What Does Democracy Mean to Tunisians?

Why this study

When people say they support democracy, it is usually assumed that they mean the same thing by it. The research does not support that assumption. Across the world the word carries different weights: elections, civil liberties, redistribution, order, religion, or a strong leader unbothered by parliament. Tunisia has its own history of that argument, from the constitution of 2014 to the years after 2021, and no public instrument has yet asked Tunisians in detail what they mean when they use the word.

This study asks the question directly, in Modern Standard Arabic, French and English, and it does four things the existing surveys do not. It asks respondents to rate each component of democracy. It then forces a choice when everything cannot be equally important. It asks how present each component is in Tunisia today, which produces a gap between the ideal and the reality. And it measures how much the answers move when the order of the questions changes. The instrument, the data and the analysis are published whether the findings are comfortable or not.

The point is not to grade anyone's understanding of democracy. The point is to find out whether Tunisians disagree about politics, or whether they are talking past each other with the same word.

How it works

The survey takes about twelve minutes and runs in a browser. It asks for no account, no email address and no name, and it stores no IP address with the answers. Any question can be skipped, and a respondent can stop at any point before submitting. Participation is open to anyone aged 18 or over, inside Tunisia or outside it.

There are four parts: a set of components to rate from 0 to 10, with an explicit option for components considered to be against democracy; a forced-choice task that asks which components matter most when they cannot all be chosen; a set of ratings for how present each component is in Tunisia today; and a few questions about government and about the respondent, used to weight the sample and to compare it with national surveys.

At the end the respondent receives a random receipt code. It is the only way to delete a response while the study is open. Once the study closes the dataset is frozen and a response cannot be removed, and the consent screen states that before participation begins.

What happens to the answers

The answers are stored without any identifier and used for one purpose: this study. Nothing is sold, shared with a political party, or handed to an advertising system. There is no analytics script on the survey page. When the study closes, the raw rows are replaced by an anonymized dataset in which location is grouped into regions, age is widened into bands and education is reduced to three levels, so that no published table can point at a person. Free-text answers are coded into categories and the raw text is deleted.

The results are published in full, including negative ones. If the instrument fails to measure what it intended, or the sample is too thin for a planned analysis, that is reported first rather than buried. The data is released under CC BY 4.0, and every number on the results page carries the number of respondents it is based on and the fact that they were self-selected.