Content uploaded by Carol Galais
Author content
All content in this area was uploaded by Carol Galais on Dec 15, 2015
Content may be subject to copyright.
Full Terms & Conditions of access and use can be found at
http://www.tandfonline.com/action/journalInformation?journalCode=fwep20
Download by: [95.23.229.64] Date: 15 December 2015, At: 10:07
West European Politics
ISSN: 0140-2382 (Print) 1743-9655 (Online) Journal homepage: http://www.tandfonline.com/loi/fwep20
Do people feel more of a duty to vote in some
elections?
Carol Galais & André Blais
To cite this article: Carol Galais & André Blais (2015): Do people feel more of a duty to vote in
some elections?, West European Politics, DOI: 10.1080/01402382.2015.1104994
To link to this article: http://dx.doi.org/10.1080/01402382.2015.1104994
Published online: 15 Dec 2015.
Submit your article to this journal
View related articles
View Crossmark data
WEST EUROPEAN POLITICS, 2015
http://dx.doi.org/10.1080/01402382.2015.1104994
© 2015 Taylor & Francis
Do people feel more of a duty to vote in some
elections?
Carol Galais and André Blais
KEYWORDS Duty; election level; turnout; multilevel government
Quite a few people believe that voting is not only a right, but also a moral
duty that the good citizen should full. is sense of civic duty is perhaps
the most powerful factor that induces citizens to go the polling station even
when they are not that interested in politics, do not have strong preferences
among the parties or candidates, and understand that their own single vote
will not decide the outcome of an election (Blais 2000; Riker and Ordeshook
1968).
is attitude may help us to understand not only why people vote or abstain,
but also why their decision to do so diers across elections. Indeed, there is a
large literature in the eld of political behaviour showing persistent dierences
in turnout levels across types of election (see, for instance, Homann-Martinot
1992; Morlan 1984). e reasons for this gap are not entirely clear (Blais 2000:
36–39). In this study we explore one such possibility, that citizens have dierent
attitudes towards dierent elections. More specically, we investigate whether
people feel less of a duty to vote in some types of elections. Moreover, we inquire
into the causes of a ‘dierentiated’ sense of duty to vote.
ABSTRACT
Prior research on civic duty has focused on national elections, believed to be the
most salient. Evidence on turnout gaps between election levels suggests that it
is relevant to inquire whether people feel that they have the same duty to vote
in national, subnational, and supranational elections. The article investigates
this phenomenon, comparing citizens’ attitudes towards national, European,
and regional elections in ten regions from four countries. About one-quarter of
European citizens demonstrate a lesser degree of duty towards European rather
than in national elections. Dierences in duty levels for national and regional
elections are infrequent and concentrated in regions with nationalist movements.
Both rational and identity considerations explain why some individuals feel less
obliged to vote in a particular election than in another, but the latter matter more.
CONTACT Carol Galais carolgalais@gmail.com
Downloaded by [95.23.229.64] at 10:07 15 December 2015
2 C. GALAIS AND A. BLAIS
Previous research on civic duty has focused on national elections, which are
usually deemed to be the most relevant elections (Franklin 2001; Van der Eijk
and Franklin 1996), and for which the sense of civic duty is relatively widespread
Blais (2000). Nevertheless, the static conception of territorial organisation upon
which this assumption is based is dubious. On the one hand, decisions made
by the European Union have become more and more signicant (Craig 2010;
Piris 2010). On the other hand, there has been a substantial transfer of author-
ity from the central government to regional governments in many countries,
and as a consequence the claim that regional elections are secondary has been
seriously challenged (Henderson and McEwen 2015; Schakel and Jeery 2013).
It is thus relevant to inquire whether people feel that they have the same duty
to vote in national, subnational, and supranational elections. As far as we can
tell, this question has never been addressed in any systematic fashion, and we
hope to ll a gap in the literature and in the process contribute to a deeper
understanding of the sources of civic duty.
