Methodologie of the survey

The Maptionnaire tool

As the timeframe of the project did not allow for the development of a tailor-made survey tool, a phase of research and comparative analysis of existing survey tools was carried out in response to a number of constraints:

Pandemic context and health protocols specific to each country surveyed
Adaptation to different digital devices (computer, tablet, smartphone)
Technical survey capabilities (type of questions, multilingual functionality)
Technical capacity in terms of cartography (assisted and georeferenced drawing tool)
Georeferenced and automatic digitisation of mental maps (no manual processing)
Tool compliant with RGDP and European regulations
Cost of setting up and maintaining the survey tool over a long period
Automatically generated database format

The IMAGEUN Student Survey was constructed and distributed using the digital cartographic questionnaire Maptionnaire. This participatory GIS application was created by a team of Finnish researchers and planning professionals to meet the need for collaborative mapping as part of planning projects (Kahila and Kyttä 2009). The construction of the tool was supported by the Nordic Research Institute Nordregio. Designed for participatory urban projects, the application can be used to answer traditional questionnaire questions, as well as mental maps with a “pencil-like” drawing tool. It allows to choose to frame the map on a global scale, using an OpenStreetMap API (Figure 1). This highly ergonomic mapping tool does have a few limitations, however: it is impossible to modify the interactive background map and its projection system (Pseudo-Mercator).

This survey application also allows the questionnaire to be designed and offered in several languages, and to be accessed on computers, smartphones and tablets.

Figure 1: Example of the application proposed by the distributor

The Questionnaire

Issues and Questions Raised by the Test Phase

A test phase of the questionnaire identified a series of issues that were taken into account in preparing the final version. During this phase, the main challenge was to get the best possible grasp of the complexity involved in expressing spatial imaginaries. Several question orientations were tested to meet technical constraints, constraints of comprehension, of respondents’ attention throughout the questionnaire, together with ethical issues.

Taking Account of Different National Contexts

Carrying out a survey on an international scale means taking into account the different national contexts when designing and administering the questionnaire: languages of translation and languages spoken, socio-economic contexts, level of access to digital devices and the Internet, sensitivity in certain national contexts to issues relating to minorities (Kurds) and gender¹, which necessitated adaptations to the overall structure or to the translations concerned (Bréchon 2002).

Harmonisation of the Level of Analysis

The state level is used as a reference in questions concerning countries of birth, citizenship or mobility, despite its limitations and the desire in this survey to take account of intermediate spatial scales. Nevertheless, this choice has the advantage, in the context of the survey, of allowing comparison and facilitating the translation of spatial terms (e.g. region or city) between the different languages of the questionnaire.

Simplification of the Questionnaire

Particular attention was paid to the structure of the questionnaire to limit the drop-out and non-response rates² and to the expressions used so as not to present it as a simple test of knowledge (feedback from a student work placement by Roman Derlich) or to put respondents in an uncomfortable situation³.

The mental map question required particular attention. It was formulated in several ways, for example by asking them to represent macro-regions by theme (cultural, political, economic space) or to draw a “core” region for the space represented. These questions appeared to be redundant, complex or too geographically expert. The exercise was therefore geared towards tracing the “region of the world” that the student was in at the time of completing the questionnaire. The question is deliberately left open-ended, as it is simpler, but allows for the possibility of representing several macro-regions, depending on personal interpretation and understanding of the complexity of spatial affiliations.

1With regard to the question of gender in the questionnaire, the aim is to understand the influence of gendered socialisation on the representations of regions of the world. It is also to produce an inclusive questionnaire (The GenIUSS Group 2014). The wording chosen is deliberately vague ,“You are:”, implying gender or sex for contexts where the question would create tension or risk during dissemination. The terms “other” and “do not wish to answer” are also not included in the concerned translation.

2For example, by reducing the number of cartographic questions which, according to the test phases, produce a major increase in the non-response rate, using drop-down menus (lists of countries) or sorting questions to avoid fatigue/exclusion (mobility and cultural practices).︎

3Question on (family) income level: attempts at self-positioning on a subjective income scale were tested to get around the impossibility of finding a quantitative indicator that was both simple and comparable in each national survey context. The tests showed that the question is often misunderstood or disturbing for respondents, posing ethical problems and a risk of questionnaire abandonment. These reasons led to the deletion of this question.︎

Adaptation to the Drawing Tool

Due to the technical constraints of the Maptionnaire survey tool, the drawing of a zone automatically leads to the representation of a space by excluding other spaces. This is why the possibility of drawing several spaces has been maintained. This is why the option of drawing several areas has been retained. This makes it possible to grasp the different scales of macro-regions, the links and spatial logics that emerge between these spaces and their imaginary worlds.

