







































1    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 

 

 

 

 

Explaining Change in Citizens’ Preferences About Intergovernmental 

Responsibilities During the COVID-19 Crisis: The Case of Spain 

 
César Colino, Spanish National Distance-Learning University (UNED) Madrid, ORCID: 0000-0002-9352-705X1 

Gibrán Cruz-Martínez, Complutense University of Madrid, Madrid, ORCID: 0000-0002-4583-29142 

Eloísa del Pino, Institute of Public Goods and Policies, CSIC, Madrid, ORCID: 0000-0001-5497-13023 

Jorge Hernández-Moreno, Institute of Public Goods and Policies, CSIC, Madrid, ORCID: 0000-0002-2085-12034 

 

 

Abstract 

The COVID-19 pandemic brought about some extraordinary shifts in citizens’ 

preferences about intergovernmental responsibilities in several federal states and has 

therefore provided an especially interesting context to contribute to the ongoing debate 

about the scope, direction, and determinants of attitudinal change in citizens’ preferences 

in situations of protracted crisis. Although there is evidence of the role of partisanship 

and some other factors during normal times, the importance that partisanship may have 

with respect to other factors in accounting for changes in citizens’ preferences during 

these crises still needs to be established. Does partisanship account for attitudinal changes 

during a crisis, or do citizens have other predispositions, such as individual core beliefs 

about federalism, perceptions of government performance, or trust in government, which 

could account for the scope and direction of these changes? The article relies on an 

original national survey of 7,175 respondents collected during the transition from the first 

to the second wave of the pandemic in Spain and examines the shift in citizens’ 

preferences in three policy domains: healthcare, nursing homes, lockdown declaration 

and management. It finds that partisanship and attribution of responsibility are relevant 

to explaining shifts in preferences for intergovernmental responsibilities, whereas, 

contrary to expectations, individual beliefs about autonomism are not significant. The 

authors’ findings contribute to the broader literature on the configuration of public 

preferences for multilevel governments and to understanding blame management and 

accountability during crisis situations in federal democracies. 

 

 

  

 

1 César Colino is an Associate Professor at the Department of Political Science and Public Administration 

at the Spanish National Distance-Learning University (UNED) in Madrid 
2 Gibrán Cruz-Martínez is an Assistant Professor at the Department of Political Science and Public 

Administration at Complutense University of Madrid, Spain (corresponding author) 
3 Eloísa del Pino is a Senior Research Fellow at the Institute of Public Goods and Policies (CSIC) in 

Madrid, Spain 
4 Jorge Hernández-Moreno is a researcher at the Institute of Public Goods and Policies (CSIC) in Madrid, 

Spain. 

https://orcid.org/0000-0002-9352-705X
https://orcid.org/0000-0002-4583-2914
https://orcid.org/0000-0001-5497-1302
https://orcid.org/0000-0002-2085-1203


2    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 

1. Introduction 

The flourishing observational and experimental survey research undertaken during the 

pandemic indicates some extraordinary shifts in preferences among citizens concerning 

intergovernmental responsibilities in federal states at different phases of the pandemic. 

For example, a large amount of criticism of the federal system in Germany at the start of 

the pandemic led to an increase in citizens’ support for strengthened powers for the 

federal government, which dropped sharply after a few weeks, to rise again during the 

second wave of the pandemic (Eckhard and Lenz 2020; Juhl et al. 2022). In the US, an 

apparent shift in preferences for federal responsibility due to the pandemic reveals a 

complete reversal in the traditional preferences for federal responsibility among 

Republicans and Democrats (Dinan and Heckelman 2020; Jacobs 2021; Schildkraut, 

Berry, and Glaser 2020). In Spain, the beginning of the pandemic apparently produced 

increased support for the central government (CG), which was preferred to the regional 

government by 73 percent of citizens, but this figure declined to 56 percent by May 2020 

(Barbet Porta 2022; CIS 2020a, 2020b, 2020c; White et al. 2021). In September 2020, the 

same national survey—with a slightly altered wording for the relevant question—found 

that collaboration between central and regional governments was preferred by 72 percent 

of respondents, leaving the central government acting alone with only 16 percent of 

support and the regional governments acting alone with only 5 percent.  

Changes in preferences for policy functions in a multilevel context during a prolonged 

crisis have been insufficiently studied. This is a very relevant issue for the better 

understanding of accountability in democratic and federal systems and the extent to which 

crises complicate this further. Crises tend to reinforce multilevel system features, such as 

collaboration or conflict between federal and subnational governments, 

intergovernmental blame games or buck-passing, competition, outright bickering or open 

confrontation, and even defiance. In this context, citizens can take refuge in heuristics or 

partisan cues to interpret what is happening and structure their beliefs on who should do 

what in the territorial system. However, other studies lead us to believe that during a 

crisis, other factors might gain greater weight in opinion formation due to the exceptional 

nature of the situation, manifested in the issue’s high salience, high individual anxiety, 

and highly available policy information, which allows partisan cues to be ignored. With 

increased media attention, debates in the public sphere focus on what should be done and 

on who should lead responses to the crisis but also on more specific key issues, such as 

who should be managing different crisis-related policies (e.g. lockdowns, curfews, and 

re-openings), who should protect older adults in nursing homes, and who should be 

guaranteeing healthcare for all citizens (Blackburn et al. 2023; Congleton 2021; Del Pino 

et al. 2021; Vicentini and Galanti 2021). Finally, other individual factors, such as deep-

rooted beliefs in government, trust in a specific level of government, or contextual factors, 

such as polarization, the timing of the measures, or trust in national leaders, have been 

found to explain the attitudes of the general public towards policy responsibilities and 

government performance (Altiparmakis et al. 2021).  

This pandemic has thus provided an especially interesting context to contribute to the 

ongoing debate about the scope, direction, and determinants of attitudinal change in 

citizens towards government responsibilities in decentralized systems. This research 

seeks to establish the importance of partisanship with respect to other factors in 

accounting for this change of preference during the crisis. Does partisanship explain the 

changes in citizens’ preferences, or do citizens have other predispositions, such as 

individual core beliefs about federalism or trust in government, which might have been 

activated during the pandemic in response to elite discourses or expert or media 



3    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 

information? Does the evaluation of government policies and growing satisfaction or 

dissatisfaction affect citizens’ preferences, or is this evaluation wholly determined by 

partisan cues and motivated reasoning?  

Spain is a decentralized country which offers us a persuasive case for investigating which 

level of government citizens believe should manage the crisis and to what extent these 

preferences have changed during the pandemic. The Spanish model of federalism, known 

as Estado Autonómico, has several unique features (Colino 2020; Del Pino and Colino 

2024). The Constitution outlines only broad principles such as autonomy and unity, 

leaving the specific power distribution vague but safeguarded by the Constitutional Court. 

It has evolved into a highly decentralized system concerning public spending and policies, 

although the central government maintains significant control through shared powers and 

revenues. The system exhibits both centrifugal forces, encouraging regional 

differentiation and centripetal forces, thus promoting uniformity. Its political dynamics 

are influenced by strong nationalist movements in specific regions, which have led to 

recognized asymmetries and impacted the system’s evolution. However, the system has 

transitioned from an asymmetrical to a symmetrical cooperative model without formal 

constitutional amendments through political agreements, judicial interpretations, and 

updates to regional statutes. 

Concurrently, Spain has constructed a modern welfare state, with regional and local 

governments now managing over half of the public expenditures and employing 75 

percent of public employees. Since 2002, healthcare governance in Spain has been 

decentralized, giving power to the 17 regional governments, which now handle 92.5 

percent of the country’s healthcare expenditure. The central government maintains 

regulatory oversight, setting the fundamental entitlements, organization, and financing of 

the National Health System (NHS), while the autonomous communities manage the 

system’s operation. Coordination between these levels is facilitated by an 

intergovernmental body, the Inter-Territorial Council of the NHS, comprising central and 

regional health ministers (Mattei and Del Pino 2021). Spain was one of the countries most 

severely affected by the COVID-19 health crisis, which claimed many victims. It 

witnessed different political constellations and degrees of intergovernmental agreements 

and conflict in response to the crisis during the pandemic’s various phases (Hernández-

Moreno and Harguindeguy, 2024; Hernández-Moreno, Pereira-Puga, and Cruz-Martínez 

2023; Navarro and Velasco 2022; Pereira-Puga, Hernández-Moreno, and Cruz-Martínez 

2023). As a relevant contextual element, Spaniards have experienced increasing 

polarization of political attitudes in general, which also influenced their attitude towards 

crisis management in several phases of the health crisis.  

Several hypotheses are tested, which aim to establish the scope and determinants of the 

changes in citizen’s preferences. An original national representative web-based survey of 

7,175 respondents was used. The dependent variable is the change in preferences for the 

level of government which should be responsible for managing the COVID-19 pandemic 

in three policy domains: healthcare, nursing homes, lockdown declaration and 

management. It is found that during the transition from the first to the second wave of the 

pandemic, only a small share of the population changed their minds about government 

responsibilities (i.e., most people did not change their opinion on preferred governmental 

responsibility during the initial period of very serious public health emergency). 

Partisanship and trust in institutions matter, but other individual factors also become more 

critical, displacing the importance of partisanship, depending on the policy initiative. 



4    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 

This study contributes to the broader literature on how the public assigns preference for 

policy responsibility to various government levels (Del Pino and van Ryzin 2013; 

Schneider, Jacoby, and Lewis 2011). It is also valuable for studying democratic 

accountability and understanding how citizens in federal democracies can hold their 

governments accountable. It helps us understand under what conditions citizens will 

simply seek refuge in heuristics or partisan cues to make sense of what is happening in 

the multilevel political context, during a crisis where politicians and governments will 

typically frame issues of policy as questions of federal governance and will play blame 

games to avoid accountability for bad policy results. 

The article is organized as follows: in the next section, the existing literature on changing 

preferences about government responsibility in normal and crisis times and the 

explanatory factors which have been identified are reviewed. Then, some theoretical 

expectations for the Spanish case are presented, followed by a description of the data 

collected and the empirical strategy applied. After introducing the statistical analysis, the 

results and the implications of this research are discussed, suggesting some future 

research avenues. 

 

2. Explaining Public Preferences Regarding Intergovernmental Responsibility in 

Normal Times and Crises  

The literature provides three contrasting, inconclusive sets of expectations about the 

factors behind citizens’ preferences about government responsibility and how they 

change in normal times and during crises (Schneider and Jacoby 2011, 2013; León, 

Jurado, and Garmendia-Madariaga 2018). The first view contends that preferences are 

often driven by partisanship and motivated reasoning (Kunda 1990). Party positions move 

opinions, leading citizens to become more supportive of their own party’s policy position 

(Slothuus and Bisgaard 2020; Viskupic and Wiltse 2023). Partisanship could act as a 

“judgmental shortcut, efficient ways to organize and simplify political choices” 

(Sniderman, Brody, and Tetlock 1991, 19). Regarding government-level preferences in 

the US, for example, Democrats become more supportive of decentralization when the 

Republicans control the federal government (Wolak 2016), which means that support for 

decentralization tends to decrease when an individual’s party is in office at the central 

level and increases when the other party governs. When attributing blame and 

responsibility across territorial levels of government, citizens will assign more blame to 

the state or regional government when it is controlled by the opposing party and will 

unevenly punish regional officeholders for unpopular federal or central presidents and 

hold their own party and opposing parties to different standards of performance (Brown 

2010). Individuals will ignore damaging information when it challenges an individual’s 

partisan priors (Druckman, Peterson, and Slothuus 2013). Motivated reasoning based on 

partisan identities and government preferences influences individuals’ assessment of 

information about governmental performance and evidence of strength (James and Van 

Ryzin 2017; Torcal and Mota 2013). 

According to these studies, few people can be motivated to incorporate knowledge of 

policies or federal division of power into their judgements. Instead, they rely on easy and 

ready heuristics to inform their evaluations. Partisan identification also becomes stronger 

and less ambivalent under polarized conditions, leading to stronger party cue effects and 

increased motivated reasoning. Polarized environments fundamentally change how 

citizens make decisions by intensifying the impact of party endorsements on opinions and 



5    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 

decreasing the impact of substantive information on opinions (Druckman and Lupia 2016; 

Druckman, Peterson, and Slothuus 2013). 

Much of the research that appeared during the pandemic supports the idea that the effect 

of partisanship can be even more pronounced in times of crisis in federal countries 

(Glaser, Berry, and Schildkraut 2023; Goidel et al. 2024; Jacobs 2021; Lobera, Santana, 

and Gross 2024; Snow and Evans 2024; VanDusky-Allen et al. 2022). Rodriguez et al. 

(2020) found consistent evidence of partisan divergence in pandemic response policy 

preferences across the first six months of the crisis: Republicans supported national 

control measures, whereas Democrats supported welfare policies, and these interparty 

differences grew over time. Milosh, Van Dijcke, and Wright (2020) found only limited 

evidence that exposure or experience moderates these partisan differences. León and 

Garmendia-Madariaga (2020) observe a centralizing trend in opinion in Spain, with 

partisanship exerting an apparent moderating effect. 

A second view has been developing in recent years, with many studies increasingly 

showing citizens as capable of reasoning without bias (Arceneaux and Vander Wielen 

2017) and presenting a more complex picture of the public and its ability to process 

information both heuristically and systematically (Ciuk and Yost 2016). Given certain 

conditions, motivated partisan reasoning can be limited (Groenendyk 2013). Elite partisan 

influence and, therefore, support for governments or leaders and levels of government 

might be limited by individual factors and the perception of efficacy (Arceneaux and 

Vander Wielen 2017; Mullinix 2016), by emotions, such as anxiety and risk perception 

(Albertson and Kushner Gadarian 2015), competition for policy information and 

arguments (Bullock 2011), polarization (Druckman, Peterson, and Slothuus 2013), as 

well as issue salience (Ciuk and Yost 2016). Party cues do not inhibit thinking about 

policy, and people’s attitudes seem to be affected at least as much by information, 

reference groups, and cues from other party elites (Bullock 2011; Grofman and Norrander 

1990). In this line, some research has also studied how people evaluate government and 

allocate blame in normal times or crisis and whether a preference shift towards more 

centralization or decentralization occurs (Arceneaux 2008; Arceneaux and Stein 2006). 

Due to the high degree of risk perception, this research finds less partisan bias and more 

accuracy-based reasoning during a crisis than in normal times. Many citizens, 

emotionally moved by the health crisis, engage in accuracy-based reasoning to inform 

and update their opinions, suppressing some of their predispositions and attributing blame 

that may contradict their usual partisan biases (Atkeson and Maestas 2012). 

Finally, a third view has consistently argued that citizens have beliefs or values about the 

territorial system, the right amount of decentralization, and how the territorial division of 

powers should work. This expression of a federalist culture or set of beliefs has been 

called citizens’ “intuitive federalism” (Schneider and Jacoby 2013). Wolak (2016) finds 

that these deeply rooted values or core beliefs can anchor individuals’ federalism and 

policy devolution preferences. These beliefs in federalism can inform reactions to policy 

debates, and thus make people less likely to rely on policy concerns and more likely to 

form preferences based on their beliefs about federalism (Jacobs 2017; Kam and Mikos 

2007; Rendleman and Rogowski 2024). This attitudinal configuration is not always 

manifested but can be activated in cases of crisis and amplified by elite and media 

discussions of system failures (Kam and Mikos 2007; Pears and Sydnor 2022). 