Building on the literature on the turnout gap, we identify two types of factors
that may explain why one would feel less of a duty to vote in one particular
election. e rst is inspired by Reif and Schmitt’s (1980) characterisation of
a second-order election being one in which there is less at stake. According to
this perspective, some levels of government have more impact than others on
citizens’ lives, and people would logically consider that they have less of a duty
to vote when the representatives to be elected are perceived to have little power.
e second explanation for dierentiated duty borrows from the literature on
dierential abstention, which argues that an election may be less appealing for
symbolic reasons, because the person does not identify with the community
that the election is about. We therefore expect a weaker sense of duty to vote
in an election when the individual does not feel particularly attached to the
territorial community in which the election takes place.
In the following pages, we investigate the phenomenon of dierentiated duty
from a comparative perspective, confronting perceptions and attitudes about
national and European elections, on the one side, and national and regional
elections, on the other. Our research relies on a new and original dataset that
covers 10 regions from four European countries plus Canada. e descriptive
evidence indicates that relatively few people have a dierential sense of duty,
but that this occurs more oen when it comes to comparing national with
European elections. About one-quarter of European citizens express dierent
degrees of engagement when they rate their feelings of duty to vote in national
and European elections, most of them indicating a stronger duty to vote in
national elections. Dierences in duty for national and regional elections are
less frequent but concentrated in regions with nationalist movements. Our
multivariate estimations suggest that both rational and aective/symbolic con-
siderations explain why some individuals feel less obliged to vote in a particular
Downloaded by [95.23.229.64] at 10:07 15 December 2015
WEST EUROPEAN POLITICS 3
election, although explanations based on identity work somewhat better, espe-
cially among elderly citizens.
Theoretical framework
Sense of civic duty was introduced in models of turnout when scholars noted
that the ‘rational’ explanation could not show why so many people voted despite
the fact that the probability that their vote would make a dierence was close
to nil. In order to solve this problem, some scholars began to introduce in their
voting models a duty term (Lewis-Beck et al. 2009; Riker and Ordeshook 1968;
Rosenstone and Hansen 1993). As a result, it is nowadays widely accepted in the
literature that some people vote because they believe that the good citizen has
a moral obligation to vote and thus not voting is ethically wrong Blais (2000;
Blais and Achen 2015).
Following this logic, one potential explanation for a ‘turnout gap’ between
two dierent elections (Rallings and rasher 2007) is that some people believe
there is a duty to vote in one election but not in the other. is might seem
counterintuitive because of the moral nature of the duty to vote. Indeed, sense
of duty is a belief that belongs to the moral domain (Graham et al. 2011). As a
moral attitude, it is associated with judgements of wrong and right, which tend
to be absolute (Dworkin 1978; Rawls 1971). is would imply that sense of duty
to vote will hardly vary depending on the level of the election. As Frank Sinatra
would say, you’ve either got it or you don’t. But if it varies to some extent, what
would be the reasons for a dierentiated feeling of duty?
In the absence of any literature on this specic question, we borrow from
research on a related topic, the presence or absence of ‘dierential’ turnout. e
‘turnout gap’ ‒ that is, the presence of persistent dierences in turnout rates
across election levels ‒ has fuelled two dierent research strands. e rst is
the theory of second-order elections (Reif and Schmitt 1980). According to the
classical formulation, a ‘second order’ election is one in which ‘there is less at
stake’ (Reif and Schmitt 1980: 9). ese two authors argue that the rst-order
elections are the national parliamentary elections in parliamentary systems and
the national presidential elections in presidential systems. And then ‘there is a
plethora of second-order elections: by-elections, municipal elections, various
sorts of regional elections, those to a second-chamber, and the like’ (Reif and
Schmitt 1980: 8). ey explicitly consider European and regional elections to be
second-order elections, because there is less at stake in the eyes of the elector-
ate. As a consequence, many people feel less obliged to vote in those elections.
Second, there is a less prominent research strand on dierential abstention.