Paper or Digital Drawing?

Of all the polygons drawn by the respondents, 80% were considered valid and could be analysed without any corrections. This indicates a good understanding and use of the Maptionnaire drawing tool.

The cartographic results obtained during the test phase also showed that the use of a digital tool did not seem to have any impact on the results, which were very similar to those of the EuroBroadMap project (paper survey) in terms of the methods of drawing and the appropriation of the exercise. For a student audience, the proposed drawing tool and interface did not represent a major difficulty. The practices and biases of the respondents were the same as those observed in the maps drawn on paper, from one extreme to the other, from plotting precisely along the coastline to imprecise or schematic plotting.

This can be explained by the interactive interface of the tool. The base map, which cannot be changed, allows users to find themselves in a familiar tool (since it is the same base map as traditional GPS tools such as OpenStreetMap or Google Maps). The interactivity of the base map also enables respondents to find their way around with precision, thanks to an adaptive generalisation (place names, borders, buildings, road network) according to the scale (zoom), enabling the map to be tailored to their mapping needs.

Framing, Practices and Representations

The final, distributed version of the questionnaire contains seven thematic sections designed to capture students’ spatial practices and representations of macro-regions and to collect data that can be compared with the other thematic areas of the project. Without claiming to go into detail on these themes, each of which could be the subject of a separate survey, the idea is to analyse the construction of students’ representations of the world on the basis of their experiences.

Framing questions (1) about the surveyed student
Mobilities achieved and imagined
Representation of the macro-region(s)
Media practices
Cultural practices
Representations of the EU and Europe
Framing questions (2) on parents or tutors

The Figure 2 below shows the sequence of the sections and their questions:

Figure 2: Structure of the questionnaire

Spatial practices are covered in three parts of the questionnaire:

Achieved mobility (tourism, work, university) - Part 2
Media practices (languages, media, themes) - Part 4
Cultural practices (languages, countries of origin of films or readings) - Part 5

The aim of the “Media” section is twofold. Firstly, it intends to identify and question the socio-spatial representations arising from students’ media practices and present in macro-regional imaginations. The aim is to collect and analyse a range of information, such as centres of interest, the languages in which these media are consulted, and the scales and spaces involved. The second objective is methodological, since this part on media practices is also linked to the “Media” thematic axis of the IMAGEUN project and is based on the nomenclatures of the surveys carried out by the French Ministry of Culture (see The cultural practice survey))(Coulangeon 2011).

Understanding how space is represented is central to this questionnaire. It is addressed in two stages, firstly in the section on mobility, with a focus on the mobility desired by students in different contexts (Part 2). These questions are presented on a non-interactive background map, allowing students to describe their mobility and introducing the next question, which will require the use of the interactive map. Students’ spatial representations of macro-regions are then questioned using a mental map. Respondents are asked to draw and name the region(s) of the world (up to five) in which they are located at the time of completing the questionnaire (Part 3).

Figure 3: “Mental map” question - Draw and name the region to which you belong

As the countries surveyed are diversely linked to the EU and Europe, the aim is to study the way in which students position their region of the world in relation to the European area, in this case using the question proposed in the European Commission’s Eurobarometer) survey (Part 6). The use of this resource enables us to mobilise an already well-tried formulation, and above all to compare the responses of our sample and position it in relation to a large-scale survey conducted over the same period.

Figure 4: “Eurobarometer” question - What is your image of the EU?

Final distributed version of the questionnaire

Figure 5: Screenshots of the English version of the questionnaire, on computer.

The Surveyed Public

The public surveyed was targeted so as to obtain, for each country, feedback from students in several disciplines (Humanities and Arts, Humanities and Social Sciences, Natural Sciences, Law and Business, etc.), from institutions of varying status, adapted to each national or territorial context (public and/or private universities or schools, under contract or not) and located in cities of varying sizes.

Table 1: Number of respondents (after filtering, see Processing and enrichment section, by surveyed country or territory)

Country	Nb of respondants
Ireland (Dublin)	99
Tunisie (Sfax, Sousse, Tunis)	165
French Caribbean communities (Martinique, Guadeloupe)	231
Turkey (Ankara, Antalya, Istanbul)	233
Germany (Erlangen, Frankfurt am Main, Halle)	582
Mainland France (Lyon, Paris, Rouen)	720
Total	2 030

Six sets of disciplines have been targeted, according to the UNESCO classification (UNESCO Institute for Statistics 2013):

Arts and humanities
Health and welfare
Engineering, manufacturing and construction
Business, administration and law
Social sciences, journalism and information
Natural sciences, mathematics and statistics

Figure 6: Distribution of respondents from all countries by UNESCO discipline

Although each of these groups tends to encompass very diverse fields, this classification provides a basis for comparability between countries with different school systems.

