7a. Country aggregates and geometries

Overview

This page builds the data ingredients that every choropleth, ranked bar chart, radar, country snapshot, and annex table in the policy report depends on. It writes one CSV per analysis to outputs/policy_report/tables/ and one RDS to outputs/policy_report/geometries/. No figures are drawn here — that happens in pages 7c – 7f.

Section	Function	Output
§1	compute_country_means_higherorder()	s4_country_means_higherorder.csv
§2	compute_country_means_fimi()	s4_country_means_fimi.csv
§3	compute_country_means_long()	s4_country_means_byconstruct_long.csv
§4	compute_country_fimi_difference()	s4_country_fimi_difference.csv
§5	compute_country_item_means()	s4_country_item_means.csv
§6	compute_s23_first_order_means() + compute_s23_higher_order_means()	s23_first_order_means.csv + s23_higher_order_means.csv
§7	compute_eu_pooled_benchmarks()	s4_eu_pooled_benchmarks.csv
§8	compute_sample_composition()	s4_sample_composition_by_country.csv
§9	build_country_geometries()	policy_report_country_geometries.rds

Design source: docs/reports/2026-05-21_policy-report-content/03_new-analyses-roadmap.md §§1, 2, 5, 6, 9.

1. Higher-order country means (S4)

Per-country means, SDs, and bootstrapped 95 % CIs for the five higher-order DisInforMeter composites (threat, abandonment, fear, foreign-power admiration, general) plus an overall DisInforMeter composite (mean of the five z-scored, rescaled to 1–7).

Table 1: Preview — per-country higher-order means (first 10 rows of s4_country_means_higherorder.csv).

country_iso3	country_name	cluster	construct_id	construct_label	n	mean	sd	se	ci_lower	ci_upper	scale_min	scale_max
SRB	Serbia	Southern	aband	Betrayal / abandonment	606	5.396	1.359	0.056	5.288	5.498	1	7
TUR	Turkey	Southern	aband	Betrayal / abandonment	599	5.097	1.559	0.063	4.969	5.218	1	7
BGR	Bulgaria	Central-East	aband	Betrayal / abandonment	592	4.960	1.444	0.061	4.840	5.080	1	7
ITA	Italy	Southern	aband	Betrayal / abandonment	606	4.545	1.342	0.054	4.437	4.651	1	7
HUN	Hungary	Central-East	aband	Betrayal / abandonment	599	4.455	1.319	0.055	4.346	4.564	1	7
POL	Poland	Central-East	aband	Betrayal / abandonment	600	4.408	1.494	0.061	4.286	4.529	1	7
FRA	France	Western	aband	Betrayal / abandonment	596	4.359	1.337	0.055	4.253	4.471	1	7
BEL	Belgium	Western	aband	Betrayal / abandonment	595	4.253	1.412	0.057	4.134	4.358	1	7
AUT	Austria	Western	aband	Betrayal / abandonment	599	4.244	1.570	0.064	4.122	4.367	1	7
DEU	Germany	Western	aband	Betrayal / abandonment	598	4.191	1.485	0.059	4.080	4.305	1	7

2. FIMI country means — all four operationalisations

Mean rated plausibility per country for Full (8 items), Russian-origin (5 items), Chinese-origin (3 items), and Short (4 items) FIMI.

Table 2: Preview — FIMI country means (first 10 rows of s4_country_means_fimi.csv).

country_iso3	country_name	cluster	fimi_version	n	mean	sd	se	ci_lower	ci_upper
AUT	Austria	Western	full	599	4.765	0.879	0.037	4.689	4.835
BEL	Belgium	Western	full	595	4.824	0.909	0.037	4.751	4.892
BGR	Bulgaria	Central-East	full	592	4.749	1.011	0.043	4.659	4.830
DEU	Germany	Western	full	598	4.616	0.954	0.039	4.537	4.692
EST	Estonia	Baltic	full	604	5.035	0.982	0.041	4.956	5.116
FRA	France	Western	full	596	4.711	0.949	0.038	4.639	4.789
HUN	Hungary	Central-East	full	599	4.784	0.998	0.041	4.704	4.865
ITA	Italy	Southern	full	606	4.784	0.942	0.038	4.709	4.860
LTU	Lithuania	Baltic	full	593	5.254	1.020	0.043	5.170	5.338
LVA	Latvia	Baltic	full	596	5.141	0.890	0.035	5.072	5.211

3. Canonical long-form country × construct table

Single source consumed by F-07, F-08, F-09, F-16, F-17, and the country snapshots (F-22..F-29). Long format makes it trivial to facet by construct or by country in ggplot.