‘Dierential abstention’ was rst dened, to the best of our knowledge, by
Shively (1992: 309) as ‘the movement of citizens into and out of abstention
across a pair of elections’. Although initially focused on the US context, the role
of the media, and the availability of information about the candidates, research
Downloaded by [95.23.229.64] at 10:07 15 December 2015
4 C. GALAIS AND A. BLAIS
on dierential abstention has moved to Spain in considering the dierences
between national and regional elections, or, more specically, the dierent
attitudes that voters hold towards the dierent levels of government, which
determine their propensity to vote less at one level, usually the subnational
one (Colomer and Padró-Solanet 1992; Liñeira and Vallés 2014; Montero and
Font 1991; Riera 2011, 2012; Vallès 2009). erefore, a dierential abstainer is
an individual who systematically votes in the national elections and abstains
in the regional elections (Riba 2000: 64).
Unlike the second-order perspective, the explanation for dierential absten-
tion is socio-psychological. e most oen referred cause for this behaviour
is the individual’s origin and attachment (or lack thereof) to the subnational
territory, as a function of his/her birthplace and how many years he/she has
lived in the region. e assumption is that a personal story of (intrastate) migra-
tion hampers identication with subnational political institutions, including
government, parliament, and parties (Riba 2000). It is also a matter of social
capital, as non-native residents are supposed to be less integrated into public
life (Liñeira and Vallés 2014).
e ‘dierential abstention’ research strand pays more attention to the aec-
tive connection between the individual and the territorial community than
to the relative power of the oce to be elected. is approach has solid bases
in theories of social identity. e latter, dened as a ‘part of an individual’s
self-concept which derives from his knowledge of his membership of a social
group or groups together with the value and emotional signicance attached to
that membership’ (Tejfel 1978: 372), greatly inuences political participation.
Indeed, the greater the identication with that community, the greater the
chances of adopting the participatory norms of that group (Terry et al. 1999).
ese two strands of literature provide two competing hypotheses about why
people may feel they have a weaker sense of duty in some elections. e rst is
the ‘rational’ interpretation that people feel less of a duty when they perceive
that less is at stake at a given level, because one level of government is less
‘powerful’ than the other. e second is the ‘socio-psychological’ interpretation
according to which some people feel less integrated into a given community,
which in turn produces a weaker sense of duty to participate in that commu-
nity’s elections.
Recent works have included both explanations in their voting models, nd-
ing the second one to have a slightly more prominent role (Henderson and
McEwen 2015).
1
Following these works, we expect attachment to have a greater
inuence on dierential sense of duty than perceived impact. e rst reason
for this expectation relies on the moral nature of duty. Indeed, there is a research
strand which argues that what dierentiates morality from convention is aec-
tion and emotions (Haidt 2003), oriented both towards the self and towards
others (Prinz 2008). erefore, duty will be conceptually closer to aective
and group considerations than to rational ones. Also, from a purely political
Downloaded by [95.23.229.64] at 10:07 15 December 2015
WEST EUROPEAN POLITICS 5
psychology perspective, the belief that voting is a duty is part of the ‘symbolic
politics’ dimension (Sears et al. 1979). ese attitudes, more stable, formed
earlier, and more aective (e.g. ideology, party identication, values), inu-
ence more specic outlooks, such as issue opinions and candidate evaluations,
and, nally, individuals’ political behaviour.2 If civic duty develops early in life
just like other political attitudes, including orientations towards representative
institutions (Dennis 1973) or notions of citizenship (Van Deth et al. 2011), it
should be relatively impervious to more specic orientations about the polit-
ical context, such as evaluations and perceptions about government levels or
the relative importance of elections.3 In sum, the identication with a political
community, as well as the duty to vote, will come rst in the causal chain rather
than rational considerations about the degree of inuence of a government.
Although this is not the main goal of our article, the prominence of iden-
tication with a political community requires further reection. ere is a
growing literature on the causes and eects of territorial identity (Berg 2007;
Bruter 2005; Hooghe and Marks 2001; Keating 2001). Yet in order to unravel
the consequences of territorial identity on other attitudinal or behavioural
outcomes ‘it is important to go beyond merely underscoring the importance
of group “attachment” and “othering” by specifying who those groups are ...
whether identity should be measured at a low level of abstraction (for example,
regional identity) or higher abstraction level (for example, national identity)
and how this level of identity relates to other levels of identication’ (Mols and
Weber 2013: 511). Following this advice, we adopt a comparative perspective
and contrast the degree of attachment to the region, the country, and (where
applicable) the European Union (EU). In this sense, we know that identication
with the EU, the nation and the region tends to be positively, rather than neg-
atively, associated (Castano 2004). As a consequence, dierential attachment
to dierent communities may not be very frequent.