In sum, the second and third views generally contend that in extraordinary times, such as 

natural disasters or pandemics, the role of partisanship in preference formation and blame 

attribution diminishes, meaning that a non-partisan shift of preferences can be expected, 



6    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 

giving priority to other individual short-term predispositions or perceptions, deep-rooted 

beliefs, and contextual factors. Information, especially from expert and non-partisan 

sources, can prompt citizens to overlook their party affiliation and change preferences or 

to assign blame in a more performance-based manner by freeing individuals from partisan 

biases on how they find information or by reinforcing their individual beliefs or 

predispositions. 

 

3. Toward a Theory of Citizen Preference Changes in the COVID-19 Pandemic 

This section proposes a set of testable propositions about the conditions under which 

citizens change their preferences about intergovernmental policy responsibilities during 

a pandemic like COVID-19. It is reasonable to expect that some citizens would change 

their preferences over the course of the crisis. It is well known that in the early stages of 

a crisis, there is a rally-around-the-flag effect. Public opinion and leaders momentarily 

set aside their ideological differences and political interests and support the government, 

although in the long run, the rally-around-the-flag effect diminishes (Johansson and 

Shehata 2021; Kritzinger et al. 2021). The same literature suggests that after an initial 

alignment with the (central) government, citizens’ attitudes can change. As the crisis 

evolves, citizens will begin to evaluate the results of the various authorities’ responses to 

different policies and hold them accountable.  

From the literature on blame management, it is known that governments themselves 

change strategies during crises and use discursive or presentational strategies, such as 

scapegoating, spinning, or framing (Blackburn et al. 2023; Porumbescu et al. 2022), to 

claim credit from or shift blame to other levels of government. Citizens’ attitudes react to 

these strategies, and therefore, their preferences may change over the course of the crisis. 

One can also expect a certain negativity bias. Regardless of the government’s policy 

responses, any action by the authorities can be viewed critically, since a pandemic will 

bring with it a high degree of chaos typical of such crises, combined with death and 

personal suffering (Flinders 2020).  

3.1 The Influence of Partisanship on Preferences 

In troubled waters, confusion will grow over the responsibilities and the adequacy of 

solutions. As the media will offer multiple perspectives on a health crisis, one can expect 

individuals to gravitate toward narratives expressing attitudes congruent with their 

existing personal predispositions. With increasing awareness of coordination failures, 

blame games, and the resulting lack of clarity about governmental performance, 

individuals’ sources for opinion formation about responsibility will tend to be based 

increasingly on partisan cues rather than on performance evaluations, federalist or 

autonomist beliefs, or feelings of trust in the different levels of government (Biddle, Gray, 

and McAllister 2024). The likely change in individuals’ preferences will then be shaped 

by the party control of different governments and by their preferred parties’ cues through 

media discourse. The party which controls the government of the region in which an 

individual lives will determine whether their preferences shift towards greater 

centralization. It is argued that partisanship could act as a heuristic that simplifies 

complex decisions during a crisis, leading to a preference for a certain government level. 

The capabilities of party leaders are essential to shifting public citizens’ preferences on 

questions of central versus regional authority. Jacobs (2021) has shown how shifting 

rhetoric and strategic framing led citizens to favour different levels of government during 



7    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 

the pandemic. Moreover, in polarized systems, citizens are more sensitive to public 

officials or partisan cues across party lines (Kincaid and Leckrone 2020).  

H1: Partisanship and party positions will determine the shift in citizens’ preferences 

towards centralized or decentralized pandemic management. 

H1a: Right-wing voters (PP and Vox5) will support further decentralization of 

responsibility, regardless of their individual preference for autonomism or 

satisfaction with their government’s performance. 

H1b: Left-wing voters (PSOE and UP6) will support the further centralization of 

responsibility towards the central government regardless of their regional 

incumbent or satisfaction with their government. 

H1c: When a voter’s party is in office at the regional level, there will be an 

intergovernmental shift in preferences towards decentralization. 

3.2 Satisfaction with Performance and Citizens’ Government Responsibility Preferences 

Citizens are sometimes partially able to set aside their partisan biases. Due to the high 

issue salience of the COVID-19 pandemic, it is expected that most people will be 

sophisticated enough to process increasing information (Arceneaux and Vander Wielen 

2017). Citizens in Spain were continuously exposed to comparative information on 

COVID-19 cases by country and region. This informed their views on government 

effectiveness, making it easier for them to evaluate the performance of different levels of 

government and ultimately conditioned their preferences. Following Ciuk and Yost 

(2016), it is expected that during the health crisis, the increased issue salience that the 

pandemic generated motivated people to go beyond heuristics and engage in the 

systematic processing of policy-relevant information. That means that we can expect 

citizens to evaluate and credibly process information in the media, combined with a need 

to assess threatening events fully and accurately (Albertson and Kushner Gadarian 2015; 

White et al. 2021). 

It is expected that only the people who are satisfied with the government’s management 

performance and consider it effective will maintain their preferences. In contrast, the 

authors expect that dissatisfaction with the management of the pandemic could lead 

citizens to change their preferences or produce a certain desire for change (Arceneaux 

and Vander Wielen 2017; Mullinix 2016).  

H2. The more satisfied citizens are with the management of a policy initiative, the 

less likely they are to want another level of government to perform this task. 

3.3 Autonomist Beliefs, the Crisis and Preferences About Government Responsibilities 

As recent studies by Wolak (2016) and Rendleman and Rogowski (2024) show, 

individuals with stronger federalist or autonomist values, as measured by their preference 

for subnational power over national power, are more likely to support policy devolution 

across various policy areas. Federalist or autonomist values, such as beliefs about the 

appropriate balance of power between national and regional governments, play a crucial 

 

5 PP stands for Partido Popular (Popular Party). It is the main center-right party in Spain and the leader 

of the opposition to the current government. VOX is a populist radical right party and the third-largest 

political force, ranking just below PP. 
6 PSOE stands for Partido Socialista Obrero Español (Spanish Socialist Workers' Party). It is the main 

center-left party in Spain and the leader of the coalition government with Unidas Podemos (UP)—a 

coalition of left-wing parties in its own right. 



8    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 

role in shaping individuals’ preferences towards policy devolution. Such beliefs can shape 

attitudes toward various policy areas, including fiscal management, education, healthcare, 

and cultural policies. This support for shifting authority from the national level to 

subnational levels is influenced by an individual’s core beliefs about the role of 

government at different levels. For example, according to the characteristics of different 

federations, these core beliefs or values could derive from an individual’s support for 

historical states’ rights, limited government, subsidiarity, or the identification with a 

minority nation within a federal multinational state. These deeply rooted orientations are 

anchored in a substantive desire for smaller, closer government or ethnonational self-

government at the subnational level and are not merely a reflection of short-term partisan 

interests or considerations. They form a principled basis for reasoning about federalism. 

People’s preferences for devolution are thus responsive to the existing balance of power, 

demonstrating a nuanced understanding of federalism that goes beyond partisan 

heuristics. Individuals with federalist and autonomist beliefs are more likely to advocate 

for policy devolution, seeking to shift authority from the national level to subnational 

levels and guiding people to consider the alignment between the structure of government 

and their broader values. 

As Wolak and Kelleher Palus (2010) have also pointed out, these autonomist beliefs can 

intensify during periods of crisis, such as the COVID-19 pandemic, when the 

effectiveness of local responses becomes more visible and relevant to citizens and not just 

in partisan matters. The COVID-19 crisis and the high media attention may have also 

triggered federalist and autonomist beliefs, which have been described as citizens’ 

intuitive federalism. Strong autonomist beliefs would push citizens towards greater 

support for decentralization or the sharing of responsibilities between different levels of 

government in the management of the pandemic. 

However, it is also known that citizens’ preferences for one or another government level 

may not necessarily be uniform across different policies aimed at combating the pandemic 

(Connolly et al. 2020). Citizens’ views about the assignment of policy responsibilities 

may vary by policy sector and even within a sector (Thompson and Elling 1999). 

Additionally, citizens frequently exhibit contradictory attitudes, generally favouring 

federalism and decentralization yet advocating for centralization in numerous policy areas 

(Jedwab and Kincaid 2018). Citizens also show an inability to assign responsibility 

correctly (Kennedy, Sayers, and   2022). 

Some differences in preferences are then expected among areas, mostly consistent with 

the constitutional distribution of power. These deep-rooted preferences are expected to 

be pretty consistent with each other and, therefore, to vary both according to the pandemic 

policy area and to its different phases, adapting to the circumstances as people perceive 

them to change. Given the extent of media discussion during a crisis, people are expected 

to become more sophisticated at pinpointing the responsible levels of government and at 

assigning to each what they do and what they should do. 

H3. The stronger an individual’s autonomist beliefs, the greater the shifts in support 

towards     subnational or shared responsibilities, with variations according to the 

type of pandemic policies 

3.4 Attribution of Responsibility and Changes of Preferences 

Even if in some disasters or crises there is no culprit, citizens end up blaming the 

government, which they hold responsible for managing the crisis (Skarżyńska, 

Urbańska, and Radkiewicz 2021). This is especially true when crises become protracted 



9    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 

(Arceneaux and Stein 2006). A protracted crisis is characterized by a prolonged period 

of instability involving complex and multifaceted challenges which are difficult to 

solve, e.g., the COVID-19 crisis (Boin 2024; Boin and ‘t Hart 2022; Boin, McConnell, 

and ‘t Hart2021; Boin and Rhinard 2023). In fast-burning crises or in their early 

moments, there is less opportunity to find someone to blame. Citizens and even the 

opposition need time to understand the magnitude of the crisis and for reasons of 

loyalty, solidarity, or in order not to hinder efforts to end the crisis, they avoid looking 

for culprits at first. However, in a protracted crisis, the severity of the situation as well 

as the political and media dynamics may “significantly alter levels of political support 

for public officeholders and public policies” (Boin, ‘t Hart, and McConnell 2009, 83) 

even if experienced only vicariously through media coverage (Atkeson and Maestas 

2012). A decentralized political system provides the opportunity for a change of 

preferences from one level to another (Heinkelmann-Wild et al. 2020), so the following 

is expected: 

H4. In a protracted crisis, there is a higher probability that an intergovernmental 

shift of preferences in the opposite direction of the level of government to which 

responsibility is attributed for the policy area will be observed. Those attributing 

responsibility to the central government will shift their preferences towards 

decentralization and vice versa.  

From a longitudinal perspective, this hypothesis allows us to understand the evolution 

of citizens’ attitudes during the crisis. 

3.5 Control Variables 

The authors considered both theoretical frameworks and empirical evidence in the 

selection of control variables for the multivariate analysis to ensure a comprehensive 

analysis. Education, social class, and political variables are incorporated as control 

variables based on both their empirical relevance to the selected dependent variable and 

foundational theories in political science and sociology. 

Theoretically, education is often linked to increased political awareness and cognitive 

skills, which enable individuals to understand and engage with complex governmental 

structures. This is aligned with theories suggesting that higher education levels equip 

individuals with the cognitive tools necessary to evaluate the benefits of centralization in 

policy implementation, particularly in nuanced or complex policy areas (Dalton 2018). 

This theoretical perspective supports Konisky’s (2011) empirical finding that individuals 

with higher levels of education tend to prefer centralization for certain policies. 

Theoretical perspectives on fiscal federalism suggest that economic resources influence 

political preferences, including governance structures. Higher-income groups with more 

at stake in terms of property and taxes can prefer decentralized government systems, 

which often allow more localized and potentially favourable fiscal policies (Banzhaf and 

Walsh 2008). This theoretical framework supports the inclusion of social class as a 

control variable in examining preferences for governmental decentralization. This aligns 

with Thompson and Elling’s (1999) findings that individuals from higher income brackets 

are more likely to support state and local governance. Similarly, Schneider, Jacoby, and 

Lewis (2011) contribute to this discussion by indicating that the unemployed and those 

facing economic vulnerability, prefer the national government’s capacity for mobilizing 

extensive resources during crises. 

Beyond socioeconomic factors, political theories emphasizing trust and legitimacy argue 

that trust in governmental institutions fundamentally shapes citizens’ governance 



10    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 

preferences (Wolak and Kelleher Palus 2010). This theoretical basis supports the 

inclusion of trust as a control variable, reinforced by Blackburn et al.’s (2023) empirical 

finding that trust in government is a significant predictor of approval of pandemic 

governance. These theoretical frameworks provided a robust foundation for the selection 

of the control variables utilized, ensuring that our empirical investigation is grounded in 

a nuanced understanding of the political and social dynamics influencing preferences for 

governmental decentralization and centralization. 

 

4. Research Design and Measurement 

4.1 Data Collection 

To examine these hypotheses, a representative sample of the Spanish population was 

selected, and an original survey instrument was designed to assess the intergovernmental 

shift of preferences to manage the pandemic in three policy areas. The sample consisted 

of 7,175 respondents aged over 18, with a 99 percent confidence level and a 2 percent 

margin of error.7 

The data was gathered by Netquest (www.netquest.com) using a web-based survey with 

a nonprobability sampling approach and op-in panels between August 30 and September 

23, 2020 (i.e., coinciding with the protracted period of the second wave of COVID-19 in 

Spain).8 The questionnaire took an average of 11 minutes to complete, which is similar 

to the median period for Netquest questionnaires in Spain (Revilla 2017). 

Netquest sent 39 rounds of invitations to their respondents’ panel, making sure to have 

national-level representative quotas for gender, education, and age. The response rate was 

52.6 percent. The resulting sample has a similar demographic composition to the 

prestigious and extensive representative surveys conducted by the Spanish Centre for 

Sociological Research (CIS) regarding gender, age, and education quotas. Sample 

statistics and a translated version of the questionnaire for dependent and explanatory 

variables appear in Tables A.7 and A.8 of Appendix I. 

4.2 Dependent variable  

The dependent variables are binary variables measuring whether there was a centralizing 

or decentralizing intergovernmental shift in preference for the management of the 

lockdown, nursing homes, and healthcare. They were constructed from the following two 

survey questions: 

1) Imagine that today is March 1 – that is, at the beginning of the coronavirus 

crisis – and you are given the choice of which level of government should be 

primarily responsible for managing the following issues. Which level of government 

would you have chosen on March 1? 

2) If now – 5 months after the crisis began, and knowing what you know – you 

could choose again, which level of government should be primarily responsible for 

managing the coronavirus crisis? 

 

7 Each of the 17 autonomous communities (ACs) has a representative sample for its population with a 95 

percent confidence level and a 5 percent margin of error, except for the deliberate oversample of Catalonia 

and the Basque Country (each case with 700 respondents). 
8 A pre-test was launched on August 14–16, 2020, with a sample of 363 individuals. 



11    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 

The two questions were displayed in two checkbox grids, and respondents could choose 

from the central government, the autonomous community (ACs) regional government, 

or shared responsibility between both levels.  

The binary variable is constructed first using a 5-point scale (2 to –2) with the following 

formula: 𝐷 = 𝑋𝑡1 − 𝑋𝑡2; D is the dependent variable, 𝑋𝑡1 refers to the preferred 

government level in March 2020 and 𝑋𝑡2 is the preferred government level at the time of 

completing the survey (September 2020). The central government is operationalized with 

a 1, both government levels with a 2, and regional government with a 3. Once it was 

determined whether there was a shift of preferences towards centralized (2 or 1), no 

change (0), or decentralized (–1 or -2), the two binary variables were operationalized. 

Table 1 shows the operationalization.  

Table 1: Operationalization of the dependent variable

 

The change in preferences was thus measured with retrospective survey questions 

concerning the preferred government level to manage the crisis in early March, before the 

COVID-19 outbreak. The cross-sectional survey data was transformed into longitudinal 

data to measure the change in preferences. Although retrospectively measuring changes 

in preferences or attitudes is contentious due to memory bias and because prior 

preferences can be confounded with current ones, research validates the use of 

retrospective questions because of recall accuracy and its limited impact on the validity 

of research findings (Hipp et al. 2020). When measuring attitude or preference changes, 

recall data yields similar results to longitudinal data (Jaspers, Lubbers, and De Graaf 

2009). Moreover, the use of retrospective questions six months into the pandemic allowed 

respondents to be informed about their preferences, as the population tends to be more 

informed during prolonged crises. Using recall data is also a normal practice in areas such 

as electoral studies. 