Table 3: Preview — canonical long-form country × construct table (first 10 rows of s4_country_means_byconstruct_long.csv).

country_iso3	country_name	cluster	construct_id	construct_label	construct_family	n	mean	sd	se	ci_lower	ci_upper	scale_min	scale_max
LTU	Lithuania	Baltic	fimi_chinese_origin	FIMI: Chinese-origin (news6-8)	fimi	593	4.845	1.200	0.049	4.743	4.937	1	7
POL	Poland	Central-East	fimi_chinese_origin	FIMI: Chinese-origin (news6-8)	fimi	600	4.699	1.195	0.048	4.606	4.793	1	7
LVA	Latvia	Baltic	fimi_chinese_origin	FIMI: Chinese-origin (news6-8)	fimi	596	4.681	1.095	0.043	4.598	4.764	1	7
EST	Estonia	Baltic	fimi_chinese_origin	FIMI: Chinese-origin (news6-8)	fimi	604	4.598	1.106	0.046	4.509	4.688	1	7
SRB	Serbia	Southern	fimi_chinese_origin	FIMI: Chinese-origin (news6-8)	fimi	606	4.530	1.158	0.047	4.439	4.617	1	7
BEL	Belgium	Western	fimi_chinese_origin	FIMI: Chinese-origin (news6-8)	fimi	595	4.477	1.130	0.047	4.383	4.568	1	7
BGR	Bulgaria	Central-East	fimi_chinese_origin	FIMI: Chinese-origin (news6-8)	fimi	592	4.446	1.213	0.050	4.346	4.544	1	7
AUT	Austria	Western	fimi_chinese_origin	FIMI: Chinese-origin (news6-8)	fimi	599	4.445	1.143	0.047	4.355	4.541	1	7
TUR	Turkey	Southern	fimi_chinese_origin	FIMI: Chinese-origin (news6-8)	fimi	599	4.378	1.325	0.056	4.265	4.481	1	7
HUN	Hungary	Central-East	fimi_chinese_origin	FIMI: Chinese-origin (news6-8)	fimi	599	4.364	1.156	0.047	4.266	4.458	1	7

4. Paired Russian − Chinese FIMI difference

Per-respondent paired difference (Russian-origin mean − Chinese-origin mean) aggregated by country, with bootstrapped 95 % CI for the difference. Country snapshots and F-12 use this as their primary contrast.

Table 4: Preview — paired Russian − Chinese FIMI difference by country (first 10 rows of s4_country_fimi_difference.csv).

country_iso3	country_name	cluster	n	mean_R	mean_CH	difference	se	ci_lower	ci_upper	ci_excludes_zero
FRA	France	Western	596	5.058	4.134	0.924	0.059	0.809	1.049	TRUE
ITA	Italy	Southern	606	5.093	4.268	0.824	0.050	0.732	0.922	TRUE
LVA	Latvia	Baltic	596	5.417	4.681	0.736	0.048	0.641	0.830	TRUE
EST	Estonia	Baltic	604	5.298	4.598	0.700	0.048	0.605	0.795	TRUE
HUN	Hungary	Central-East	599	5.037	4.364	0.673	0.052	0.561	0.770	TRUE
LTU	Lithuania	Baltic	593	5.500	4.845	0.655	0.052	0.558	0.756	TRUE
SRB	Serbia	Southern	606	5.093	4.530	0.563	0.050	0.466	0.667	TRUE
BEL	Belgium	Western	595	5.032	4.477	0.555	0.049	0.460	0.655	TRUE
AUT	Austria	Western	599	4.957	4.445	0.512	0.048	0.420	0.612	TRUE
BGR	Bulgaria	Central-East	592	4.931	4.446	0.485	0.053	0.379	0.583	TRUE

5. Country × FIMI item heatmap data

Per-country mean plausibility for each of the eight news items in S4, joined with item metadata (origin, truth value, short label) from outputs/item_mapping.csv. Drives F-13.

Table 5: Preview — per-country FIMI-item plausibility means (first 10 rows of s4_country_item_means.csv).