Finally, we have to consider the possibility that dierentiated attachment
does not exert a homogeneous eect across the population. On the one hand,
we consider that this factor will have a larger eect among the elderly. According
to the self-categorisation theory, people can identify with dierent social groups
(Turner et al. 1994). Indeed, multiple loyalties have increased in Western Europe
over the last three decades (Klandermans et al. 2004; Medrano and Gutiérrez
2001). Yet this is a recent phenomenon (McManus-Czubińska et al. 2003),
which leads us to expect dierentiated territorial attachment to be more fre-
quent among older citizens. is should be particularly true for newer territorial
units and political institutions (i.e. the EU), as the elderly have been consistently
found to be less supportive of Europe (see Castano 2004), probably because
the older generations have come into contact with these institutions too late
to aect the core of their crystallised political identity. Furthermore, age has
previously been found to be positively related to the duty to vote, because
older people identify more with their community and are more exposed to
Downloaded by [95.23.229.64] at 10:07 15 December 2015
6 C. GALAIS AND A. BLAIS
the dominant social norm that voting is a duty (Santana-Leitner 2008: 76). As
a result, dierential attachment may have a stronger eect on dierentiated
duty among the elderly.
We also consider the possibility that territorial attachment moderates the
relationship between perceived impact and duty. More specically, when an
individual feels more attached to one territorial community they may not care
much whether the territorial government has more or less power. erefore,
the eect of perceived impact should be lower among those who score high
on dierential attachment.
Our general expectations are thus as follows. First, since duty to vote is rooted
in the belief and moral system of the individual, most citizens do not discrimi-
nate between levels of elections. Mostly, they have a sense of civic duty, or they
do not. Second, when people feel a weaker sense of duty at one level, territorial
identity considerations (one’s sense of attachment to a given community) should
matter more than ‘rational’ ones (perceptions of impact). ird, the eects of
dierential attachment should be greater among older citizens. Fourth, there
should be a negative interaction eect between impact and attachment.
Methods and data
We use 21 election (European/national/sub-national) studies conducted
in 10 regions within ve countries (Canada, France, Germany, Spain, and
Switzerland) in the framework of the Making Electoral Democracy Work
(MEDW) project (Blais 2010). e surveys are online quota-based polls that
guarantee the representativeness of the samples of the population under study
regarding age, gender, education, and region. Each study includes a campaign
and a post-election survey. is article relies on the campaign surveys, which
contain all the relevant information.4
e surveys were conducted in each of the 10 regions at the time of regional,
national, and European elections (see Appendix 1 for the list of surveys and
elections). Each regional survey has a sample size of around 1,000 respondents,
making a total of 2,000 to 3,000 observations per region, with the exception of
Bavaria, where we have about 6,000, and the Canadian provinces of Quebec
and Ontario, where we have only one survey and about 1,000 respondents.5
For the purposes of our study, we merge all the surveys conducted in the same
region to present an overview of the patterns in each region. We then proceed
to a global examination of the sources of dierential duty, running multivariate
estimations with a pooled dataset combining all the regions.
In each election survey respondents were asked to indicate whether they con-
sider voting rst and foremost as a duty or as a choice in European, national, and
regional elections, to express their relative attachment to each of these three terri-
torial communities, and to evaluate the relative impact of each level of government
on their own well-being (see Appendix 2 for the exact wording of questions).6
Downloaded by [95.23.229.64] at 10:07 15 December 2015
WEST EUROPEAN POLITICS 7
We rst determine whether individuals dierentiate across levels regarding
their sense of duty, attachment, and perceptions of policy impact. In each case,
we rst determine how many individuals perceive each level as being less salient
(in terms of duty, attachment or perceived impact), and how many make no
dierence. For this purpose, we make two sets of comparisons: national against
European elections and national against regional elections. ese comparisons
determine the number of individuals and regions in our analyses, as we lack
data regarding the European level for Canada and Switzerland and for the
regional level in France.
