Research Methodology: The Survey 

GraphThis note briefly summarises the background of the survey that was carried out to collect data on open society attitudes across the six countries in the project (Hungary, Poland, Greece, Italy, France and Germany). The questionnaire was developed to address two main areas of concerns: first, to evaluate the composition of open and closed society attitudes conceptually and second, to examine how individuals respond when asked to decide whether a particular open society attribute was more, less or equally important than an attribute associated with closed societies.


Open society constructions

In the first part of the survey respondents were presented with 14 items (in random order) and asked to indicate how important they thought the respective item was for a good society. Half of the items (7) were attributes defined as open society characteristics and half (7) were attributes more closely associated with closed societies.

The data collected in this first section allows us to apply dimension reduction techniques to examine how the items relate to each other. The items were:

Based on the answer options we computed two scores, one for the rating of open and one for closed society attributes. To calculate the score, the order of the answer scales was reversed (so higher values indicated higher levels of rating an item as essential). Then all seven respective item scores were added up, resulting in a score between 7 and 28. We then standardised the scores between 0 and 1. So both scores respectively measure how essential respondents rated the open and closed society attribute items respectively. A score of 0 means respondents said “not at all essential” to all seven items, a score of 1 means they said “absolutely essential” to all seven items.Trade-off experimentsIn the second part of the survey respondents were presented with 14 direct comparisons between two items and asked to evaluate their relative importance. For each of the seven open society attributes from the first part of the survey, respondents were presented with two alternative items and asked to make a decision in each of those comparisons. The order in which the comparisons were presented was randomised. The full set of comparisons was as follows:

Table 2


Other questions

In addition to the instruments presented above, we also included a number of socio-economic questions and correlate questions about attitudinal domains, such as attitudes towards immigration, civil society and their political affiliations across all countries. Additionally, a few country-specific questions were added in each country to support the analyses within each country-specific context.


Survey development

The survey was developed by dpart’s core team for the Situation Room project and in close cooperation with all five country partners. After initial scoping about issues and instruments, a draft questionnaire was developed that was then discussed in a workshop with representatives of all country partners, d|part and OSEPI. Based on that workshop a second draft questionnaire was developed that went through further iterations with feedback from all partners. The final survey (in an English master version) was then translated by professional translators into the six country languages.. Country partners then checked the translation, in addition to checks carried out by the core team. Feedback was then given to the translators to further improve the translations. Country partners formulated draft questions specific to their own contexts that were reviewed and edited by the core team before going through a final round of checks together with the master questionnaire.


Fieldwork and data

The fieldwork was carried out by Lightspeed Germany in close cooperation with the core team. The survey was administered through an online panel in all countries. Programming of the survey was pre-tested by several people in each country to check for user experience, correct routing and the implementation of translations. Quotas for age, gender, geography, education and income and several cross quotas were applied, to achieve good representation in the samples. Quotas were only relaxed at later stages in the fieldwork and in case they could not be filled adequately otherwise. Before commencing with the fieldwork, a soft launch pilot was carried out with 50 respondents in each country to test the survey instruments and check initial distributions and participation. Subsequently the full launch took place with over 1000 respondents recruited in each country. The survey was carried out between 12 February 2018 and 5 March 2018. Where achieved sample distributions deviated from actual population distributions, weights were calculated and applied to account for those deviations.