country_iso3	country_name	cluster	item_code	item_short_label	origin	truth_value	short_form	n	mean	sd	se
AUT	Austria	Western	news1	Zelensky / King Charles mansion	Russian-origin	FALSE	FALSE	599	5.573	1.665	0.068
AUT	Austria	Western	news2	France / Russia in Ukraine	Russian-origin	FALSE	TRUE	599	5.259	1.717	0.070
AUT	Austria	Western	news3	Milan-Cortina 'openly LGBT' Olympics	Russian-origin	FALSE	FALSE	599	3.898	1.849	0.076
AUT	Austria	Western	news4	Ukrainian army tolerance mentor	Russian-origin	FALSE	FALSE	599	4.740	1.621	0.066
AUT	Austria	Western	news5	Kyiv LGBT brigades	Russian-origin	FALSE	TRUE	599	5.314	1.683	0.069
AUT	Austria	Western	news6	Li-Meng Yan COVID rumour	Chinese-origin	FALSE	FALSE	599	4.364	1.835	0.075
AUT	Austria	Western	news7	CGTN: China-Vietnam friendship	Chinese-origin	TRUE	TRUE	599	4.406	1.602	0.065
AUT	Austria	Western	news8	Banned-leader Li Hongzhi 'plague-rule'	Chinese-origin	FALSE	TRUE	599	4.564	1.776	0.073
BEL	Belgium	Western	news1	Zelensky / King Charles mansion	Russian-origin	FALSE	FALSE	595	5.613	1.619	0.066
BEL	Belgium	Western	news2	France / Russia in Ukraine	Russian-origin	FALSE	TRUE	595	5.007	1.733	0.071

6. S2 (Lithuania) and S3 (Germany) means

First-order and higher-order means with 95 % CIs for the Lithuanian (S2) and German (S3) samples. These are what F-14 (radar) and F-15 (paired bars) draw from, and they also feed the LT and DE entries in the country snapshots.

Table 6: Preview — S2 / S3 higher-order means (full s23_higher_order_means.csv).

study	country_iso3	country_name	construct_id	construct_label	n	mean	sd	se	ci_lower	ci_upper	scale_min	scale_max
S2	LTU	Lithuania	threat	Ideological threat	582	4.054	1.629	0.070	3.916	4.189	1	7
S2	LTU	Lithuania	aband	Betrayal / abandonment	582	3.798	1.667	0.070	3.663	3.941	1	7
S2	LTU	Lithuania	prag	Fear / pragmatic accommodation	582	4.329	1.922	0.083	4.162	4.498	1	7
S2	LTU	Lithuania	supf	Foreign-power admiration (Russia)	582	1.920	1.527	0.062	1.801	2.039	1	7
S3	DEU	Germany	threat	Ideological threat	782	3.349	1.670	0.059	3.236	3.461	1	7
S3	DEU	Germany	aband	Betrayal / abandonment	782	3.721	1.727	0.061	3.599	3.838	1	7
S3	DEU	Germany	prag	Fear / pragmatic accommodation	782	4.043	2.017	0.070	3.905	4.183	1	7
S3	DEU	Germany	supf	Foreign-power admiration (Russia)	782	2.364	1.627	0.060	2.245	2.486	1	7

Table 7: Preview — S2 / S3 first-order means (first 10 rows of s23_first_order_means.csv).

study	country_iso3	country_name	construct_id	construct_label	n	mean	sd	se	ci_lower	ci_upper	scale_min	scale_max
S2	LTU	Lithuania	expl	Exploitation by foreign powers	582	3.851	1.697	0.073	3.719	3.995	1	7
S2	LTU	Lithuania	gay	Opposition to LGBT+ values	582	4.347	2.135	0.090	4.176	4.521	1	7
S2	LTU	Lithuania	migr	Opposition to migration	582	5.113	1.610	0.067	4.980	5.242	1	7
S2	LTU	Lithuania	abanf	Abandoned by foreign allies	582	3.835	1.669	0.067	3.708	3.976	1	7
S2	LTU	Lithuania	abanh	Abandoned by own state	582	3.888	1.843	0.078	3.741	4.047	1	7
S2	LTU	Lithuania	abanm	Abandoned by media	582	3.566	1.940	0.080	3.402	3.720	1	7
S2	LTU	Lithuania	pragr	Pragmatism toward Russia	582	3.020	1.855	0.078	2.859	3.169	1	7
S2	LTU	Lithuania	pragch	Pragmatism toward China	582	3.738	1.614	0.068	3.602	3.874	1	7
S2	LTU	Lithuania	supr	Superiority of Russia	582	2.342	1.562	0.064	2.217	2.468	1	7
S2	LTU	Lithuania	supch	Superiority of China	582	3.303	1.413	0.059	3.185	3.420	1	7

7. EU pooled benchmarks

One row per construct: pooled-EU mean, CI, and the min/max country mean. Country snapshots overlay these as the EU range and EU mean reference lines.