With regard to the measure of our dependent variable, respondents were
asked the Blais and Achen (2015) duty/choice question about whether they
view voting rst and foremost as a duty or as a matter of personal choice. ose
who answer ‘choice’ are given a score of 0 on the duty scale. ose who answer
‘duty’ are given a score of 1, 2, or 3 depending on whether they feel that duty
not very strongly, somewhat strongly, or very strongly. A person has a weaker
duty in a given level of election when their duty score is lower for that election
than for the other. erefore, by subtracting duty in the non-national election
from duty in the national election we have created two variables, one for each
comparison (national against European, national against regional) that yields
two three-category variables. e middle category corresponds to those who
give the same scores of duty to the two elections. e two other categories refer
to those who give a higher (positive value) or lower (negative value) score to
the national election.
We proceed similarly for the variables tapping dierences between levels
regarding impact and attachment. On the basis of the questions on feelings of
attachment and perceived impact (with a 0 to 10 scale), we have created four
variables, which tap the dierences between the national and the non-national
(either European or regional) levels for both impact and attachment. As we
did for dierentiated duty, we subtract the European or regional score from
the national, and we then collapse positive and negative values. e value 0
indicates no dierence, positive values indicate that the individual feels more
attached to the national level, or that they feel that national policies aect them
more than those of the non-national (European or regional) level. Negative
values refer to the opposite ‒ that is, the individual perceives more impact for
the European/regional policies on their life or that they feel more attached to
the non-national territorial community. e main independent variables thus
compare the national level with the other two levels.7
We estimate the eects of dierential attachment and impact on dieren-
tiated duty by means of ordinal logistic regressions, as the dependent variable
is an ordinal variable, ranking from less dutiful at the national level to more
dutiful at the national level.8 We also include as control variables gender, age,
education, the number of years one has lived in the country or the region
(depending on the type of comparison), the level of general political interest,
Downloaded by [95.23.229.64] at 10:07 15 December 2015
8 C. GALAIS AND A. BLAIS
dummies for each of the regions (Bavaria is the reference category), and nally
a dummy for whether the survey was conducted at the time of a national elec-
tion.9 Our estimations also include two interactions with age (one with the
attachment variable, the other with the impact variable), to determine whether
dierentiated attachment has a stronger eect among older citizens. We nally
test the possibility that attachment moderates the eects of impact on duty by
means of an interaction between our two main independent variables.
Comparing European and national elections
Table 1 shows the proportion of citizens, in each of the six European Union
regions covered by the MEDW project, who indicate that they have a weaker
or stronger sense of duty in European or national elections. e most striking
result is that an overwhelming majority feels the same way about the two elec-
tions: around 80 per cent in Spain and Germany and 70 per cent in France. Only
a minority, about 20 per cent in France and 15 per cent in Spain and Germany
express a weaker sense of civic duty in European than in national elections.10
e table also shows how many citizens indicate that the European Union
has more or less impact on their well-being as compared to their national
government, and who feel more or less strongly attached to their country than
to Europe. e data reveal that the European Union is perceived by many to
have less inuence on their daily well-being. e proportion of respondents
who feel that the policies of the European level of government have less impact
on their lives is about 50 per cent in France, 60 per cent in Germany (more in
Bavaria than in Lowen Saxony), and 40 per cent in Spain. All in all, about half
of the citizens in these three countries consider the EU as a less relevant level of
government. ese are of course signicant numbers, though we should keep in
mind that many people do not make such a distinction. It is also interesting to
Table 1.Distribution (%) of differential duty, attachment and impact, per region: National
vs European (column percentages).