The pandemic generated a natural experiment of unparalleled proportions (Bjørnskov and 

Voigt 2022; Rosen 2021). It would, therefore, be relevant to consider its potential impact 

on the shift in preferences for intergovernmental responsibilities and its determinants. 

Logit models were conducted with and without controls to test the hypotheses. The first 

six models examine the reasons behind the shift in preferences towards centralization. If 

respondents shifted their preference towards centralization (e.g., from regional 

government in March to central government in September), their shift was coded as 1, 

while respondents with no change in preference or with a shift towards decentralization 

were coded 0. The last six models assess the same for respondents which displayed a shift 

in preferences towards decentralization. Therefore, if respondents shifted their preference 

towards decentralization (e.g., from central government to shared responsibility), their 

Table 1: Operationalization of the dependent variable 

Preferred level of government     

March September 
Additive 
formula 

Operationalisation Interpretation 

Regional (3) Central (1) 3-1 2 
Centralized shift two levels 

up 

Regional (3) 
Shared (2) 

Shared (2) 
Central (1) 

3-2 
2-1 

1 
Centralized shift one level 

up 

Central (1) 

Shared (2) 
Regional (3) 

Central (1) 

Shared (2) 
Regional (3) 

1-1 

2-2 
3-3 

0 
No intergovernmental 

change of preference 

Shared (2) 

Central (1) 

Regional (3) 

Shared (2) 

2-3 

1-2 
-1 

Decentralized shift one level 

down 

Central (1) Regional (3) 1-3 -2 
Decentralized shift two 

levels down 

 



12    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 

shifts were coded as 1, while respondents who did not change or who shifted preferences 

towards centralization were coded as 0. This construct allows the factors behind the shift 

in preferences towards centralized and decentralized government to be measured and 

tested separately for each policy area.   

 

4.3 Explanatory Variables  

The first variable (to test H1) is partisanship, and it is operationalized using three dummy 

variables. The first indicator shows whether the respondent voted for any of the political 

parties in the central government. The second tests for the same but at the regional level. 

The third indicator assesses whether the respondent voted for either of the two main 

opposition parties in the central government (i.e., the right-wing parties of PP and VOX).  

The second variable (to test H2) is satisfaction with the performance of the government 

managing the pandemic for each of the three policy areas. Respondents were asked for 

their assessment using a 5-point Likert scale: How do you assess the work the government 

has done concerning each of these three tasks? A 5-point Likert scale from very good (5) 

to very bad (1) was used.  

The third variable (to test H3) is operationalized using a composite indicator between 

three kinds of structural attitudes towards the state’s territorial organization (autonomist 

beliefs). The questionnaire included three questions for this purpose. 

1) a 10-point scale question on the degree of preferred centralization or 

decentralization for Spain, where 0 represents “maximum centralization” and 10 

represents “maximum decentralization.” 

2) a 5-point Likert scale on satisfaction with the creation and development of the 

current model of territorial organization in Spain (distribution of competencies 

between 17 ACs), where 1 is very low and 5 is very high.  

3) a 5-point Likert scale Linz-Moreno question, asking respondents to self-identify 

with an exclusive or dual territorial identity (1=only self-identified as Spanish, 2–

4=different degrees of dual identity, 5=only self-identified as being from the AC).  

All three indicators were normalized and then assigned the same weight using an 

arithmetic mean to construct the composite indicator.  

The fourth variable (to test H4) assesses the impact of the attribution of responsibility for 

each of the three tasks on the shift in intergovernmental preferences. Respondents were 

asked which level of government—central, regional, or both—they thought was mainly 

responsible for managing each of the three tasks. The responses were then transformed to 

a Likert scale (–1 to 1) to distinguish in the model respondents who favoured national, 

shared, or subnational authority. Respondents choosing central government were coded 

as 1, shared responsibility as 0, and regional responsibility as –1.  

4.4 Control variables  

Data on various indicators were collected to control for individual political and 

socioeconomic factors. The political factors are the individuals’ ideological 

identifications (0–10 scale), interest in politics (1–5 Likert scale), and trust in government 

(1–5 scale).9 The socioeconomic factors are net household income (1–5 scale), health-

 

9 Confidence in regional and central government was highly correlated in the pre-test, the same with 

regional and central parliaments. The authors, therefore, kept the questions at the regional level.  



13    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 

related risks (a dummy variable to measure whether the respondent contacted the health 

authority with the suspicion of having COVID-19), vulnerable employment status (a 

dummy variable to measure whether the respondent was unemployed or covered by a 

temporary lay-off scheme), subjective social class (1–7 scale) and education (1–5 scale). 

A context-level variable was also included in the model: seven-days accumulated 

incidence of COVID-19 in the region the day the survey was taken. Incidence data was 

built using data from the Spanish National Epidemiological Surveillance Network 

(RENAVE) managed by the National Epidemiology Center. See Table A.7 in the 

Appendix I for details on the operationalization of the control variables.  

 

5. Results: Shift in Intergovernmental Preference During the Pandemic and Its 

Determinants 

Citizens preferred the lockdown to be managed by the central government in March and 

September (52.5 and 47.7 percent of respondents). This choice is consistent with the 

central government’s power to manage the pandemic during the state of emergency, 

meaning that it is possible that citizens considered that the central government should 

manage the pandemic response given the scope of the measures needed. However, 

preferences were more dispersed as regards nursing homes and healthcare. Less than a 

third of citizens in both cases preferred the central government to be primarily responsible 

for these tasks. In these cases, citizens’ preferences are consistent with the existing 

distribution of powers in Spain, where the regional governments manage these two policy 

areas (Table 2).  

Table 2: Citizens preferred level of government to manage the pandemic 

 

Overall, 37 percent of citizens changed their preferences during the crisis for at least one 

of the three tasks examined here. Figure 1 shows the share of the population with a 

decentralizing, centralizing, and no shift in intergovernmental preferences. Figure 2 

shows the share of the population with a centralizing and decentralizing shift and 

excluding those with no change. Clearly, the vast majority of the population (almost four 

out of every five respondents) did not change their preferred level of government between 

the first and second wave of the pandemic. This near-zero variance might make finding 

the determinants for the shift in preferences more difficult. Fifty-eight percent of those 

who changed went towards decentralization, especially in the management of the 

lockdown. 

  

Table 2: Citizen’s preferred level of government to manage the pandemic 

Prefered level Lockdown Nursing homes Healthcare 

 
March September March September March September 

Central government 52.50% 47.72% 31.38% 30.69% 31.03% 29.84% 

Shared responsibility 31.95% 32.69% 38.16% 38.57% 41.95% 42.20% 

Regional government 15.55% 19.59% 30.46% 30.73% 27.03% 27.96% 

 



14    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 

Figure 1: Intergovernmental shift of preferences across the three policy areas 

 
Note: The sample size is 7,022 for the lockdown, 6,957 for nursing homes, and 6,962 for healthcare. 

 

Figure 2: Intergovernmental shift of preferences across the three policy areas 

(excluding no changes of preferences) 

 

 
Note: After excluding participants without changes of preferences, the sample size is 1,460 for the 

lockdown, 1,620 for nursing homes, and 1,463 for healthcare. 

Before proceeding to the logit models to test the hypotheses, the most relevant trends will 

be summarized in the descriptive analysis (see Tables A1 to A5 in Appendix I for details 

on the descriptive statistics). The relationships between explanatory variables and 

intergovernmental shifts in preferences seem to vary according to the policy area. 

Partisanship appears to be relevant for H1a and H1b. Of those with a shift in preferences, 

right-leaning (PP and VOX) voters shifted towards decentralization and left-leaning 

(PSOE and UP) voters shifted towards more centralization. This is clear for the 

management of healthcare and nursing homes but not for the lockdown. Moreover, if the 

respondents’ preferred party was in office at the AC where the respondent lived, the 

chances for preferring more decentralization increased for all three policy areas. An 

unprecedented concentration of executive power was immediately reflected in the near-

daily media appearances of several government figures, such as the Prime Minister (Pedro 

13.88 12.25 11.38

79.21 76.86 78.99

6.91
10.90 9.64

0.00

10.00

20.00

30.00

40.00

50.00

60.00

70.00

80.00

90.00

Lockdown Nursing homes Healthcare

%
 P

o
p

u
la

ti
o

n

Decentralization shift No change of preferences Centralization shift

66.78

52.92 54.14

33.22

47.08 45.86

0.00

10.00

20.00

30.00

40.00

50.00

60.00

70.00

80.00

Lockdown Nursing homes Healthcare

%
 P

o
p

u
la

ti
o

n

Decentralization shift Centralization shift



15    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 

Sánchez), the Minister of Health (Salvador Illa), and the director of the Centre for the 

Coordination of Health Alerts and Emergencies (Fernando Simón). Following the usual 

pattern for political crises, this initial reconsolidation encountered little resistance despite 

not being explicitly provided by the Constitution (Boin and t’Hart 2003). Similar to the 

Great Recession of 2008, competencies were recentralized. However, parliamentary 

support for the state of emergency eroded with each of the six votes in the lower chamber. 

Particularly, the demands to restore regional autonomy grew increasingly strong among 

certain ethnonationalist parties. On one side, the Basque Nationalist Party (PNV), Canary 

Coalition, Bildu, New Canaries, Galician Nationalist Bloc (BNG), Regionalist Party of 

Cantabria (PRC), and Teruel Existe supported the central government or abstained 

throughout the six votes, while continuously criticizing the concentrated centralism 

implied by the state of emergency. On the other side, the Republican Left of Catalonia 

(ERC) toughened its stance, as did Together for Catalonia (JxCat), Popular Unity 

Candidacy (CUP), Compromís for the Valencian Country, Navarrese People's Union 

(UPN), and Asturias Forum. Thus, the sixth vote was held on June 3, 2020, with 177 votes 

in favour, 155 against, and 18 abstentions: a very close result for the government during 

the de-escalation phase. 

Satisfaction with the government’s performance in managing the crisis (H2) seems to be 

associated with respondents changing their preferences towards decentralization of the 

management of nursing homes. Interestingly, the share of the population favouring 

decentralization exceeds those favouring centralization for all degrees of satisfaction. 

This is particularly marked for lockdown management. The degree of autonomism (H3) 

does not seem to be a relevant factor in explaining the general population preference 

changes. There is no clear trend or difference in the autonomism scores between those 

shifting towards centralization or decentralization. The attribution of responsibility (H4) 

seems to be a relevant determinant of intergovernmental shift. Those shifting in all three 

areas who attributed responsibility to the central government mainly shifted towards 

decentralization, while those attributing responsibility to the regional government shifted 

towards centralization. Therefore, the shift in intergovernmental preferences seems to 

also be motivated by blame or retribution. 

5.1 Explaining the Changes in Preference: Logit Models 

In this section, the logit models are presented (Tables 3 and 4). First, the authors’ analysis 

appears to confirm H1a and H1b regarding partisanship and the incumbency of national 

governments. Having voted for PSOE-UP—which leads the left-wing governing 

coalition— for the central government is associated with a centralizing shift in the 

management of nursing homes and healthcare but not in the management of the 

lockdown. This holds true after controlling for political, socioeconomic, and contextual 

factors. The odds of having a centralizing preference shift for the management of nursing 

homes is 43 percent higher (odds ratio=1.43, 95% CI=1.13 – 1.82, p=0.003) and 62 

percent higher (odds ratio=1.62, 95% CI=1.26–2.10, p<0.001) for the management of 

healthcare if the subjects had voted for any of the two parties in the central government 

coalition (PSOE or UP). After including controls in models 5 and 6 (Table 3), the 

likelihood of a centralizing shift in the management of nursing homes and healthcare 

increased for PSOE-UP voters. The likelihood of having a centralizing preference shift 

for the management of nursing homes is 59 percent higher (odds ratio=1.59, 95% 

CI=1.21–2.11, p<0.001) and 64 percent higher for the management of healthcare (odds 

ratio=1.64, 95% CI=1.22 – 2.21, p<0.001) if the subjects had voted for the left-wing 

central government coalition. 



16    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 

Table 3: Logit models explaining the change in preferences for centralized decision-

making in the management of the lockdown, nursing homes and healthcare 

 
 

 

 

 

 

 

 

 

Table 3. Logit models explaining the change of preferences for centralized decision-
making in the management of the lockdown, nursing homes and healthcare 

 (1) (2) (3) (4) (5) (6)    
Satisfaction with 

govt. performance 

.044 

(.061) 

.067 

(.051) 

-.028 

(.048) 

.078 

(.069) 

.072 

(.058) 

.002 

(.054) 

Partisanship        
Central government 
  

-.066 
(.149) 

.358** 
(.122) 

.485*** 
(.131) 

.114 
(.175) 

.466** 
(.142) 

.495** 
(.153) 

Regional government 

  

-.129 

(.127) 

-.043 

(.099) 

.003 

(.105) 

-.065 

(.148) 

-.069 

(.115) 

-.039 

(.123) 

Vote to right-wing 
opposition parties in 

the CG (PP+VOX)  

-.172 

(.190) 

-.074 

(.166) 

.113 

(.171) 

-.137 

(.232) 

-.074 

(.197) 

.044 

(.207) 

Degree of 
autonomism 

  

-.033 

(.037) 

-.034 

(.030) 

-.042 

(.031) 

.028 

(.044) 

-.023 

(.035) 

-.032 

(.037) 

Attribution of 
responsibility 

  

-.561*** 

(.085) 

-.330*** 

(.070) 

-.411*** 

(.072) 

-.523*** 

(.094) 

-.295*** 

(.077) 

-.402*** 

(.080) 

Incidence of 
COVID-19 (7 days)    

-.001 
(.001) 

.0003 
(.001) 

.0007 
(.0007) 

Subjective social 

class    

.020 

(.062) 

-.018 

(.050) 

-.003 

(.053) 

Education 
    

.032 
(.068) 

.057 
(.056) 

-.011 
(.058) 

Ideology 

    

.048 

(.034) 

.024 

(.028) 

.016 

(.030) 

Interest in politics    
-.076 
(.067) 

.097 o 
(.055) 

.117* 
(.059) 

Trust in regional 

government    

-.125o  

(.069) 

-.089 

(.054) 

-.036 

(.058) 

Contact health 
authority  

suspicion COVID-19     

-.043 

(.197) 

-.044 

(.153) 

.039 

(.159) 

Vulnerable 
employment status    

.131 
(.175) 

-.087 
(.149) 

-.108 
(.156) 

Income level     

-.136o 

(.076) 

.010 

(.057) 

-.129* 

(.062) 

       
Constant 
  

-2.147*** 
(.314) 

-2.227*** 
(.232) 

-2.244*** 
(.250) 

-2.154*** 
(.540) 

-2.71*** 
(.415) 

-2.437*** 
(.441) 

N 4290 4197 4227 3562 3481 3507 

McFadden PseudoR2 .021 .019 .021 .027 .022 .022 

Nagelkerke 

PseudoR2 .026 .027 .029 .034 .031 .031 

Likelihood ratio test 44.39*** 58.95*** 59.99*** 47.39*** 56.62*** 58.51*** 

Notes: Model 1 and 4 refers to the management of the lockdown, while model 2 and 5 refers to the 
nursing homes and model 3 and 6 to healthcare. Standard errors are in parenthesis. (

o
)p<0.1 *p<0.05. 

