2.1 Discourses

2.1.1 Estimated Frequency of Discourses by group

The heatmap illustrates the extent to which groups address the different discourse clusters across the countries included in the analysis.

Methodology
The content produced by the groups included in the sample was analysed and classified into predefined discourse clusters. For each group, a binary coding scheme based on "yes/no" values was applied:

  • "yes" if the group addressed a given discourse category;
  • "no" if the category was not addressed.
Each group may be associated with multiple discourse clusters. To ensure the comparability of results across countries with different sample sizes, the data were normalised and aggregated at country level. For each country and for each discourse category, the proportion of positive codings ("yes") was calculated as the ratio between: the number of groups coded as "yes" for the specific category, and the total number of groups analysed in the country. These proportions were then expressed as percentage values, which constitute an estimated measure of the relative prevalence of each discourse category among influencers within each national context.
This approach allows for a consistent cross-country comparison while accounting for differences in the number of groups included in each national sample.

2.1 Discourses

2.1.2 Estimated Frequency of Discourses by Groups (All Countries)

The bar chart illustrates the extent to which groups address different clusters of discourses across all countries by aggregating the groups interviewed in each country.

Methodology
The content produced by the influencers included in the sample was analysed and classified into a set of predefined discourse clusters. For each influencer, a binary coding scheme was applied for every category:

  • "Yes" if the influencer addressed the given discourse clusters;
  • "No" if the category was not addressed.
Each influencer could be associated with multiple discourse clusters. In addition to the country-level analysis, an aggregate cross-country measure was calculated in order to capture the overall prevalence of each discourse category across the entire multinational sample. For each discourse category, all positive codings ("yes") recorded in the different countries were summed. This total was then divided by the total number of groups analysed across all countries.
The resulting values were expressed as percentages and represent an estimate of the relative prevalence of each discourse category across the full set of countries considered. This aggregate measure provides a synthetic overview of dominant discourse patterns beyond national contexts, while remaining consistent with the binary coding framework adopted for the analysis.

2.2 Timeline of Mapped Groups

The chart line shows the growth of the FR Groups during the last 45 years.

The line chart illustrates the timeline of activity of the groups mapped in this research.
Data normalization is a crucial step in this analysis, as it ensures comparability across countries with differing and non-uniform numbers of observations.
Specifically, the data were normalized by calculating, for each year, the percentage of active groups relative to the total recorded in 2025.

2.3 Group Typologies

The stacked bar chart illustrates the percentage of different group types within each country.
For each country, the number of groups belonging to each category was divided by the total number of groups in that country. The resulting values were then expressed as percentages, such that the sum of all group types for each country equals 100%.
This method allows for a meaningful comparison of the relative composition of group types across countries, regardless of differences in the absolute number of groups.