer on the phone to such prodding questions then, after two decades of conducting surveys, we can figure it out as well — especially since we have shown several times at critical junctures that our projections have been correct. Besides, do they think that the accumulated experience of social polling science has found no solutions for such issues, or that we are ignorant of them?

 

Can you tell us what methods you used to determine the accurate  answers and claim your findings are correct? Did you ask direct questions about the Green Movement or indirect ones?

One of the tried and true techniques for establishing trust between the questioner and respondent is to start the interview and conduct it in a way so that that respondent enters the conversation “naturally” and is gradually steered  toward sensitive questions — in a manner that most people would not notice, or because the level of trust has risen so high that they give you a more honest answer.

To this end, we observe many delicate points when both structuring our questionnaires and posing the questions. There are professional secrets to what we do so I cannot talk about every detail, like many other organizations that do not disclose such information.

But let me point to something that I can discuss. In this poll we presented 11 main questions about the Green Movement. But to get the answers for these 11 questions we asked each respondent 22 questions. The extra questions ranged from topics such as familiarity with Ali Karimi [the footballer], Fatima Gul [a Turkish TV series] and Morteza Pashaei [pop singer and musician], to other questions designed to relax the respondent. So the 11 extra question are either for warming up the conversation and building up trust or for leading to more sensitive questions.

Besides, in all our questionnaires, we take into consideration the political psychology of language, meaning that we — at least as much as anybody else — know what questions drive people away and make them guarded. When wording the questions, we stay away from formal language and terms and use common everyday language.

More importantly, our questioners are trained to present the questionnaire in a way that suggests they are asking the questions out of personal curiosity, in the same way that in a normal conversation each question can lead to another question. In the process, they evaluate the trustworthiness and the honesty of the respondent. In cases where the respondents have a low level of trust and honesty and their answers are biased, the answers are taken out of the sample.

Nevertheless, you would be mistaken to believe that we think, even with all this preparation,  that everyone gives us their true views with trust and honesty. To be more certain, we evaluate the level of honesty with internal and external control variables and warn the consumers of the polls about why and where they should avoid hasty conclusions — for example, we included observations about the analysis of certain results at the end of this survey.

 

In this survey, what percentage of respondents were educated to college level? What percentage did not hold a high school diploma? What percentage lived in villages or underdeveloped towns? Did the samples match the diversity of Iranian society?

Note that, if necessary, we only give weight to two personal variables —age and gender — and to where they live — village or town— based on the 2011 census. In our sample, around nine percent had an educational qualification higher than a high school diploma but lower than a university degree, and 26 percent held a Bachelor’s degree or higher. In other words, around 65 percent of those polled were educated to a standard that was less than the equivalent of a high school diploma.

In our sampling, 74 percent lived in cities and 26 percent in villages. The poll included responses from two people from tribal backgrounds. In addition, four percent of our respondents were from provinces with low human development indices [which measure well-being across a number of different variables such as education, income, and health] — Sistan and Baluchistan and Kurdistan — and 31 percent from provinces with high indices — Tehran, Isfahan, Yazd and Semnan. This corresponds with the distribution of the parameters in Iranian society. So we can say that our sampling was a small-model representative of the Iranian population based on attributes such as age, gender, education, living in towns or villages, and the level of human development, which is itself  composed of three factors: Health, education and income.

 

What criteria did you have for choosing respondents?

Theoretically, any Iranian who lives in Iran, is over 18 years of age and has a landline or a cellphone (from any service provider) has an equal chance to be included in our sampling.

 

Considering the sensitive nature of this survey and the fact that the questions cross the government’s red lines or taboo subjects, can answers like “I don’t know” or “I have no opinion” and truncated interviews be treated as meaningful?

I agree up to a point that this is problematic. But the question is about the degree of this effect and its direction. I will clarify.

First of all, I have to point out that the level of truncated interviews in this survey is not very different from others conducted by iPOS. For example, in this survey, when it came to people dropping out of interviews, we had only two percent more dropouts than when we conducted surveys about the approval ratings of President Rouhani or about nuclear negotiations — even though the questions in this survey were more sensitive and the questionnaires were longer. One of the reasons that a respondent might cut short an interview is that he or she becomes tired of a lengthy interview or other chores intervene — it is not necessarily the sensitive or political nature of the question. At whatever level the initial trust of the respondent was, 95 percent of respondents completed the interview.

Secondly, by reviewing the responses of those who did not answer sensitive questions, we found that there were no serious differences between them and those who did answer them. We classified the 11 main questions into five categories of sensitivity. From a political point of view and from the standpoint of how safe the respondent felt about answering them, the most sensitive questions were about electoral fraud and what should be done about Messrs. Moussavi and Karroubi. But when we compare those who did and those who did not answer these two questions, we find out that the two groups are not much different when it comes to personal and social attributes such as age, gender and education. Besides, by taking into account the control questions, we notice that even from a political perspective, the answers did not differ greatly, meaning that respondents’ political tendencies did not necessarily make them scared of answering. Even if there were serious differences, the size of differences are not enough to change the general outcome. But of course, the consumer must be prudent when it comes to using or applying the results.

 

Related Article:

What do you think of Rouhani?

 

To read more stories like this, sign up to our weekly email.