**p<0.01. ***p<0.001 

 



17    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 

Table 4: Logit models explaining the change of preferences for centralized decision- 

making in the management of the lockdown, nursing homes and healthcare 

 

 

 

On the other hand, having voted for PP or VOX (right-wing opposition parties) is 

associated with a decentralizing preference shift in all three areas, although these factors 

seem to lose significance after controls are incorporated into the models. The odds of 

having a decentralizing preference shift are 45 percent higher for the management of the 

lockdown (odds ratio=1.45, 95% CI=1.12–1.88, p=0.005), 49 percent higher for the 

Table 4. Logit models explaining the change of preferences for decentralized decision-
making in the management of the lockdown, nursing homes and healthcare 

 (1) (2) (3) (4) (5) (6)    
Satisfaction with 

govt. performance 

-.049 

(.043) 

.092 o 

(.049) 

.011 

(.044) 

-.081 o 

(.049) 

.075 

(.056) 

-.024 

(.051) 

Central government 
  

.016 
(.113) 

-.070 
(.122) 

-.118 
(.044) 

-.0002 
(.130) 

.055 
(.142) 

-.015 
(.142) 

Regional government 

  

.036 

(.091) 

.139 

(.098) 

.236* 

(.100) 

-.087 

(.107) 

.026 

(.115) 

.045 

(.117) 

Vote to right-wing 
opposition parties in 

the CG (PP+VOX)  

.371** 

(.134) 

.400** 

(.139) 

.392** 

(.143) 

.142 

(.163) 

.192 

(.171) 

.149 

(.175) 

Degree of 
autonomism 

  

.067* 

(.027) 

.035 

(.029) 

.042 

(.030) 

.052 

(.032) 

.030 

(.035) 

.028 

(.036) 

Attribution of 
responsibility 

  

.027 

(.075) 

.305*** 

(.062) 

.182** 

(.063) 

.024 

(.082) 

.253*** 

(.070) 

.216** 

(.071) 

Incidence of 
COVID-19 (7 days)    

-.0001 
(.001) 

-.00008 
(.0006) 

-.001 
(.001) 

Subjective social 

class    

-.139** 

(.046) 

-.100* 

(.049) 

-.073 

(.051) 

Education 
    

-.066 
(.049) 

.003 
(.053) 

.057 
(.055) 

Ideology 

    

.025 

(.025) 

.056* 

(.027) 

.055* 

(.028) 

Interest in politics    
-.029 
(.050) 

-.023 
(.054) 

-.091 o 
(.055) 

Trust in regional 

government    

.128* 

(.050) 

.073 

(.054) 

.164** 

(.056) 

Contact health 
authority  

suspicion COVID-19     

.079 

(.140) 

.192 

(.147) 

.169 

(.151) 

Vulnerable 
employment status    

.260* 
(.129) 

.257 o 
(.138) 

.216 
(.142) 

Income level     

.107* 

(.054) 

.012 

(.059) 

-.003 

(.059) 

       

Constant 
  

-2.206*** 
(.234) 

-2.450*** 
(.225) 

-2.464*** 
(.236) 

-1.785*** 
(.392) 

-2.463*** 
(.407) 

-2.51*** 
(.423) 

N 4290 4197 4227 3562 3481 3507 

McFadden PseudoR2 .005 .019 .011 .013 .022 .019 

Nagelkerke 

PseudoR2 .007 .026 .015 .019 .030 .027 

Likelihood ratio test 16.03* 58.63*** 32.29*** 37.5** 54.71*** 47.68*** 

Notes: Model 1 and 4 refers to the management of the lockdown, while model 2 and 5 refers to the 
nursing homes and model 3 and 6 to healthcare. Standard errors are in parenthesis. (

o
)p<0.1 *p<0.05. 

**p<0.01. ***p<0.001 

 



18    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 

management of nursing homes (odds ratio=1.49, 95% CI=1.14–1.96, p=0.004) and 48 

percent higher in the management of healthcare (odds ratio=1.48, 95% CI=1.12–1.96, 

p=0.006) if the subjects had voted for any of the main right-wing parties in the 

opposition (PP or VOX). Having voted for one of the governing parties at the regional 

government level increases the probability of having a decentralizing preference shift 

only for the management of healthcare (27 percent higher; odds ratio=1.27, 95% 

CI=1.04–1.54, p=0.018). Therefore, H1c seems to be only partially confirmed. 

H2 is rejected. Satisfaction with government performance to manage the policy areas does 

not have significant explanatory power for intergovernmental preference change. This 

variable is only significant at the 0.1 level to explain a decentralization shift for nursing 

homes and lockdown. For every one-unit increase in government performance 

satisfaction in the management of nursing homes, the odds of having a decentralizing 

preference shift was 10 percent higher (odds ratio=1.1, 95% CI=1.0–1.21, p=0.058), 

while it was 8 percent lower (odds ratio=0.92, 95% CI=0.84–1.01, p=0.094) for the 

management of the lockdown.10  

Individual core beliefs, which are here termed autonomism, do not appear to play a clear 

role in most scenarios. High autonomism increases the likelihood of a decentralizing 

preference shift only for the management of the lockdown (Model 1, Table 4). However, 

the inclusion of controls erases its explanatory power. For every one-unit increase in 

autonomism, the odds of having a decentralizing preference shift for the management of 

the lockdown was 7 percent higher (odds ratio=1.07, 95% CI=1.01–1.13, p=0.013). 

Attribution of responsibility is very useful in determining the likelihood of an 

intergovernmental preference shift in either direction. This variable is statistically 

significant in 10 out of 12 models. The odds of a centralizing preference shift for the 

management of the lockdown fall 43 percent (odds ratio=0.57, 95% CI=0.48–0.67, 

p=0.001) as the attribution of responsibility moves up one level towards central 

government (i.e., it drops 43 percent as the attribution of responsibility moves from 

regional government to shared responsibility or from shared responsibility to central 

government). The likelihood of a centralized preference shift is 28 percent lower for the 

management of nursing homes (odds ratio=0.72, 95% CI=0.63 – 0.82, p=0.001) and 33 

percent lower in the management of healthcare (odds ratio=0.66, 95% CI=0.58–0.76, 

p<0.001). After including controls, the attribution of responsibility continued to be 

relevant in explaining the increase in the centralizing shift of preferences.11  

The attribution of responsibility is also relevant in explaining the likelihood of having a 

decentralizing preference shift for the management of nursing homes and healthcare but 

not the lockdown. As the attribution of responsibility moves up one level in the direction 

of the central government, the likelihood of a decentralizing preference shift for the 

management of nursing homes is 36 percent higher (odds ratio=1.36, 95% CI=1.2–1.53, 

 

10 What if H2 and H4 were merged? The expectation would be to find a higher likelihood of no change 

of preference in individuals satisfied with government performance who attributed responsibility for each 

of the three tasks and vice versa with dissatisfied individuals. However, in Table A6 in Appendix I, no clear 

relationship can be observed between satisfaction with the performance of the government to which 

responsibility is attributed and a preference shift for the management of any of the three areas. The authors 

confirmed this with OLS models. Satisfaction with government performance to which responsibility is 

attributed for the management of the pandemic was only relevant in two out of the 18 models, one of which 

offered an incoherent result and the second lost significance after incorporating controls. 
11 Here are the odds ratios for models 4-6 in Table 3: Model 4 (odds ratio=0.59, 95% CI=0.49–0.71, 

p<0.001); Model 5 (odds ratio=0.74, 95% CI=0.64–0.87, p<0.001); Model 6 (odds ratio=0.67, 95% 

CI=0.57–0.78, p<0.001). 



19    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 

p<0.001) and 20 percent higher for the management of healthcare (odds ratio=1.2, 95% 

CI=1.06–1.36, p=0.004).12 

Interest in politics and income level appear to be relevant in predicting the probability of 

having a centralizing preferences shift but only for the management of nursing homes. 

For every one-unit increase in interest in politics, the odds of having a centralized 

preference shift for the management of nursing homes was 12 percent higher (odds 

ratio=1.12, 95% CI=1.00–1.26, p=0.045). The odds of having a centralized preference 

shift for the management of the lockdown is 12 percent lower (odds ratio=0.88, 95% 

CI=0.78–0.99, p=0.038) for every unit increase in income. 

In contrast, there are control variables which help explain the likelihood of a 

decentralizing preference shift in all three policy areas. The odds of having a 

decentralizing preference shift for the management of the lockdown is 14 percent higher 

(odds ratio=1.14, 95% CI=1.03–1.25, p=0.011) for every one unit increase in trust in 

regional government, 13 percent lower (odds ratio=0.87, 95% CI=0.8–0.95, p=0.002) for 

every level increase in social class, 30 percent higher (odds ratio=1.3, 95% CI=1.01–1.67, 

p=0.047) if the subject is in a situation of employment vulnerability (unemployed or 

covered by a temporary layoff scheme in Spain), and 11 percent higher (odds ratio=1.11, 

95% CI=1.0–1.24, p=0.048) for every unit increase in income. 

For every one unit to the right in the ideology spectrum, the odds of having a 

decentralizing preference shift for the management of nursing homes are 6 percent higher 

(odds ratio=1.06, 95% CI=1.0–1.12, p=0.039), and for every level increase in social class, 

the odds of a decentralized shift are 10 percent lower (odds ratio=0.9, 95% CI=0.82–1.0, 

p=0.042). The odds of a decentralized preference shift for the management of healthcare 

is 18 percent higher (odds ratio=1.18, 95% CI=1.06–1.31, p=0.003) for every unit 

increase in trust in the regional government, and 6 percent higher (odds ratio=1.06, 95% 

CI=1.0–1.12) for every unit to the right in the ideology spectrum. 

These models were replicated for a robustness check but with an ordinary least square 

regression analysis (see Table A.10 in Appendix I). The dependent variable was 

operationalized with a 5-point scale (–2 to +2). Although the dependent variable in the 

robustness check is not continuous, previous literature recognizes that if it is 5 points or 

more (like the ones constructed), statistical treatments such as OLS can be applied. This 

is a common and accepted practice in the field (Gomila 2021; Kromrey and Rendina-

Gobioff 2003). The results were confirmed: attribution of responsibility is the only 

explanatory variable with a statistically significant relationship with the preference shift 

for all models. As the attribution of responsibility moves towards central government, the 

population preferences move towards decentralization and vice versa. The relevance of 

partisanship was also confirmed for some models, with PSOE-UP voters having a higher 

probability of a centralizing shift and VOX-PP voters a decentralizing shift. 

The McFadden and Nagelkerke PseudoR-squared showed low values for all models, 

which many might point out as an indicator of having models with low predictive values, 

low quality, or poor goodness-of-fit. First, low PseudoR-squared in logit models or R-

squared in linear models are expected in behavioural political science, and it was 

particularly expected with the dependent variable due to the near-zero variance. A large 

part of the population (63 percent) did not change in its preference regarding the level of 

government to manage the pandemic in any of the areas examined. However, statistically 

 

12 Here are odds ratios for models with controls: Model 5 (odds ratio=1.29, 95% CI=1.12–1.48, p<0.001) 

and Model 6 (odds ratio=1.24, 95% CI=1.08–1.43, p=0.002). 



20    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 

significant predictors for a near-zero variance-dependent variable were nonetheless 

found. Even with a small effect size, important conclusions can still be drawn with 

statistically significant predictors. As Peng, Lee, and Ingersoll (2002, 6) point out: “a 

researcher can treat these two R2 indices as supplementary to other, more useful 

evaluative indices, such as the overall evaluation of the model, [and] tests of individual 

regression coefficients.” King (1986) also confirmed that R-squared indices can be highly 

misleading and that there are better ways to test the predictive power or goodness-of-fit 

in a model. The significance of the overall model was confirmed using the chi-square test 

in ANOVA and the likelihood ratio test. 

In the results section, the individual regression coefficients that were statistically 

significant are highlighted. The statistical significance of these terms is further confirmed 

through additional analyses (not shown here), using odds ratios to assess the relevance 

and power of the predictors. Moreover, Greenhill, Ward and Sacks (2011, 993) proposed 

the separation plot as a “more nuances and nonscalar, visual yardstick” for assessing 

binary model fit. As Table A.11 in Appendix I shows, none of the 12 models are perfect, 

but they do a reasonably good job of describing the data, given that large parts of the 

events (i.e., centralizing or decentralizing intergovernmental preference shifts) are on the 

right-hand side of the graphs.  

 

6. Discussion and Conclusion 

Based on a large and original national survey of the Spanish population conducted during 

the second wave of the COVID-19 crisis, this article has studied the determinants of the 

shifts in the intergovernmental preferences of Spanish citizens for managing three public 

policy areas related to the pandemic. The article makes three contributions: it adds to the 

literature on how and why the general population change their preferences for 

government-level responsibility, it presents some immediate policy and democracy-

related implications for accountability and blame management in the crisis for advanced 

multilevel democracies, and it contributes to our understanding of the Spanish 

population’s opinion on federal governance. 

According to the hypotheses, a part of the Spanish public shifted its preferences about 

which level of government should be in charge of crisis management as the crisis 

progressed and in particular— in line with the literature about rally-around-the-flag—

support for central government suffered more than support for sub-national government 

after the initial phase of the crisis. The article shows that 37 percent of the general 

population surveyed shifted their government responsibility preferences during the health 

crisis in at least one policy area. More than half of those who changed their preferences 

did so towards decentralization for the three selected policy areas. Therefore, it could be 

argued that the crisis affected citizens’ perceptions of power distribution across three 

policy areas, although the perceptions of the majority did not change within the short 

period analyzed. 

The authors’ hypotheses suggested three main factors for explaining the changes in 

attitudes during crises. Even during a crisis, when citizens have more information from 

the media, are much more sensitive to it, and are probably very interested in the 

effectiveness of policies, they employ partisan cues to help make sense of policies. 

However, Spanish citizens do not appear to be sensitive to the perceived effectiveness of 

these policies during the pandemic. They tend to shift their preferences regarding 

responsibility towards the government they voted for, even in some cases when they think 



21    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 

the government level responsible for the policy is doing well (e.g., those who attributed 

nursing home management to the region and considered it was well managed). Of those 

who did not vote for the PSOE-UP coalition in central government, 9.8 percent shifted 

their preferences towards centralization; however, 14.2 percent of PSOE-UP voters 

reported a centralizing shift. It is possible that this effect may be different depending on 

the different phases of the crisis. Certainly, the partisan effect will be less acute during 

the initial phase, in which the public ‘rallies around the flag’ and the opposition refrains 

from attacking the government outright (Flinders 2020). However, this effect can be 

observed in the second phase, where blame-shifting and blame avoidance reached their 

peak in Spain. 

Interestingly, in Spain, unlike in the US or other federations, right-wing parties at the 

national level are traditionally associated with centralization, while the opposite is true 

for parties on the left. The data in this article demonstrates that the right-wing, especially 

extreme-right voters, shifted their preferences towards decentralization in this crisis, 

while voters from the left supported the central government when their party controlled 

this level. Surprisingly for the Spanish context, the vote for VOX (far-right party) was 

associated with a preference for decentralization. 

The authors’ analysis supports the conclusion that even in crises like the one analyzed in 

this article, a global pandemic not initially attributable to the government of Spain, the 

government initially responsible for policy management is ultimately penalized by its 

citizens. Citizens tend to prefer another level of government after months of failing to 

find a solution to the crisis. Multilevel governments offer this escape route. The results in 

Table A.5 in the appendix clearly support this conclusion. 

Constituents cannot easily change their government outside of electoral periods, but at 

least they can change their preferences about who should be responsible for certain tasks. 

