1 Purpose and specification

This report separates the service-quality component used in Core 2 into the two item families requested by the coauthor. It estimates weighted OLS associations with the territorial political-solidarity index (municipal, national, and EU willingness to contribute). All models use the Core 2 controls, country fixed effects, and NUTS2-clustered HC1 standard errors.

Each predictor is rescaled in its original direction, completed with its country mean when missing, and standardized. Coefficients therefore represent the change in the 1–5 political-solidarity outcome associated with a one-standard-deviation difference in the predictor. Because predictor and outcome are contemporaneous self-reports, the estimates describe associations rather than causal effects.

2 Family indices

Associations of the two requested family indices with political solidarity.
Index	b	95% CI	p	N	NUTS2 clusters
Six-item domain-evaluation index	0.307	[0.274, 0.341]	< .001	8066	128
Four-item integrity-criteria index	0.293	[0.259, 0.327]	< .001	8066	128

3 Disaggregated item associations

The separate models estimate each item’s association using the common adjustment set. The mutually adjusted models include every item from the same family simultaneously; these coefficients describe each item’s unique association conditional on the other closely related assessments.

Standardized item associations. Entries are b [95% CI]; p.
Family	Item	Separate item models	Mutually adjusted item models
Service domains	Tax authorities	0.276 [0.239, 0.313]; p < .001	0.126 [0.078, 0.173]; p < .001
Service domains	Local administration	0.260 [0.222, 0.299]; p < .001	0.121 [0.082, 0.160]; p < .001
Service domains	Health care	0.233 [0.201, 0.264]; p < .001	0.063 [0.026, 0.099]; p < .001
Service domains	Benefits and employment services	0.229 [0.199, 0.258]; p < .001	0.062 [0.025, 0.100]; p = 0.001
Service domains	Schools	0.208 [0.179, 0.236]; p < .001	0.021 [-0.011, 0.053]; p = 0.198
Service domains	Public transport	0.204 [0.170, 0.238]; p < .001	0.032 [-0.001, 0.065]; p = 0.058
Integrity criteria	Fair treatment	0.256 [0.223, 0.288]; p < .001	0.159 [0.121, 0.198]; p < .001
Integrity criteria	Integrity	0.238 [0.203, 0.273]; p < .001	0.119 [0.085, 0.153]; p < .001
Integrity criteria	Prompt help	0.211 [0.182, 0.239]; p < .001	0.071 [0.037, 0.104]; p < .001
Integrity criteria	Efficiency	0.192 [0.154, 0.230]; p < .001	0.027 [-0.014, 0.069]; p = 0.200

Separate and mutually adjusted standardized associations with political solidarity.

4 Strongest and weakest associations

Descriptive coefficient ranking within each item family and specification.
Family	Specification	Strongest association	Weakest association
Service domains	Separate item models	Tax authorities (b = 0.276)	Public transport (b = 0.204)
Service domains	Mutually adjusted item models	Tax authorities (b = 0.126)	Schools (b = 0.021)
Integrity criteria	Separate item models	Fair treatment (b = 0.256)	Efficiency (b = 0.192)
Integrity criteria	Mutually adjusted item models	Fair treatment (b = 0.159)	Efficiency (b = 0.027)

These rankings are descriptive. The separate-model coefficients capture overall adjusted associations; the mutually adjusted coefficients answer a different question about unique associations among correlated items. Differences in coefficient size should not be inferred from whether one p-value falls above or below a conventional threshold.

5 Item-level missingness before completion

Item-level missingness before deterministic country-mean completion.
Family	Item	Missing responses	Missing percent
Service domains	Health care	100	1.2%
Service domains	Schools	560	6.9%
Service domains	Tax authorities	904	11.2%
Service domains	Public transport	298	3.7%
Service domains	Benefits and employment services	1823	22.6%
Service domains	Local administration	496	6.1%
Integrity criteria	Fair treatment	0	0.0%
Integrity criteria	Efficiency	0	0.0%
Integrity criteria	Prompt help	0	0.0%
Integrity criteria	Integrity	0	0.0%