Fuzzy cluster profiles.

Impartiality in public services provision is an important dimension that explains the quality of government (QoG). The analysis of impartiality has boomed in recent years at different territorial levels, like countries or regions. The impartiality measures depend on several attributes that are aggregated using different methods. However, little attention has been given to the effects of negative wording attributes and the number of format answers, despite the efforts made by previous studies to build robust composite impartiality indices. This study corrects this existing gap partly using one of the most extensive surveys (the European Quality of Government Index 2021) that include attributes related to impartiality (six attributes and 129,991 citizens). The method will be based on a fuzzy clustering approach, the extended Apostle model and an ordinary binary probit model. The results show that the type of wording and the number of answer options affect impartiality. The analysis of the main differences observed is affected by some insightful covariates such as country, gender, being native, town size, occupation, and the perception of the economic situation.