However, the authors’ results did not find that satisfaction with task management, nor 

with the level of government to which responsibility for the task is attributed, is relevant 

in explaining the intergovernmental preference shifts. Among the citizens surveyed, 

satisfaction with government action weakly explains a decentralization preference shift 

for the management of nursing homes and lockdowns. One possible solution for this 

might be that citizens do not properly differentiate between authorities at the different 

levels of government when it comes to the responsibility for handling the crisis, as 

Blackburn et al. (2023) found for Russia. 

Additionally, preferences about federal governance or beliefs about how the state 

territorial arrangements should be structured have shown no significant power in 

predicting the preference shifts in citizens. This is, to a certain extent, surprising for Spain, 

where the territorial organization has been one of the big issues in the public debate since 

the restoration of democracy in the 1970s and one that citizens seem to have elaborate 

views on, particularly after the secessionist tensions in Catalonia since 2017. 

Furthermore, this somehow challenges previous literature on general preferences towards 

decentralization which had found some links between these preferences and identity. It 

will be necessary to explore the extent to which this may relate to the special context in 

which this study was conducted or, on the contrary, it may mark a period of 

transformation in the sophisticated conception that citizens have on the issue of their 

preferred system of territorial organization in Spain. 

This research is relevant for studying other cases. For example, it proposes a way to 

examine if similar shifts in public preferences regarding government responsibilities 

occur in other federal countries with increasing polarization between left and right, 



22    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 

increasing geographical concentration of partisan interests, and in different federations 

with similar degrees of devolution of health responsibilities. It would also be relevant to 

assess broader applications across different types of crises (e.g., previous health crises 

such as the Ebola or Zika virus outbreaks, economic downturns, natural disasters, and 

even future crises). Future research should replicate this analysis using a longitudinal or 

panel design to track changes in public opinion about government responsibilities across 

various stages of a crisis and beyond. 

Among the restrictions of this study, one could highlight the usual limitations in natural 

experiments regarding external validity. Future research should be done to confirm the 

validity of these results in other waves of the pandemic, other external shocks, or at post-

pandemic times. A second limitation concerns the non-probability sample design inherent 

in opt-in panel surveys. Quotas were used to smooth this limitation, although it should be 

recalled that the lack of representativeness is not a problem because it is not the intention 

to explain the distribution of preferences. Thirdly, although the methodology used 

retrospective survey questions to assess changes in preferences, we acknowledge the 

inherent limitations associated with recall bias and the potential conflation of past and 

present attitudes. Nonetheless, supported by the literature mentioned in the methodology 

section, it has been argued that the use of recall data provides valuable insights, 

particularly in dynamically evolving situations like the COVID-19 pandemic. 

 

  



23    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 

REFERENCES 

Albertson, Bethany, and Shana Kushner Gadarian. 2015. Anxious Politics. Democratic 

Citizenship in a Threatening World. Cambridge, UK: Cambridge University Press. 

Altiparmakis, Argyrios, Bojar, Abel, Brouard, Sylvain, Foucault, Martial, Kriesi, 

Hanspeter, and Richard Nadeau. 2021. “Pandemic Politics: Policy Evaluations of 

Government Responses to COVID-19.” West European Politics 44 (5–6): 1159–

1179. https://doi.org/10.1080/01402382.2021.1930754. 

Arceneaux, Kevin, and Robert M. Stein. 2006. “Who Is Held Responsible When Disaster 

Strikes: The Attribution of Responsibility for a Natural Disaster in an Urban 

Election.” Journal of Urban Affairs 28 (1): 42–53. https://doi.org/10.1111/j.0735-

2166.2006.00258.x. 

Arceneaux, Kevin, and Ryan J. Vander Wielen. 2017. Taming Intuition: How Reflection 

Minimizes Partisan Reasoning and Promotes Democratic Accountability. New 

York: Cambridge University Press. 

Atkeson, Lonna Rae, and Cherie D. Maestas. 2012. Catastrophic Politics: How 

Extraordinary Events Redefine Perceptions of Government. Cambridge, UK: 

Cambridge University Press. 

Banzhaf, H. Spencer, and Randall P. Walsh. 2008. “Do People Vote with Their Feet? An 

Empirical Test of Tiebout.” American Economic Review, 98 (3): 843–63. 

https://doi.org/10.1257/aer.98.3.843. 

Barbet Porta, Berta. 2022. Gestió de la COVID-19 i preferències de (des)centralització a 

Catalunya [COVID-19 management and (de)centralisation preferences in 

Catalonia]. Barcelona: Generalitat de Catalunya, Institut d'Estudis de 

l'Autogovern. 

Biddle, Nicholas, Gray, Matthew, and Ian McAllister. 2024. “Federalism and Confidence 

in Australian Governments During the COVID-19 Pandemic.” Publius: The 

Journal of Federalism 54 (2): 257–282. https://doi.org/10.1093/publius/pjad032.  

Bjørnskov, Christian, and Voigt, Stefan. 2022. “Emergencies: On the Misuse of 

Government Powers.” Public Choice 190: 1–32. https://doi.org/10.1007/s11127-

021-00918-6. 

Blackburn, Matthew, Hutcheson, Derek, Tsumarova, Elena, and Bo Petersson. 2023. 

“Covid-19 and the Russian Regional Response: Blame Diffusion and Attitudes to 

Pandemic Governance.” Canadian Journal of European and Russian Studies 16 

(1): 29–54. https://doi.org/10.22215/cjers.v16i1.3955. 

Boin, Arjen. 2024. “Pandemic Challenges for Public Managers: Juggling Parallel Crisis 

Playbooks.” In Research Handbook on Public Management and COVID-19, edited 

by Helen Dickinson, Sophie Yates, Janine O’Flynn, and Catherine Smith, 19–30. 

Cheltenham, UK: Edward Elgar Publishing. 

Boin, Arjen, and Paul ‘t Hart. 2022. “From Crisis to Reform? Exploring Three Post-Covid 

Pathways.” Policy and Society 41 (1): 13–24. https://doi.org/10.1093/

polsoc/puab007. 

Boin, Arjen, ‘t Hart, Paul, and Allan McConnell. 2009. “Crisis Exploitation: Political and 

Policy Impacts of Framing Contests.” Journal of European Public Policy 16 (1): 

81–106. https://doi.org/10.1080/13501760802453221. 

https://doi.org/10.1093/publius/pjad032
https://doi.org/10.1093/polsoc/puab007
https://doi.org/10.1093/polsoc/puab007


24    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 

Boin, Arjen, McConnell, Allan, and Paul ‘t Hart. 2021. Governing the Pandemic: A Crisis 

Management Perspective. London: Palgrave Macmillan. 

Boin, Arjen, and Mark Rhinard. 2023. “Crisis Management Performance and the 

European Union: The Case of COVID-19.” Journal of European Public Policy 30 

(4): 655–675. https://doi.org/10.1080/13501763.2022.2141304. 

Brown, Adam R. 2010. “Are Governors Responsible for the State Economy? 

Partisanship, Blame, and Divided Federalism.” The Journal of Politics, 72 (3): 

605–615. https://doi.org/10.1017/S0022381610000046. 

Bullock, John G. 2011. “Elite Influence on Public Opinion in an Informed Electorate.” 

American Political Science Review 105 (3): 496–515.  https://doi.org/

10.1017/S0003055411000165. 

Centro de Investigaciones Sociológicas (CIS) 2020a. “Estudio 3279. Barómetro de abril” 

[Study 3279. April Barometer]. Accessed October 1, 2020. https://www.cis.es/-

/avance-de-resultados-del-estudio-3279-barometro-de-abril-2020-. 

———. 2020b. “Estudio 3281. Barómetro de mayo” [Study 3281. May Barometer]. 

Accessed October 27, 2020. https://www.cis.es/-/avance-de-resultados-del-

estudio-3279-barometro-de-abril-2020-  

———. “Estudio 3292. Barómetro de septiembre” [Study 3292. September Barometer]. 

Accessed October 19, 2020. https://www.cis.es/-/avance-de-resultados-del-

estudio-3279-barometro-de-abril-2020-  

Ciuk, David J., and Berwood A. Yost. 2016. “The Effects of Issue Salience, Elite 

Influence, and Policy Content on Public Opinion.” Political Communication 33 

(2): 328–45. https://doi.org/10.1080/10584609.2015.1017629. 

Colino, César. 2020. “Decentralisation in Spain: Federal Evolution and Performance of 

the Estado Autonómico.” In Oxford Handbook of Spanish Politics, edited by Diego 

Muro and Ignacio Lago, 62–81. Oxford: Oxford University Press. 

Congleton, Roger D. 2021. “Federalism and Pandemic Policies: Variety as the Spice of 

Life.” Public Choice 195: 73–100. https://doi.org/10.1007/s11127-021-00915-9.  

Connolly, Jennifer, Klofstad, Casey, Uscinski, Joseph, and Jonathan West. 2020. “Public 

Preferences for Zika Policy and Responsibility in the Absence of Partisan Cues.” 

Journal of Public Policy 40 (3): 402–427. https://doi.org/10.1017/

S0143814X19000011. 

Dalton, Russell J. 2018. “Realignment and Beyond.” In Political Realignment: 

Economics, Culture, and Electoral Change,” 215–236. Oxford: Oxford University 

Press. https://doi.org/10.1093/oso/9780198830986.003.0010 

Del Pino, Eloísa, Moreno Fuentes, Francisco Javier, Cruz-Martínez, Gibrán, Hernández-

Moreno, Jorge, Moreno, Luis, Pereira-Puga, Manuel, and Roberta Perna. 2021. 

“Governmental Response to the COVID-19 Pandemic in Long-Term Care 

Residences for Older People: Preparedness, Responses and Challenges for the 

Future.” MC COVID-19 Working Paper No. 13. Madrid: Institute of Public Goods 

and Policies (IPP-CSIC). http://dx.doi.org/10.20350/digitalCSIC/13688. 

Del Pino, Eloísa, and Gregg van Ryzin. 2013. “Intergovernmental Responsibility for 

Social Policy: An Analysis of Public Preferences in Spain.” Publius: The Journal 

of Federalism 43 (4), 676–700. https://doi.org/10.1093/publius/pjs051. 

https://doi.org/10.1007/s11127-021-00915-9
https://doi.org/10.1093/publius/pjs051


25    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 

Del Pino, Eloísa, and César Colino. 2024. “Subnational Governance in Spain.” In 

Handbook on Subnational Governments and Governance, edited by Claudia 

Avellaneda and Ricardo A. Bello-Gómez, 375–392. Cheltenham, UK: Edward 

Elgar Publishing. 

Dinan, John, and Jac C. Heckelman. 2020. “Stability and Contingency in Federalism 

Preferences.” Public Administration Review 80 (2): 234–243. https://doi.org/

10.1111/puar.13157. 

Druckman, James N., Erik Peterson, and Rune Slothuus. 2013. “How Elite Partisan 

Polarization Affects Public Opinion Formation.” American Political Science 

Review 107 (1): 57–79. https://doi.org/10.1017/S0003055412000500. 

Druckman, James N., and Arthur Lupia. 2016. “Preference Change in Competitive 

Political Environments.” Annual Review of Political Science 19: 13–31. 

https://doi.org/10.1146/annurev-polisci-020614-095051. 

Eckhard, Steffen, and Alexa Lenz. 2020. Die öffentliche Wahrnehmung des 

Krisenmanagements in der Covid-19 Pandemie [Public perception of crisis 

management in the Covid-19 pandemic]. Working Paper 05, Konstanzer Online-

Publikations-System (KOPS). https://kops.uni-konstanz.de/server/api/core/

bitstreams/dbcf1459-5c94-4e21-9acc-28c943f6aeb6/content. 

Flinders, Matthew. 2020. “Democracy and the Politics of Coronavirus: Trust, Blame and 

Understanding.” Parliamentary Affairs 74 (2): 483-502. https://doi.org/

10.1093/pa/gsaa013. 

Glaser, James M., Berry, Jeffrey M., and Deborah J. Schildkraut. 2023. “Ideology and 

Support for Federalism in Theory—And in Practice.” Publius: The Journal of 

Federalism 53 (4): 511–535. https://doi.org/10.1093/publius/pjad003. 

Goidel, Kirby, Callaghan, Timothy, Nuzhath, Tasmiah, Scobee, Julia, Washburn, David, 

and Matthew Motta. 2024. “Professional Partisans? Primary Care Physicians, State 

Governments, and COVID-19 Responsibility and Response.” State Politics & 

Policy Quarterly 24 (2): 145–66. https://doi.org/10.1017/spq.2023.27. 

Gomila, Robin. 2021. “Logistic or Linear? Estimating Causal Effects of Experimental 

Treatments on Binary Outcomes Using Regression Analysis.” Journal of 

Experimental Psychology: General 150 (4): 700–709. 

https://doi.org/10.1037/xge0000920. 

Greenhill, Brian, Michael D. Ward, and Audrey Sacks. 2011. “The Separation Plot: A 

New Visual Method for Evaluating the Fit of Binary Models.” American Journal 

of Political Science 55 (4): 991–1002. https://doi.org/10.1111/j.1540-

5907.2011.00525.x. 

Groenendyk, Eric W. 2013. Competing Motives in the Partisan Mind: How Loyalty and 

Responsiveness Shape Party Identification and Democracy. Oxford University 

Press. https://doi.org/10.1093/acprof:oso/9780199969807.001.0001. 

Grofman, Bernard, and Barbara Norrander. 1990. “Efficient Use of Reference Group 

Cues in a Single Dimension.” Public Choice 64: 213–227. 

https://doi.org/10.1007/BF00124367. 

Heinkelmann-Wild, Tim, Kriegmair, Lisa, and Berthold Rittberger. 2020. “The EU 

Multi-Level System and the Europeanization of Domestic Blame Games.” Politics 

and Governance 8 (1): 85–94. https://doi.org/10.17645/pag.v8i1.2522. 

https://doi.org/10.1017/S0003055412000500
https://doi.org/10.1093/pa/gsaa013
https://doi.org/10.1093/pa/gsaa013
https://doi.org/10.1093/publius/pjad003
https://doi.org/10.1017/spq.2023.27
https://psycnet.apa.org/doi/10.1037/xge0000920
https://doi.org/10.1111/j.1540-5907.2011.00525.x
https://doi.org/10.1111/j.1540-5907.2011.00525.x
https://psycnet.apa.org/doi/10.1093/acprof:oso/9780199969807.001.0001
https://doi.org/10.17645/pag.v8i1.2522


26    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 

Hernández-Moreno, Jorge, Pereira-Puga, Manuel, and Gibrán Cruz-Martínez. 2023. 

“COVID-19 and Nursing Homes in Decentralised Spain: How Governance and 

Public Policies Influence Working Conditions and Organisational Responses to the 

Crisis.” Administration & Society 55 (9): 1651–1679. 

https://doi.org/10.1177/00953997231187075. 

Hernández-Moreno, Jorge, and Jean-Baptiste Harguindéguy. 2024. “COVID-19 and 

territorial politics: towards a redefinition of Spanish intergovernmental relations?” 

Regional Studies, Regional Science 11 (1), 387–405. https://doi.org/

10.1080/21681376.2024.2370420. 

Hipp, Lena, Bünning, Mareik, Munnes, Stefan, and Armin Sauermann. 2020. “Problems 

and pitfalls of retrospective survey questions in COVID-19 studies.” Survey 

Research Methods, 14 (2): 109–114. doi:10.18148/srm/2020.v14i2.7741. 

Jacobs, Nicholas. 2017. “An Experimental Test of How Americans Think about 

Federalism.” Publius: The Journal of Federalism 47 (4): 572–598. 

https://doi.org/10.1093/publius/pjx018. 

