Annex tables & downloads

The consolidated annex reference and a complete machine-readable download index

What this page is

This is the supplement’s annex and data hub. It indexes every annex table, embeds the glossary, and provides a complete machine-readable download list — every country-level aggregate (CSV) and every figure (print-ready vector PDF) used in the report and this supplement.

Annex table index

Each annex table has a stable identifier matching the printed report. Most are embedded in context on a thematic page (linked below); all can be downloaded as LaTeX/HTML fragments from the repository.

Table 1: Complete annex-table index.
ID Annex table Shown in context on
A-01 Construct reference card (every construct) Constructs & items
A-02 Short DisInforMeter language matrix Constructs & items
A-03 S4 sample composition by country Measurement quality
A-04 Reliability and AVE by construct and study Measurement quality
A-05 Measurement-invariance ladders Measurement quality
A-06 Full structural path estimates Predictors & anchors
A-07 Full external-anchor correlations (174 rows) Predictors & anchors
A-08 S5 preregistered-hypothesis verdicts Predictors & anchors
A-09 FIMI-variant intercorrelations and reliability Measurement quality
A-10 Glossary below

Glossary (Annex F)

Table 2
Annex F — Glossary
Short technical and policy terms used in the report and this supplement, cross-checked against the methods documentation and the EU policy instruments referenced throughout (EDMO, DSA, EEAS FIMI Defender).
Term Definition
AVE Average Variance Extracted. Proportion of variance in a construct's indicators captured by the latent factor; ≥ .50 conventional threshold.
Chinese-origin FIMI Mean across the three Chinese-origin news items (`news6`–`news8`).
Configural invariance Same factor structure holds across groups; loadings and intercepts free to vary.
DG CONNECT Directorate-General for Communications Networks, Content and Technology — the European Commission service that owns the FIMI / disinformation policy portfolio.
DSA Digital Services Act (Regulation (EU) 2022/2065). Sets risk-assessment obligations under Art. 34/35 and platform-data access under Art. 40 for Very Large Online Platforms (VLOPs).
DisInforMeter The instrument introduced in this report: a 27-item Short scale + 11 first-order belief items + 5 higher-order affective composites measuring population FIMI receptivity.
EDMO European Digital Media Observatory — the EU's network of independent fact-checking and disinformation-research hubs.
FDR False Discovery Rate (Benjamini–Hochberg). Used here to adjust the p-values of the 174 external-anchor correlations.
FIMI Foreign Information Manipulation and Interference. Coordinated activity by a foreign actor (state or non-state) designed to manipulate the information environment of another country.
FIMI Defender EEAS / StratCom's analytical framework for cataloguing FIMI incidents and attribution.
FIMI item One of the 8 news-headline criterion items (`news1`-`news8`) that constitute the FIMI battery. In S1, higher means higher sharing intent; from S2 onward, higher means better detection of foreign manipulation.
First-order belief One of 11 narrower belief sub-dimensions (Exploitation, LGBT+ threat, Immigration, Abandonment by allies / own state / media, Pragmatism toward Russia / China, Superiority of Russia / China, Poisonous ethnocentrism).
Full FIMI Mean across all 8 news items (`news1`-`news8`).
Higher-order composite An affective–cognitive construct (Threat, Betrayal / abandonment, Fear / pragmatism, Foreign-power admiration, General anchor) measured directly in S2-S5 via dedicated items.
Marginal r (predictor rank) Country-specific marginal correlation between a higher-order factor score (metric-invariant) and Full FIMI; bounded in [-1, +1]. Used for the country-snapshot 'top-3 predictors' bar (F-22..F-28).
Metric invariance Loadings are equal across groups (relationships between constructs are comparable).
Russian-origin FIMI Mean across the five Russian-origin news items (`news1`–`news5`).
Scalar invariance Intercepts also equal across groups (absolute means are comparable).
Short FIMI Source-balanced 4-item subset `news2 / news5 / news7 / news8`.
VLOP / VLOSE Very Large Online Platform / Very Large Online Search Engine. Designated under DSA Art. 33 once monthly active EU recipients exceed 45 million.
ΔCFI Change in the Comparative Fit Index between two nested invariance models. ΔCFI ≥ -0.010 supports the more constrained model (Cheung & Rensvold 2002).
ω (McDonald's omega) Composite reliability under the CFA loadings, more appropriate than α when loadings are unequal.
Source: Author judgement; cross-checked against methods annex and EU policy documents (EDMO, DSA, EEAS FIMI Defender). Reproduction code: scripts/policy_report/A-10_glossary.R.

Download index — aggregate data (CSV)

Every country-level aggregate behind the maps, profiles, and country cards. Re-derivation code is on the Data pipeline page.

Download index — figures (print-ready vector PDF)

Every figure is available as a ≥ 300 dpi PNG (embedded throughout the supplement) and as a vector PDF for print reuse. PDFs are listed here.

All artefacts carry stable identifiers (F-, A-, T-) so the report, this supplement, and any downstream reuse refer to the same object. When reusing a figure, carry the three caveat statements from the Home page.