Table 8: EU pooled benchmark per construct (full s4_eu_pooled_benchmarks.csv).

construct_id	construct_label	construct_family	n_total	n_countries	mean	se	ci_lower	ci_upper	sd	min_country_mean	max_country_mean	min_country_iso3	max_country_iso3
aband	Betrayal / abandonment	higher_order	7783	13	4.430	0.128	4.186	4.689	0.488	3.537	5.396	EST	SRB
admir	Foreign-power admiration (composite)	higher_order	7783	13	2.858	0.173	2.544	3.203	0.652	1.875	4.032	EST	BGR
fear	Fear / pragmatic accommodation	higher_order	7783	13	4.671	0.117	4.433	4.886	0.454	3.535	5.215	EST	BGR
fimi_chinese_origin	FIMI: Chinese-origin (news6-8)	fimi	7783	13	4.476	0.051	4.381	4.577	0.194	4.134	4.845	FRA	LTU
fimi_full	FIMI: Full (news1-8)	fimi	7783	13	4.857	0.050	4.765	4.960	0.190	4.616	5.254	DEU	LTU
fimi_russian_origin	FIMI: Russian-origin (news1-5)	fimi	7783	13	5.085	0.056	4.986	5.201	0.211	4.792	5.500	DEU	LTU
fimi_short	FIMI: Short (news2/5/7/8)	fimi	7783	13	4.863	0.055	4.759	4.974	0.208	4.495	5.250	TUR	LTU
gen	General disinformation acceptance	higher_order	7783	13	4.049	0.099	3.858	4.247	0.378	3.287	4.730	EST	SRB
overall	Overall DisInforMeter composite	higher_order	7783	13	4.074	0.117	3.855	4.310	0.445	3.221	4.847	EST	SRB
supf	Foreign-power admiration (Russia)	higher_order	7783	13	2.683	0.200	2.318	3.092	0.756	1.593	4.004	EST	BGR
supfch	Foreign-power admiration (China)	higher_order	7783	13	3.033	0.147	2.764	3.326	0.558	2.157	4.060	EST	BGR
threat	Ideological threat	higher_order	7783	13	4.400	0.113	4.193	4.633	0.429	3.834	5.209	EST	SRB

8. S4 sample composition by country

Per-country N, age (mean, SD), gender breakdown, education level, urban share, daily-internet share. Feeds T-02 (body of the report) and Annex C.

Table 9: S4 sample composition per country (full s4_sample_composition_by_country.csv).

country_iso3	country_name	cluster	n	age_mean	age_sd	age_min	age_max	pct_female	pct_male	pct_other_gender	pct_pref_not_say	pct_tertiary_education	pct_lower_education	pct_vocational_education	pct_media_daily
AUT	Austria	Western	599	45.62	16.51	23.5	67	51.75	47.91	0.33	0.00	28.21	28.88	42.90	7.68
BEL	Belgium	Western	595	46.42	15.71	23.5	67	48.91	51.09	0.00	0.00	45.04	41.18	13.78	10.42
BGR	Bulgaria	Central-East	592	43.76	15.10	23.5	67	54.22	45.78	0.00	0.00	44.76	32.94	22.30	8.61
DEU	Germany	Western	598	45.82	15.75	23.5	67	50.00	50.00	0.00	0.00	36.62	17.06	46.32	8.86
EST	Estonia	Baltic	604	47.11	15.87	23.5	67	54.30	45.36	0.33	0.00	37.09	43.21	19.70	4.80
FRA	France	Western	596	45.97	15.43	23.5	67	48.66	51.01	0.34	0.00	40.77	32.72	26.51	6.04
HUN	Hungary	Central-East	599	44.03	15.36	23.5	67	52.25	47.75	0.00	0.00	23.87	50.58	25.54	5.84
ITA	Italy	Southern	606	48.58	15.01	23.5	67	46.37	53.63	0.00	0.00	35.97	50.33	13.70	11.22
LTU	Lithuania	Baltic	593	48.07	15.93	23.5	67	55.14	44.18	0.17	0.51	50.08	22.43	27.49	7.42
LVA	Latvia	Baltic	596	47.90	15.99	23.5	67	55.70	43.96	0.17	0.17	38.26	27.01	34.73	3.19
POL	Poland	Central-East	600	42.62	15.27	23.5	67	53.50	46.33	0.00	0.17	31.50	57.83	10.67	11.17
SRB	Serbia	Southern	606	43.47	11.65	23.5	67	67.66	32.34	0.00	0.00	52.64	33.50	13.86	7.59
TUR	Turkey	Southern	599	41.79	14.59	23.5	67	47.25	52.75	0.00	0.00	56.09	23.87	20.03	27.38

9. Country geometry layer

Reusable sf object: Natural Earth admin-0 at 1:50 m, filtered to the 13 S4 countries plus a grey neighbourhood, reprojected to EPSG:3035 (ETRS89 / Lambert Azimuthal Equal-Area Europe — the convention every EU report uses). Cached as RDS so map scripts don’t re-download / re-project.

Note

Edge cases to verify on first build: Kosovo (XKX) and Serbia (SRB) — confirm Natural Earth’s handling. Georgia (GEO) and Turkey (TUR) must be present in the filtered set.

The cached RDS is consumed by every choropleth in 7c.