———. 2021. “Federalism, Polarization, and Policy Responsibility during COVID-19: 

Experimental and Observational Evidence from the United States.” Publius: The 

Journal of Federalism 51 (4): 693–719. https://doi.org/10.1093/publius/pjab014. 

James, Oliver, and Gregg. G. Van Ryzin. 2017. “Motivated Reasoning about Public 

Performance: An Experimental Study of How Citizens Judge the Affordable Care 

Act.” Journal of Public Administration Research and Theory 27 (1):197–209. 

https://doi.org/10.1093/jopart/muw049. 

Jaspers, Eva, Lubbers, Marcel, and Nan D. De Graaf. 2009. “Measuring Once Twice: An 

Evaluation of Recalling Attitudes in Survey Research.” European Sociological 

Review 25 (3): 287–301. https://doi.org/10.1093/esr/jcn048. 

Jedwab, Jack, and John Kincaid. 2018. Identities, Trust, and Cohesion in Federal 

Systems: Public Perspectives. Kingston, ON: McGill-Queen’s University Press.  

Johansson, Bengt, Hopmann, David Nicolas, and Adam Shehata. 2021. “When the rally-

around-the-flag effect disappears, or: when the COVID-19 pandemic becomes 

‘normalized.’” Journal of Elections, Public Opinion and Parties 31 (1): 321–334. 

https://doi.org/10.1080/17457289.2021.1924742. 

Juhl, Sebastian, Lehrer, Roni, Blom, Annelies G., Wenz, Alexander, Rettig, Tobias, 

Krieger, Ulrich, Fikel, Marina, Cornesse, Carina, Naumann, Elias, Möhring, Katja, 

and Maximiliane Reifenscheid. 2022. “Preferences for Centralized Decision-

Making in Times of Crisis: the COVID-19 Pandemic in Germany.” In Jahrbuch 

für Handlungs- und Entscheidungstheorie, edited by Jan Sauermann, Markus 

Tepe, and Marc Debus, 1–27. Wiesbaden: Springer Fachmedien Wiesbaden. 

Kam, Cindy, and Robert A. Mikos. 2007. “Do Citizens Care about Federalism? An 

Experimental Test.” Journal of Empirical Legal Studies 4 (3): 589–624. 

https://doi.org/10.1111/j.1740-1461.2007.00100.x. 

Kennedy, John, Sayers, Anthony, and Christopher Alcantara. 2022. “Does Federalism 

Prevent Democratic Accountability? Assigning Responsibility for Rates of 

COVID-19 Testing.” Political Studies Review 20 (1): 158–165.  

https://doi.org/10.1177/14789299211001690. 

https://doi.org/10.1177/00953997231187075
https://doi.org/10.18148/srm/2020.v14i2.7741
https://doi.org/10.1093/publius/pjx018
https://doi.org/10.1093/publius/pjab014
https://doi.org/10.1093/jopart/muw049
https://doi.org/10.1093/esr/jcn048


27    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 

Kincaid, John, and J. Wesley Leckrone. 2021. “Partisan Fractures in U.S. Federalism’s 

COVID-19 Policy Responses.” State and Local Government Review 52 (4): 28–

308. https://doi.org/10.1177/0160323X20986842. 

King, Gary. 1986. “How Not to Lie with Statistics: Avoiding Common Mistakes in 

Quantitative Political Science.” American Journal of Political Science 30 (3): 666–

687. https://doi.org/10.2307/2111095.  

Kritzinger, Sylvia, Foucault, Martial, Lachat, Romain, Partheymüller, Julia, Plescia, 

Carolina, and Sylvain Brouard. 2021. “‘Rally Round the Flag’: The COVID-19 

Crisis and Trust in the National Government.” West European Politics 44 (5–6): 

1205-1231. https://doi.org/10.1080/01402382.2021.1925017. 

Kromrey, Jeffrey D., and Gianna Rendina-Gobioff. 2003. “An Empirical Comparison of 

Regression Analysis Strategies with Discrete Ordinal Variables.” Multiple Linear 

Regression Viewpoints 28 (2): 30–43. Accessed September 30, 2024. 

https://www.glmj.org/archives/MLRV_2003_29_1.pdf#page=32. 

Konisky, David. 2011. “Public Preferences for Environmental Policy Responsibility.” 

Publius: The Journal of Federalism 41 (1): 76–100. https://doi.org/

10.1093/publius/pjp044. 

Kunda, Ziva. 1990. “The Case for Motivated Reasoning.” Psychological Bulletin 108 (3): 

480–498. https://doi.org/10.1037/0033-2909.108.3.480. 

León, Sandra, and Amuitz Garmendia-Madariaga. 2020. “Popular Reactions to External 

Threats in Federations.” Working Paper (Version 4). Universidad Carlos III de 

Madrid. Accessed September 30, 2024. https://doi.org/10.31235/osf.io/qyjtm. 

León, Sandra, Jurado, Ignacio, and Amuitz Garmendia-Madariaga. 2018. “Passing the 

Buck? Responsibility Attribution and Cognitive Bias in Multilevel Democracies.” 

West European Politics 41(3): 660–682. https://doi.org/10.1080/

01402382.2017.1405325. 

Lobera, Josep, Santana, Andrés, and Catherine Gross. 2024. “Are We Looking at Crises 

Through Polarized Lenses? Predicting Public Assessments of the Official Early 

Responses to the COVID-19 Pandemic in Eight Countries.” European 

Sociological Review 40 (1): 1–13. https://doi.org/10.1093/esr/jcad016. 

Mattei, Paola, and Eloísa del Pino. 2021. “Coordination and Health Policy Responses to 

the First Wave of COVID-19 in Italy and Spain.”  Journal of Comparative Policy 

Analysis: Research and Practice 23 (2): 274–281. https://doi.org/

10.1080/13876988.2021.1878886. 

Milosh, Maria, Painter, Marcus, Van Dijcke, David, and Austin L. Wright. 2020. 

“Unmasking Partisanship: How Polarization Influences Public Responses to 

Collective Risk.” University of Chicago, Becker Friedman Institute for Economics 

Working Paper No. 2020-102. https://repec.bfi.uchicago.edu/RePEc/

pdfs/BFI_WP_2020102.pdf. 

Mullinix, Kevin J. 2016. “Partisanship and Preference Formation: Competing 

Motivations, Elite Polarization, and Issue Importance.” Political Behavior 38 (2): 

383–411. https://doi.org/10.1007/s11109-015-9318-4. 

 

 

https://doi.org/10.1177/0160323X20986842
https://doi.org/10.1080/01402382.2021.1925017
https://www.glmj.org/archives/MLRV_2003_29_1.pdf#page=32
https://doi.org/10.31235/osf.io/qyjtm
https://doi.org/10.1093/esr/jcad016
https://repec.bfi.uchicago.edu/RePEc/pdfs/BFI_WP_2020102.pdf
https://repec.bfi.uchicago.edu/RePEc/pdfs/BFI_WP_2020102.pdf


28    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 

Navarro, Carmen, and Francisco Velasco. 2022. “From Centralisation to New Ways of 

Multi-Level Coordination: Spain’s Intergovernmental Response to the COVID-19 

Pandemic.” Local Government Studies 48 (2): 191–210. 

https://doi.org/10.1080/03003930.2022.2042683. 

Pears, Emily, and Emily Sydnor. 2022. “COVID-19 and the Culture of American 

Federalism.” RSF: The Russell Sage Foundation Journal of the Social Sciences 8 

(8): 181–220. https://doi.org/10.7758/RSF.2022.8.8.09. 

Peng, Chao-Ying Joanne, Lee, Kuk Lida, and Gary M. Ingersoll. 2002. “An Introduction 

to Logistic Regression Analysis and Reporting.” The Journal of Educational 

Research 96 (1): 3–14. https://doi.org/10.1080/00220670209598786. 

Pereira-Puga, Manuel, Hernández-Moreno, Jorge, and Gibrán Cruz-Martínez. 2023. “La 

coordinación institucional durante la pandemia de COVID-19: el caso de las 

residencias de mayores en España. [Institutional coordination during the COVID-

19 pandemic: the case of nursing homes in Spain].” Política y Sociedad 60 (2): 1–

20. https://doi.org/10.5209/poso.84774. 

Porumbescu, Gregory, Donald Moynihan, Jason Anastasopoulos, and Asmus Leth 

Olsen. 2022. “When blame avoidance backfires: Responses to performance 

framing and outgroup scapegoating during the COVID-19 

pandemic.” Governance 36(3): 779-803. https://doi.org/10.1111/gove.12701. 

Rendleman, Hunter, and Jon C. Rogowski. 2024. “Americans’ Attitudes Toward 

Federalism.” Political Behavior 46: 111–134. https://doi.org/10.1007/s11109-022-

09820-3. 

Revilla, Melanie. 2017. “Analyzing Survey Characteristics, Participation, and Evaluation 

Across 186 Surveys in an Online Opt-in Panel in Spain.” Methods, data, analyses 

11 (2): 135–162. https://doi.org/10.12758/mda.2017.02. 

Rosen, Julia. 2021. “Pandemic Upheaval Offers a Huge Natural Experiment.” Nature 

596: 149–151. https://doi.org/10.1038/d41586-021-02092-7. 

Rodriguez, Cristian, Gadarian, Shana Kushner, Goodman, Sara Wallace, and Thomas B. 

Pepinsky. 2020. “Morbid Polarization: Exposure to COVID-19 and Partisan 

Disagreement about Pandemic Response.” PsyArXiv. Accessed October 1, 2020. 

https://psyarxiv.com/wvyr7/.  

Schildkraut, Deborah J., Berry, Jeffrey M., and James M. Glaser. 2020. “Public Opinion 

about Federalism during the Coronavirus Pandemic.” Data for Progress, May 4, 

2020. Accessed October 1, 2020. https://www.dataforprogress.org/blog/5/4/

public-opinion-federalism.  

Schneider, Saundra K., and Jacoby, William G. 2013. “Intuitive Federalism' and Public 

Opinion Toward Government.” APSA 2013 Annual Meeting Paper, American 

Political Science Association 2013 Annual Meeting. https://ssrn.com/

abstract=2300590.  

Schneider, Saundra K., Jacoby, William G., and Daniel C. Lewis. 2011. “Public Opinion 

Toward Intergovernmental Policy Responsibilities.” Publius: The Journal of 

Federalism 41(1): 1–30. https://doi.org/10.1093/publius/pjq036. 

 

 

https://doi.org/10.1111/gove.12701
https://doi.org/10.12758/mda.2017.02
https://psyarxiv.com/wvyr7/
https://www.dataforprogress.org/blog/5/4/public-opinion-federalism
https://www.dataforprogress.org/blog/5/4/public-opinion-federalism
https://ssrn.com/abstract=2300590
https://ssrn.com/abstract=2300590
https://doi.org/10.1093/publius/pjq036


29    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 

Skarżyńska, Krystyna, Urbańska, Beata, and Piotr Radkiewicz. 2021. “Under or Out of 

Government Control? The Effects of Individual Mental Health and Political Views 

on the Attribution of Responsibility for COVID-19 Incidence Rates.” Social 

Psychological Bulletin 16 (1): 1–21. https://doi.org/10.32872/spb.4395. 

Slothuus, Rune, and Martin Bisgaard. 2020. “How Political Parties Shape Public Opinion 

in the Real World.” American Journal of Political Science 65 (4): 896–911. 

https://doi.org/10.1111/ajps.12550. 

Sniderman, Paul M., Brody, Richard A., and Philip E. Tetlock. 1991. Reasoning and 

Choice: Explorations in Political Psychology. New York: Cambridge University 

Press. 

Snow, Dan, and Geoffrey Evans. 2024. “Partisanship in a Pandemic: Biased Voter 

Assessments of Past and Present Government Performance.” Political Behavior. 

https://doi.org/10.1007/s11109-024-09929-7. 

Thompson, Lyke, and Richard Elling. 1999. “Let Them Eat Marblecake: The Preferences 

of Michigan Citizens for Devolution and Intergovernmental Service-Provision.” 

Publius: The Journal of Federalism 29 (1): 139–154. https://doi.org/10.1093/

oxfordjournals.pubjof.a030005. 

Torcal, Mariano, and Fabiola Mota. 2013. “The Role of Political Parties in Shaping 

Citizens' Political Preferences for the Territorial Organization of the State: The 

Spanish Case.” European Political Science Review 6 (3): 477–502. 

https://doi.org/10.1017/S1755773913000210. 

VanDusky-Allen, Julie A., Utych, Stephen M., and Michael Catalano. 2022. 

“Partisanship, Policy, and Americans’ Evaluations of State-Level COVID-19 

Policies Prior to the 2020 Election.” Political Research Quarterly 75 (2): 479–496. 

https://doi.org/10.1177/10659129211056374. 

Vicentini, Giulia, and Maria Tullia Galanti. 2021. “Italy, the Sick Man of Europe: Policy 

Response, Experts and Public Opinion in the First Phase of Covid-19.” South 

European Society and Politics: 459–485. https://doi.org/

10.1080/13608746.2021.1940582. 

Viskupic, Filip, and David L. Wiltse. 2023. “Political Partisanship and Trust in 

Government Predict Popular Support for COVID-19 Vaccine Mandates for 

Various Professions and Demographic Groups: A Research Note.” American 

Politics Research 51 (2): 139–146. https://doi.org/10.1177/1532673X221118888. 

White, Trenton M., Cash-Gibson, Lucinda, Martin-Moreno, Jose M, Matesanz, Rafael, 

Crespo, Javier, Alfonso-Sanchez, Jose L., Villapol, Sonia, El-Mohandes, Ayman, 

and Jeffrey V. Lazarus. 2021. “COVID-SCORE Spain: Public Perceptions of Key 

Government COVID-19 Control Measures.” European Journal of Public Health 

31: 1095–1102.  https://doi.org/10.1093/eurpub/ckab066. 

Wolak, Jennifer. 2016. “Core Values and Partisan Thinking about Devolution.” Publius: 

The Journal of Federalism 46 (4): 463–85.  https://doi.org/

10.1093/publius/pjw021. 

Wolak, Jennifer, and Christine Kelleher Palus. 2010. “The Dynamics of Public 

Confidence in U.S. State and Local Government.” State Politics & Policy 

Quarterly 10 (4): 421–445. https://doi.org/10.1177/153244001001000407. 

  

https://doi.org/10.32872/spb.4395
https://doi.org/10.1093/oxfordjournals.pubjof.a030005
https://doi.org/10.1093/oxfordjournals.pubjof.a030005
https://doi.org/10.1017/S1755773913000210
https://doi.org/10.1080/13608746.2021.1940582
https://doi.org/10.1080/13608746.2021.1940582
https://doi.org/10.1177/1532673X221118888
https://doi.org/10.1093/eurpub/ckab066
https://doi.org/10.1093/publius/pjw021
https://doi.org/10.1093/publius/pjw021


30    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 

Acknowledgments 

 

The authors would like to thank the participants of the 79th Annual Midwest Political 

Science Conference and the 2021 Winter Seminar at the Complutense University of 

Madrid, especially Inés Calzada, who provided much-appreciated comments on an 

earlier draft of this manuscript. The authors received financial support for the research 

of this article from the Spanish Ministry of Economy, Industry and Competitiveness 

(GOWPER-CSO2017-85598) and the Spanish National Research Council (Mc-

COVID19 Coordination mechanisms in Coronavirus management between different 

levels of government and public policy sectors in 15 European countries). 

Shortcomings, of course, remain our responsibility. 

 

Declarations 

 

This work was supported by the Spanish Ministry of Economy, Industry and 

Competitiveness (GOWPER-CSO2017-85598) and the Spanish National Research 

Council (Mc-COVID19 Coordination mechanisms in Coronavirus management 

between different levels of government and public policy sectors in 15 European 

countries). The authors have no competing interests to declare that are relevant to the 

content of this article. Authors are ordered alphabetically (surnames).  