IDF Provence Catalonia Madrid Low Sax Bavaria
Duty
Less national 5 4 2 0 5 5
The same 75 72 82 83 79 78
More national 20 24 16 17 16 17
Impact
Less national 12 9 15 14 6 6
The same 39 36 45 42 32 26
More national 49 55 41 44 62 68
Attachment
Less national 6 5 32 6 9 6
The same 25 18 32 23 23 20
More national 69 76 36 71 68 74
Downloaded by [95.23.229.64] at 10:07 15 December 2015
WEST EUROPEAN POLITICS 9
note that those who perceive the European Union to have less impact are more
numerous in Germany and fewer in Spain. Furthermore, a minority (about 10
per cent in France and Spain, and 5 per cent in Germany) view the European
Union as in fact more inuential.
What about attachment? We see, not surprisingly, that most citizens in the
three European Union countries (France, Germany, and Spain) have a weaker
attachment to Europe; the percentages are around 70 per cent. A tiny minority
(about 5 per cent) has a stronger attachment to Europe, and close to a quarter
Table 2.Ordered logit estimations of differential duty: National vs. European.
Standard errors in parentheses.
*p < 0.05;
**p < 0.001.
(1)Basic
(2) Age*
impact
(3)Age*
Attachment
(4)Dif.
impact* Dif.
attachment
Female −0.09* −0.09 −0.09* −0.08*
(0.04) (0.04) (0.04) (0.04)
Age 0.21 0.35 −0.26 0.20
(0.27) (0.28) (0.29) (0.27)
Education −0.33* −0.34* −0.32* −0.33*
(0.13) (0.13) (0.13) (0.12)
Interest in politics 0.42** 0.42** 0.42** 0.42**
(0.08) (0.08) (0.08) (0.04)
Dif. impact (less/same/more
national)
0.24**
(0.04)
0.39**
(0.09)
0.24**
(0.04)
0.29**
(0.06)
Dif. attachment (less/same/more
national)
0.45**
(0.04)
0.44**
(0.04)
0.08
(0.09)
0.48**
(0.04)
Level: national 0.18** 0.18** 0.17** 0.18*
(0.05) (0.05) (0.05) (0.05)
IDF 0.62** 0.63** 0.61** 0.62**
(0.09) (0.09) (0.09) (0.09)
Provence 0.56** 0.56** 0.55** 0.56**
(0.08) (0.08) (0.08) (0.08)
Catalonia 0.34** 0.34** 0.34** 0.34**
(0.09) (0.09) (0.09) (0.09)
Madrid 0.18* 0.18* 0.17* 0.17*
(0.08) (0.08) (0.08) (0.08)
Low Sax. −0.07 −0.07 −0.07 −0.07
(0.07) (0.07) (0.07) (0.07)
Bavaria (Ref.) 0.00 0.00 0.00 0.00
(.) (.) (.) (.)
Time living in country −0.50 −0.50 −0.52 −0.51
(0.29) (0.29) (0.29) (0.29)
Age * Dif. impact −0.32
(0.18)
Age * Dif. attachment 0.82**
(0.18)
Dif. impact* Dif. attachment −0.08
(0.05)
cut1_cons −2.86** −2.8** −3.09** −2.84**
(0.12) (0.13) (0.13) (0.12)
cut2_cons 2.02** 2.09** 1.81** 2.03**
(0.12) (0.12) (0.12) (0.12)
Pseudo R2 0.026 0.026 0.027 0.026
N14,827 14,827 14,827 14,827
Downloaded by [95.23.229.64] at 10:07 15 December 2015
10 C. GALAIS AND A. BLAIS
of respondents say that they are equally attached to their own country and to
Europe. ere is one huge exception to that pattern, which is Catalonia, where
about one-third of respondents identify more strongly with Europe, one-third
with Spain, and one-third have equal attachment.
Our next research question is whether dierential duty might be due to
perceived dierences in the impact of European and national policies or to dif-
ferences in territorial identication. To address this question, we regress dier-
ential sense of duty on our two main independent variables, dierential impact
and dierential attachment, taking into account all the aforementioned control
variables (Table 2). e analysis allows us to determine whether a weaker sense
of duty for a given electoral level is better explained by perceived dierences in
the impact of European and national policies or by dierences in attachment
to Europe and one’s own country.