In total, the questionnaire was distributed to thirty-three public and private institutions in the five countries:

Université des Antilles - Guadeloupe division (Point-à-Pitre, Saint-Claude and Les Abymes)
Université des Antilles - Martinique division (Schoelcher and Fort-de France)
Friedrich-Alexander-Universität (Erlangen)
Goethe-Universität (Frankfurt am Main)
Martin-Luther-Universität (Halle-Wittenberg)
Kadir Has Üniversitesi (Istanbul)
Süleyman Demirel Üniversitesi (Isparta)
Ecole normale supérieure (Lyons)
Université Lyon 2 Lumière (Lyons)
Université Lyon 3 Jean Moulin (Lyons)
Université Gustave Eiffel (Champs-sur-Marne)
Université Paris 1 Panthéon-Sorbonne (Paris)
Université Paris Cité (Paris)
Université Paris Saclay (Paris)
Ecoles de commerce (Rouen)
Ecoles d’ingénieur (Rouen)
Université de Rouen Normandie (Rouen)
Faculté de médecine (Sfax)
Faculté de Lettres et de Sciences Humaines (Sfax)
Ecole Nationale d’Ingénieurs (Sousse)
Ecole Supérieure Privée d’Ingénieurs (Sousse)
Faculté des Lettres et des Sciences Humaines (Sousse)
Faculté de Médecine de Ibn Jazzar (Sousse)
Faculté de Droit et des Sciences Politiques (Sousse)
Institut Supérieur d’Informatique et des Technologies de Communication (Sousse)
ISET (Sousse)
École nationale d’ingénieurs (Tunis)
Esprit School of Business (Tunis)
Faculté de médecine (Tunis)
Faculté des Lettres, des arts et des humanités (La Manouba)
Faculté des sciences (Tunis)
Institut de Presse et de Science de l’information (La Manouba)
University College Dublin (Dublin)

Distribution Method and Schedule

The questionnaire was sent to students via their university emails, either by a project teacher or among his or her network, or via the institution’s communication services (in the case of Université des Antilles). Development of the survey began in September 2020, with the aim for the first twelve months of working on its design, testing it and analysing the initial results. The final distribution of the survey took place over eight months, between November 2021 and June 2022, with different schedules according to countries (Figure 7). On the one hand, these variations are due to delays in getting the decisions of the ethics committees in France, Turkey and Ireland in particular. On the other hand, certain contexts, such as Tunisia and Ireland, required more contacts and reminders, sometimes because of low enrolment in certain training courses, and more or less complex access to equipment and internet connection.

Geopolitical Context

It is essential to note that a broad dissemination period exposes respondents to major national and international events which may influence their responses. At international level, we might mention the start of the Russian-Ukrainian war in February 2022, or at national level in France, the presidential elections in April 2022. When analysing this data, it is important to take into account the different timescales for distribution in different countries.

Figure 7: Change in the cumulative number of responses to the questionnaire by country or territorial unit

Raw Data Collected

The Maptionnaire tool was used to retrieve the responses to the questionnaire in the form of two Excel files: one file for the responses to the questions with modalities coded in different formats (Boolean, index and character string) and a second file just for the mental maps (geographical data) provided in WKT (Well Known Text) and JSON formats. Prior to their analysis, a significant amount of preparation and enrichment work was carried out to produce the database made available.

With regard to the question of gender in the questionnaire, the aim is to understand the influence of gendered socialisation on the representations of regions of the world. It is also to produce an inclusive questionnaire (The GenIUSS Group 2014). The wording chosen is deliberately vague ,“You are:”, implying gender or sex for contexts where the question would create tension or risk during dissemination. The terms “other” and “do not wish to answer” are also not included in the concerned translation.↩︎
For example, by reducing the number of cartographic questions which, according to the test phases, produce a major increase in the non-response rate, using drop-down menus (lists of countries) or sorting questions to avoid fatigue/exclusion (mobility and cultural practices).︎↩︎
Question on (family) income level: attempts at self-positioning on a subjective income scale were tested to get around the impossibility of finding a quantitative indicator that was both simple and comparable in each national survey context. The tests showed that the question is often misunderstood or disturbing for respondents, posing ethical problems and a risk of questionnaire abandonment. These reasons led to the deletion of this question.︎↩︎