  



31    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 

Appendix I 

Supplementary Material 

 

Table A.1: Partisanship at the central government and the intergovernmental shift of 

preferences  

  

Lockdown Management     
Decentralization 

shift 

No 

change 

Centralization 

shift  

PP (right-wing, national party) 16.47% 78.10% 5.43% 

PSOE (left-wing, national party) 13.27% 79.57% 7.15% 

Unidas Podemos (left-wing, national party) 12.32% 80.47% 7.21% 

VOX (right-wing, national party) 16.77% 76.89% 6.34% 

Ciudadanos (right-wing, national party) 13.71% 79.91% 6.38% 

Esquerra Republicana de Catalunya (left-wing, 

regional party) 

15.20% 80.12% 4.68% 

Junts (right-wing, regional party) 17.86% 78.57% 3.57% 

EH Bildu (left-wing, regional party) 13.04% 80.87% 6.09% 

PACMA (animalist party, national party) 8.20% 77.05% 14.75% 

Más País (left-wing, national party) 4.88% 85.37% 9.76% 

EAJ-PNV (right-wing, regional party) 15.89% 76.64% 7.48% 

CUP (left-wing, regional party) 2.94% 97.06% 0.00% 

MésCompromís (left-wing, regional party) 9.68% 80.65% 9.68% 

CCa-NC (right-wing regional party) 0.00% 83.33% 16.67% 

Na+ (UPN) (right-wing, regional party) 36.36% 59.09% 4.55% 

BNG (left-wing, regional party) 15.22% 71.74% 13.04% 

PRC (left-wing, regional party) 7.50% 90.00% 2.50% 

Teruel Existe (regionalist party) 28.57% 64.29% 7.14% 

    

Nursing Homes Management     
Decentralization 

shift 

No 

change 

Centralization 

shift  

PP (right-wing, national party) 17.31% 74.13% 8.57% 

PSOE (left-wing, national party) 10.95% 74.52% 14.53% 

Unidas Podemos (left-wing, national party) 9.02% 76.39% 14.59% 

VOX (right-wing, national party) 16.18% 75.93% 7.88% 

Ciudadanos (right-wing, national party) 10.64% 78.25% 11.11% 

Esquerra Republicana de Catalunya (left-wing, 

regional party) 12.21% 80.81% 6.98% 

Junts (right-wing, regional party) 16.36% 80.00% 3.64% 

EH Bildu (left-wing, regional party) 5.22% 85.22% 9.57% 

PACMA (animalist party, national party) 13.45% 73.11% 13.45% 

Más País (left-wing, national party) 10.00% 80.00% 10.00% 

EAJ-PNV (right-wing, regional party) 12.96% 78.70% 8.33% 

CUP (left-wing, regional party) 11.76% 85.29% 2.94% 

MésCompromís (left-wing, regional party) 9.68% 74.19% 16.13% 

CCa-NC (right-wing regional party) 0.00% 100.00% 0.00% 

Na+ (UPN) (right-wing, regional party) 17.39% 65.22% 17.39% 

BNG (left-wing, regional party) 13.33% 82.22% 4.44% 

PRC (left-wing, regional party) 9.76% 80.49% 9.76% 

Teruel Existe (regionalist party) 35.71% 50.00% 14.29% 

    

Healthcate Management    



32    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

  
Decentralization 

shift 

No 

change 

Centralization 

shift  

PP (right-wing, national party) 14.98% 76.13% 8.89% 

PSOE (left-wing, national party) 11.22% 76.82% 11.96% 

Unidas Podemos (left-wing, national party) 7.68% 79.51% 12.81% 

VOX (right-wing, national party) 14.46% 77.69% 7.85% 

Ciudadanos (right-wing, national party) 11.16% 79.81% 9.03% 

Esquerra Republicana de Catalunya (left-wing, 

regional party) 

12.28% 83.04% 4.68% 

Junts (right-wing, regional party) 14.55% 83.64% 1.82% 

EH Bildu (left-wing, regional party) 3.54% 90.27% 6.19% 

PACMA (animalist party, national party) 9.02% 79.51% 11.48% 

Más País (left-wing, national party) 10.00% 75.00% 15.00% 

EAJ-PNV (right-wing, regional party) 12.84% 78.90% 8.26% 

CUP (left-wing, regional party) 5.88% 88.24% 5.88% 

MésCompromís (left-wing, regional party) 3.23% 87.10% 9.68% 

CCa-NC (right-wing regional party) 0.00% 100.00% 0.00% 

Na+ (UPN) (right-wing, regional party) 26.09% 65.22% 8.70% 

BNG (left-wing, regional party) 15.22% 73.91% 10.87% 

PRC (left-wing, regional party) 9.76% 85.37% 4.88% 

Teruel Existe (regionalist party) 28.57% 57.14% 14.29% 

 

Table A.2: Partisanship at the regional government and the intergovernmental shift of 

preferences  

Lockdown Management     
Decentralization 

shift 

No 

change 

Centralization 

shift  

Did not voted for the political party ruling their 

regional government 13.67% 79.13% 7.20% 

Voted for the political party ruling their regional 

government 14.39% 79.39% 6.22% 

    

Nursing Homes Management     
Decentralization 

shift 

No 

change 

Centralization 

shift  

Did not voted for the political party ruling their 

regional government 11.97% 77.52% 10.51% 

Voted for the political party ruling their regional 

government 12.90% 75.31% 11.79% 

    

Healthcate Management     
Decentralization 

shift 

No 

change 

Centralization 

shift  

Did not voted for the political party ruling their 

regional government 11.09% 79.65% 9.26% 

Voted for the political party ruling their regional 

government 12.05% 77.44% 10.52% 

 
 

Table A.3: Satisfaction with government performance and the intergovernmental shift of preferences 

Lockdown Management     
Decentralization shift No change Centralization shift  

Very Bad 13.71% 81.09% 5.20% 

Rather Bad 14.73% 78.48% 6.79% 

Regular 15.58% 76.73% 7.69% 

Rather Good 12.97% 80.14% 6.89% 

Very Good 11.70% 81.30% 7.00% 



33    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 
    

Nursing Homes Management     
Decentralization shift No change Centralization shift  

Very Bad 12.07% 77.38% 10.56% 

Rather Bad 12.24% 77.08% 10.68% 

Regular 12.10% 76.54% 11.36% 

Rather Good 13.35% 73.59% 13.06% 

Very Good 15.74% 71.30% 12.96% 

    

Healthcate Management     
Decentralization shift No change Centralization shift  

Very Bad 11.85% 79.55% 8.59% 

Rather Bad 11.13% 79.46% 9.42% 

Regular 11.47% 78.00% 10.53% 

Rather Good 10.98% 79.99% 9.04% 

Very Good 11.90% 77.97% 10.13% 

 

 

Table A.4: Autonomic culture and the intergovernmental shift of preferences 

Lockdown Management     
Decentralization shift No change Centralization shift  

1.0-1.9 8.85% 86.73% 4.42% 

2.0-2.9 9.89% 86.08% 4.03% 

3.0-3.9 12.96% 81.86% 5.18% 

4.0-4.9 13.94% 78.75% 7.32% 

5.0-5.9 14.85% 78.26% 6.89% 

6.0-6.9 13.10% 78.89% 8.01% 

7.0-7.9 15.12% 76.40% 8.48% 

8.0-8.9 13.71% 80.26% 6.03% 

9.0-9.9 10.53% 87.37% 2.11% 

10 13.46% 82.69% 3.85% 

    

Nursing Homes Management     
Decentralization shift No change Centralization shift  

1.0-1.9 8.04% 87.50% 4.46% 

2.0-2.9 11.85% 80.74% 7.41% 

3.0-3.9 11.65% 78.02% 10.33% 

4.0-4.9 13.38% 74.88% 11.74% 

5.0-5.9 12.84% 77.04% 10.12% 

6.0-6.9 13.04% 73.71% 13.25% 

7.0-7.9 12.38% 75.13% 12.49% 

8.0-8.9 10.87% 79.74% 9.39% 

9.0-9.9 8.38% 86.39% 5.24% 

10 4.81% 91.35% 3.85% 

    

Healthcate Management     
Decentralization shift No change Centralization shift  

1.0-1.9 8.93% 86.61% 4.46% 

2.0-2.9 11.03% 79.04% 9.93% 

3.0-3.9 11.01% 77.97% 11.01% 

4.0-4.9 11.48% 77.99% 10.54% 

5.0-5.9 10.88% 80.41% 8.72% 

6.0-6.9 12.69% 77.01% 10.30% 

7.0-7.9 12.53% 75.67% 11.81% 

8.0-8.9 10.81% 80.95% 8.24% 

9.0-9.9 6.84% 88.42% 4.74% 

10 4.81% 90.38% 4.81% 



34    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 

 
Table A.5: Attribution of responsibility and the intergovernmental shift of preferences 

 
Lockdown Management     

Decentralization shift No change Centralization shift  

Regional government (-1) 14.20% 69.43% 16.37% 

Shared responsibility (0) 12.97% 78.05% 8.99% 

Central government (1) 14.14% 80.58% 5.28% 

    

Nursing Homes Management     
Decentralization shift No change Centralization shift  

Regional government (-1) 10.04% 75.86% 14.10% 

Shared responsibility (0) 12.40% 78.32% 9.28% 

Central government (1) 16.17% 76.70% 7.13% 

    

Healthcate Management     
Decentralization shift No change Centralization shift  

Regional government (-1) 9.83% 77.35% 12.82% 

Shared responsibility (0) 11.60% 80.26% 8.13% 

Central government (1) 13.67% 79.70% 6.63% 

 
Table A.6: Satisfaction with government performance considered as the responsible for managing the 

lockdown, nursing homes and healthcare during the pandemic 

 

Lockdown management 

Attributed 

responsability 

Satisfaction 

with 

government 

performance 

Decentralization 

shift 

No change Centralization 

shift  

Central 

Government 

Total 17.54 82.46 0 

Very Bad 16.52 83.48 0 

Rather Bad 20.10 79.90 0 

Regular 24.57 75.43 0 

Rather Good 16.00 84.00 0 

Very Good 13.32 86.68 0 

Shared 

responsibility 

Total 13.05 75.81 11.15 

Very Bad 16.93 72.83 10.24 

Rather Bad 15.28 75.75 8.97 

Regular 13.36 78.60 8.04 

Rather Good 12.62 75.45 11.93 

Very Good 11.11 70.90 17.99 

Regional 

Government 

Total 0 75.02 24.98 

Very Bad 0 88.00 12.00 

Rather Bad 0 80.00 20.00 

Regular 0 75.95 24.05 

Rather Good 0 73.33 26.67 

Very Good 0 67.19 32.81 

Total general 13.62 79.32 7.06 

Nursing homes management 

Attributed 

responsability 

Satisfaction 

with 

government 

performance 

Decentralization 

shift 

No change Centralization 

shift  

Central 

Government 

Total 22.60 77.40 0 

Very Bad 19.15 80.85 0 

Rather Bad 23.01 76.99 0 

Regular 29.81 70.19 0 



35    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 
Rather Good 39.39 60.61 0 

Very Good 25.64 74.36 0 

Shared 

responsibility 

Total 13.49 77.75 8.76 

Very Bad 13.49 74.85 11.66 

Rather Bad 14.06 78.86 7.08 

Regular 12.18 80.91 6.91 

Rather Good 14.07 79.26 6.67 

Very Good 15.91 75.00 9.09 

Regional 

Government 

Total 0 75.12 24.88 

Very Bad 0 75.62 24.38 

Rather Bad 0 74.82 25.18 

Regular 0 75.79 24.21 

Rather Good 0 74.26 25.74 

Very Good 0 60.00 40.00 

Total general 12.25 76.84 10.91 

Healthcare management 

Attributed 

responsability 

Satisfaction 

with 

government 

performance 

Decentralization 

shift 

No change Centralization 

shift  

Central 

Government 

Total 22.07 77.93 0 

Very Bad 18.33 81.67 0 

Rather Bad 19.19 80.81 0 

Regular 24.43 75.57 0 

Rather Good 26.55 73.45 0 

Very Good 23.44 76.56 0 

Shared 

responsibility 

Total 10.82 81.32 7.86 

Very Bad 10.67 78.22 11.11 

Rather Bad 11.45 80.00 8.55 

Regular 11.01 81.44 7.55 

Rather Good 10.22 85.30 4.48 

Very Good 9.71 80.57 9.71 

Regional 

Government 

Total 0 76.49 23.51 

Very Bad 0 77.34 22.66 

Rather Bad 0 76.70 23.30 

Regular 0 75.23 24.77 

Rather Good 0 78.28 21.72 

Very Good 0 75.00 25.00 

Total general 11.40 78.96 9.64 

 

 

 

Table A.7: Statements from the questionnaire used to operationalize the dependent and 

explanatory variables13& the operationalization of control variables 
 

Dependent variable: change of preferences in the level of governments in three policy areas. We used 

the following two statements from our questionnaire 

1) Imagine that it is March 1, that is, at the beginning of this crisis, and that you are given a 

choice about which level of government should be primarily responsible for managing the 

following issues. Which level of government would you have chosen on March 1 to manage (1) 

the lockdown during the State of Alarm; (2) the contagion of COVID-19 in nursing homes; and 

(3) the provision of healthcare to anyone in need? 

2) If, after five months, since the crisis began and knowing what you know now, you could 

choose, tell me which level of government you would prefer to be primarily responsible for 

managing each of these three issues? [Respondents could choose among the CG, the regional 

government, or shared responsibility between both government levels.] 

 

13 The original questionnaire was distributed in Spanish. Statements are translated for logical reasons.  



36    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 
 

Explanatory variables: 

 

A. Satisfaction with the management of the pandemic in each of the three policy areas.  

1) How do you assess the work carried out concerning each of these three tasks (managing the 

locjdown, nursing homes and healthcare). We use a 5-point Likert Scale in the questionnaire (from 

very good to very bad).  

B. Partisanship 

1) Which party or coalition did you vote for in the last general election in October 2019?  

2) And in the last autonomic elections, which party or coalition did you vote for? 

C. Degree of autonomism  

 1) A country can be organized territorially in several ways.  

Using a scale from 0 to 10, where 0 represents “Maximum centralization” and 10 represents 

“Maximum decentralization”, in which position would you like Spain to be? 

2) Do you think that, in general, the creation and development of the Autonomous Communities 

has been positive or negative for Spain? 

3) Which of the following phrases do you most identify with?  

01. I feel exclusively Spanish  

02. I feel more Spanish than [name of the Autonomous Community in question]. 

03. I feel as Spanish as [name of the Autonomous Community in question] 

04. I feel more [ name of the Autonomous Community in question ] than Spanish.  

05. I feel exclusively [name of the Autonomous Community in question]. 

D. Attribution of responsibility.  

1) Who do you think has been primarily responsible for managing (1) the lockdown (2) the 

contagion of COVID-19 in nursing homes; and (3) the provision of healthcare [Respondents 

could choose among the CG, the regional government, or shared responsibility between both 

levels of government].  

We then transform the responses to a likert scale (-1 to 1) to distinguish who favours national, shared or 

subnational authority in the model. If respondents chose the central government, we coded it as “1”, 

shared responsibility was coded as “0”, and regional responsibility was coded as “-1”.  

 

 

 

Control variables 

1. Political ideology is on an 11-point scale, with zero being most left-wing and ten being most 

right-wing.14  

2. Interest in politics and trust in regional governments are coded using a 5-point Likert scale (5 

=very high interest/trust. 1 =very low interest/trust).  

3. Household income is in a 5-point scale according to the following range of values: Less than 

1000 €; From 1.000 to 1800 €; From 1,801 to 2,700 €; From 2.701 to 3.900 €; More than 3.900 

€. 

4. Contact with the health system on suspicion of having coronavirus is coded as a binary variable 

(1= yes. 0=no) 

5. Vulnerable employment status is a binary variable coded as 1 for those unemployed or 

benefiting from the cash transfer due to temporary suspension of work (ERTE)15, and 0 for those 

with regular employment, students, domestic worker, and other situation.  

6. The subjective social class is coded using a 7-point scale (7 =high-class. 1 =low-class).  

7. Education is coded on a 5-point scale, with each category referring to the attainment of different 

levels of education (i.e., less than five years of school enrollment. primary education, secondary 

education, and tertiary education).  

8. Incidence of COVID-19 is the cumulative rate of COVID-19 cases in 7 days (the day of the 

survey and the six previous days). This data was multiplied by 100 and divided by the population 

on January 1, 2020 in each Autonomous Community (INE data). 

 

14 We choose an 11-point scale rather than the typical 10-point scale, so that five truly represent the 

middle value.  
15 During the pandemic, the CG created a temporary employment scheme called ERTE. The government 

assumes part of the workers' salary whose companies temporarily suspend activity due to the coronavirus.  



37    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 
 

 

 

Other treatments 

 

1. Do not know and no answer where coded as blank spaces to avoid altering the results.  

2. We tailored question-wording related to the respondent’s region (ACs) of residence to avoid 

generic statements about the region and rather mention the specific name.  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 



38    Canadian Journal of European and Russian Studies, 17(2) 2024: 1-44 
                                                               ISSN 2562-8429 

 

Table A.8: Unweighted summary statistics 
 N Median Median error Variance Minimum Maximum Q1 Q3 

Centralization shift - Lockdown 7022 0 4.526E-5 0.064 0 1 0 0 

Centralization shift – Nursing homes 6957 0 5.614E-5 0.097 0 1 0 0 

Centralization shift - Healthcare 6962 0 5.313E-5 0.087 0 1 0 0 

Decentralization shift - Lockdown 7022 0 6.172E-5 0.120 0 1 0 0 

Decentralization shift – Nursing homes 6957 0 5.906E-5 0.107 0 1 0 0 

Decentralization shift - Healthcare 6962 0 5.716E-5 0.101 0 1 0 0 

Satisfaction with govt. performance managing lockdown 7121 3 2.060E-4 1.370 1 5 3 4 

Satisfaction with govt. performance managing nursing homes 7111 2 1.712E-4 0.944 1 5 1 3 

Satisfaction with govt. performance managing healthcare 7109 3 1.998E-4 1.284 1 5 2 3 

Central government (Vote to PSOE or UP; coalition in the central government) 

  
4615 0 1.358E-4 

0.250 
0 1 0 1 

Regional government (Vote to the political party in the regional government) 

  
7175 0 7.978E-5 

0.209 
0 1 0 1 

Vote to right-wing opposition parties in the CG (PP+VOX)  4615 0 1.151E-4 0.180 0 1 0 0 

Degree of autonomism 

  
6531 6 3.412E-4 

3.162 
1.33 10 4.667 7 

Attribution of responsibility for managing lockdown 

  
7080 1 1.091E-4 

0.380 
-1 1 0 1 

Attribution of responsibility for managing nursing homes 6947 0 1.478E-4 0.671 -1 1 -1 1 

Attribution of responsibility for managing healthcare 6994 0 1.402E-4 0.612 -1 1 -1 1 

Incidence of COVID-19 (7 days) 7175 101.234 0.014 6094.319 14.113 440.465 78.081 179.554 

Subjective social class 7175 4 2.243E-4 1.649 1 7 3 5 

Education 

 
7175 4 1.992E-4 

1.300 
1 5 3 5 

Ideology 

 
7041 5 4.582E-4 

6.626 
0 10 2 5 

Interest in politics 7148 3 1.999E-4 1.299 1 5 2 4 

Trust in regional government 7130 3 1.961E-4 1.245 1 5 2 4 

Contact health authority  

suspicion COVID-19  
7131 0 5.831E-5 

0.110 
0 1 0 0 

Vulnerable employment status 6959 0 6.933E-5 0.148 0 1 0 0 

Income level  5611 2 2.365E-4 1.121 1 5 2 3 

 



       Canadian Journal of European and Russian Studies, 17(2) 2024: ##-## 
                                                               ISSN 2562-8429 

 

 

 39 

Table A.9: Characteristics of the Netquest panel, survey sample design, sample structure and validation methods 

 

Netquest panel at the time of survey 

 

• 157,535 panelists in Spain 

• Average response rate: 50-55% 

• Profile of Spain’s internet usage rate: 82% 

• Age range: 16-24 = 11%; 25-34 = 25%; 35-44 = 28%; 45-54 = 22%; +55 = 14% 

• Gender: Female 64%; Male 36% 

• Socioeconomic status: High 33%; Middle-high 18%; Middle-Middle 26%; Middle-low 9%; Low 14% 

 

Survey sample design 

 

 

• Scope: National (Spain) 

• Universe: People over 18 years of age 

• Size: 7175 interviews 

• Average survey duration: 11 minutes 

• Fieldwork: Conducted through online surveys using Netquest's online panel. 

• Field work period: August 30 to September 23, 2020 

• INITIAL SAMPLE DESIGN BY QUOTAS 

o Gender : Male 49%; Female 51%  

o Age: 18-24 12%; 25-34 15%; 35-44 22%; 45-54 20%; 55-65 18%; 65+ 13% 

o Education 

▪ No education (Unfinished primary studies) 5% 

▪ First Grade (School certificate, 1st stage EGB, more or less 10 years) 8% 

▪ Second Grade. 1st Cycle (School graduate, or EGB 2nd stage, 1st and 2nd ESO-1st cycle- up 

to 14 years) 16% 

▪ 2nd Cycle (FP Iº and IIº, Bachiller superior, BUP, 3º and 4º of ESO (2nd cycle) COU, PREU, 

1º and 2º Bachillerato, up to 18 years old) 33%. 

▪ 1st Cycle (Equivalent to Technical Engineer, 3 years, University Schools, Technical 

Engineers, Technical Architects, Experts, Teachers, ATS, University Graduates, 3 years of 

career, Social Graduates, Social Assistants, etc.) Bachelor's Degree, Degree. 2nd Cycle 

(University students, Higher graduates, Faculties, Higher technical schools, etc.) Master's 

Degree. Doctorate 38% 

 

Sample structure, participation rate and method of calculation 

 

The participation rate is the calculation between the total number of participants and the total number of Invitationi. 

 

  Date Invitations Participations Response rate 

Invitation #1 14/08/2020 17:27 799 535 67 

Invitation #2 25/08/2020 11:24 331 203 61 

Invitation #3 25/08/2020 12:42 176 83 47 

Invitation #4 31/08/2020 13:59 245 87 36 

Invitation #5 31/08/2020 16:48 371 303 82 

Invitation #6 01/09/2020 14:05 393 172 44 

Invitation #7 02/09/2020 11:34 356 150 42 

Invitation #8 02/09/2020 11:53 752 453 60 

Invitation #9 03/09/2020 13:39 325 141 43 

Invitation #10 03/09/2020 19:03 594 289 49 

Invitation #11 04/09/2020 11:30 760 500 66 

Invitation #12 04/09/2020 18:12 727 495 68 



       Canadian Journal of European and Russian Studies, 17(2) 2024: ##-## 
                                                               ISSN 2562-8429 

 

 

 40 

Invitation #13 04/09/2020 18:18 489 218 45 

Invitation #14 05/09/2020 11:36 410 203 50 

Invitation #15 07/09/2020 13:01 1024 694 68 

Invitation #16 08/09/2020 14:27 690 431 62 

Invitation #17 08/09/2020 18:24 145 62 43 

Invitation #18 09/09/2020 9:24 1521 866 57 

Invitation #19 09/09/2020 17:09 345 145 42 

Invitation #20 09/09/2020 18:30 781 455 58 

Invitation #21 10/09/2020 13:07 1311 913 70 

Invitation #22 17/09/2020 11:40 675 182 27 

Invitation #23 17/09/2020 17:55 265 103 39 

Invitation #24 18/09/2020 10:15 250 123 49 

Invitation #25 18/09/2020 10:25 110 29 26 

Invitation #26 18/09/2020 15:20 254 186 73 

Invitation #27 18/09/2020 15:30 178 108 61 

Invitation #28 19/09/2020 17:27 75 36 48 

Invitation #29 19/09/2020 17:39 86 56 65 

Invitation #30 21/09/2020 14:05 238 167 70 

Invitation #31 21/09/2020 14:05 287 182 63 

Invitation #32 21/09/2020 16:46 551 189 34 

Invitation #33 21/09/2020 17:53 96 44 46 

Invitation #34 22/09/2020 10:15 171 51 30 

Invitation #35 22/09/2020 10:20 116 59 51 

Invitation #36 22/09/2020 10:35 171 72 42 

Invitation #37 22/09/2020 16:55 53 24 45 

Invitation #38 23/09/2020 9:30 2 0 0 

Invitation #39 23/09/2020 12:45 481 273 57 

 

 

 

 

Validation methods  

 

1. Netquest includes trick questions such as "How much is one plus two" or "What year are we in?" so that if 

respondents fail, they are screen out of the survey.  

2. Netques have a control of speeders according to the average duration of the survey, if the person takes less than 

20% of the duration to answer, they are also screen out.  

3. At the beginning of the survey Netquest ask for gender and age, and we apply a control that if this does not 

match our database it is filtered out. 

4. Netquest apply recaptcha.  

5. Netquest has a control called relevant ID, which identifies each panelist as to avoid the same person can not do 

the same survey twice. 

6. The authors requested an additional validation method as to filter out straigthliners.  

 

 

Table A.10: OLS models for robustness checks  

 

 (1) (2) (3) (4) (5) (6)    
Satisfaction with govt. 

performance 

0.012 

(0.009) 

0.017 

(0.010) 

-0.005 

(0.010) 

0.002 

(0.012) 

-0.004 

(0.009) 

0.003 

(0.010) 



       Canadian Journal of European and Russian Studies, 17(2) 2024: ##-## 
                                                               ISSN 2562-8429 

 

 

 41 

Central government 

  

-0.010 

(0.022) 

0.003 

(0.026) 

0.054* 

(0.024) 

0.047 

(0.029) 

0.078*** 

(0.023) 

0.067* 

(0.026) 

Regional government 

  

-0.023 

(0.018) 

0.002 

(0.022) 

-0.022 

(0.020) 

-0.007 

(0.024) 

-0.027 

(0.019) 

-0.009 

(0.022) 

Vote to right-wing 

opposition parties in 

the CG (PP+VOX)  

-0.066* 

(0.027) 

-0.026 

(0.033) 

-0.077* 

(0.031) 

-0.035 

(0.037) 

-0.041 

(0.029) 

-0.020 

(0.035) 

Degree of autonomism 

  

-0.010 

(0.005) 

-0.004 

(0.006) 

-0.011 

(0.006) 

-0.009 

(0.007) 

-0.013* 

(0.006) 

-0.010 

(0.007) 

Attribution of 

responsibility 

  

-0.063*** 

(0.015) 

-0.055*** 

(0.016) 

-0.090*** 

(0.013) 

-0.080*** 

(0.015) 

-0.082*** 

(0.012) 

-0.085*** 

(0.014) 

Incidence of COVID-

19 (7 days)  

-0.00002 

(0.0001)  

0.00004 

(0.0001)  

0.0001 

(0.0001) 

Subjective social class  

0.027** 

(0.009)  

0.010 

(0.010)  

0.006 

(0.010) 

Education 

  

0.015 

(0.010)  

0.008 

(0.011)  

-0.006 

(0.011) 

Ideology 

  

-0.003 

(0.005)  

-0.007 

(0.006)  

-0.004 

(0.005) 

Interest in politics  

-0.002 

(0.010)  

0.019 

(0.011)  

0.025* 

(0.010) 

Trust in regional 

government  

-0.030** 

(0.010)  

-0.025* 

(0.011)  

-0.026* 

(0.011) 

Contact health 

authority  

suspicion COVID-19   

-0.020 

(0.029) 

 

-0.031 

(0.032) 

 

-0.004 

(0.030) 

Vulnerable 

employment status  

-0.034 

(0.027)  

-0.041 

(0.030)  

-0.052 

(0.028) 

Income level   

-0.025* 

(0.011)  

-0.003 

(0.012)  

-0.012 

(0.011) 

       

Constant 

  

0.0008 

(0.046) 

-0.068 

(0.080) 

0.056 

(0.046) 

-0.003 

(0.084) 

0.047 

(0.044) 

0.023 

(0.080) 

N 4290 3562 4197 3481 4227 3507 

R2 0.008 0.013 0.024 0.024 0.020 0.025 

F test 5.469*** 3.093*** 17.042*** 5.766*** 14.309*** 5.907*** 

Notes: Model 1 and 4 refers to the management of the lockdown, while model 2 and 5 refers to the nursing homes 

and model 3 and 6 to healthcare. Standard errors are in parenthesis. (o)p<0.1 *p<0.05. **p<0.01. ***p<0.001 
 



       Canadian Journal of European and Russian Studies, 17(2) 2024: ##-## 
                                                               ISSN 2562-8429 

 

 

 42 

Table A.11: Separation plots of logit models to assess goodness of fit 

 

Model 1 - Centralization Model 2 - Centralization

Model 3 - Centralization Model 4 - Centralization

Model 5 - Centralization Model 6 - Centralization



       Canadian Journal of European and Russian Studies, 17(2) 2024: ##-## 
                                                               ISSN 2562-8429 

 

 

 43 

 

 

  

Model 1 - Decentralization Model 2 - Decentralization

Model 4 - DecentralizationModel 3 - Decentralization

Model 5 - Decentralization Model 6 - Decentralization



       Canadian Journal of European and Russian Studies, 17(2) 2024: ##-## 
                                                               ISSN 2562-8429 

 

 

 44 

Published by the Centre for European Studies at Carleton University, Ottawa, Canada 

Available online at: https://ojs.library.carleton.ca/index.php/CJERS/index 

 

The Canadian Journal of European and Russian Studies (CJERS – formerly Review of European and 

Russian Affairs) is an open-access electronic academic peer-reviewed journal: articles are subject to 

double-blind peer-review. Topics relate to the European Union, its Member States, the former Soviet 

Union, and Central and Eastern Europe. The journal is published by the Centre for European Studies, 

an associated unit of the Institute of European, Russian and Eurasian Studies at Carleton University. 

CJERS aims to provide an accessible forum for the promotion and dissemination of high-quality 

research and scholarship.  

 

Contact: 

Carleton University 

The Centre for European Studies 

1103 Dunton Tower 

1125 Colonel By Drive 

Ottawa, ON K1S 5B6 

Canada 

E-mail: CJERS@carleton.ca 

 

Creative Commons License 

 

 
https://creativecommons.org/licenses/by-nc-nd/4.0/ 

 

This article is licensed under a Creative Commons Attribution-Non-Commercial-No Derivs 4.0 

Unported License (CC BY-NC-ND 4.0).  Articles appearing in this publication may be freely quoted 

and reproduced, in unadapted form only, provided the source is acknowledged. We request that for 

any online postings, you do so by providing the title and/or abstract with the DOI link back to the 

original version of the article on the journal's website. 

ISSN: 2562-8429 

 

© 2024 The Author(s) 

 

 

https://ojs.library.carleton.ca/index.php/CJERS/index
mailto:CJERS@carleton.ca
https://creativecommons.org/licenses/by-nc-nd/4.0/