Bearing in mind that the coecients are not directly interpretable (as they
are ordered logistic coecients), we can draw some interesting conclusions. e
more educated are more likely to feel less duty to vote in the national election.
e propensity to have a stronger sense of duty for national elections increases
with the level of political interest. When attitudes are measured at the time of
a national election, individuals are more prone to express more duty for the
Table 3.Predicted probabilities of differential duty: National vs European.
National vs. European
Dierence in attachment Dierence in impact
Duty Less national Same
More
national Less national Same
More
national
Less national 0.04 0.04 0.03 0.03 0.03 0.04
Same 0.86 0.81 0.74 0.83 0.79 0.74
More national 0.09 0.15 0.23 0.14 0.18 0.22
Table 4.Distribution (%) of differential duty, attachment and impact, per region: National
vs regional (column percentages).
Lucerne Zurich Catalonia Madrid
Low
Sax. Bavaria Ontario Quebec
Duty
Less national 4 2 9 1 2 2 0 7
The same 88 87 86 94 90 92 95 91
More national 8 11 5 5 8 6 4 2
Impact
Less national 29 27 21 17 16 28 18 31
The same 50 52 51 52 44 40 64 59
More national 21 21 28 31 40 32 18 10
Attachment
Less national 19 14 53 18 29 45 5 45
The same 50 43 33 62 42 37 61 38
More national 31 44 14 20 29 18 34 17
Downloaded by [95.23.229.64] at 10:07 15 December 2015
WEST EUROPEAN POLITICS 11
national level. Sex, age, and length of time living in the country (in this last
case, perhaps because it is highly collinear with age and also because there are
few non-nationals in our sample with the right to vote, about 6 per cent) do not
play any role. French and Spanish citizens are more prone to feel more dutiful
towards the national level, compared to Bavarians.
As for the main independent variables, dierences in attachment and in
perceived impact are related to dierences in duty in the expected direction.
ose with higher values on both independent variables (with higher scores
for the national level) are signicantly more dutiful towards national elections.
e same is true in the opposite direction: lower values on the two independent
variables (which correspond to higher values of attachment and impact for the
European level) predict stronger duty at the European level. It is worth noting
that separate estimations of this model in every region conrm the robustness of
our results: 11 of the 12 coecients associated with dierential attachment and
impact are positive and signicant (see Appendix 3). e size of the coecients
also suggests that the explanatory power of attachment is higher than that of
impact. Wald tests (not shown) conrm that the dierence in the magnitude
of these coecients is indeed statistically signicant.
Columns 2 and 3 in Table 2 test the interaction eects between age and
our main independent variables. Only the interaction eect with attachment
(column 3) is signicant, the positive coecient meaning that dierential ter-
ritorial identication matters more among older people.11 As for the interac-
tion between our two main independent variables (column 4), the eect is
not signicant. is suggests that the eects of dierential attachment and
dierential impact are additive; attachment does not moderate the eect of
perceived impact on duty.
Table 3 presents the predicted probabilities of being more, less, or equally
dutiful towards the national level depending on the value of these explanatory
variables. ese simulations are based on the equations used in the basic model
(column 1 in Table 4).
We can see that very few people feel less of a duty in national elections,
whatever their views about the relative power of the two institutions and their
sense of identity. e most interesting results concern the propensity towards
a weaker duty in European elections. at propensity increases from 15 per
cent when one is equally attached to Europe and one’s country to 23 per cent
when one is less attached to Europe. We can thus conclude that sense of duty
to vote is weaker in European elections in part because many people are more
attached to their country than to Europe. Likewise, the likelihood of feeling
less of a duty in European elections increases from 18 per cent when one gives
the same impact score to the European Union and the national government to
22 per cent when one believes that the European Union has less power than
the national government.
Downloaded by [95.23.229.64] at 10:07 15 December 2015
12 C. GALAIS AND A. BLAIS
We should not overstate the magnitude of these eects. On the one hand, an
overwhelming majority (74 per cent) of those who are less attached to Europe
than to their country do not have a dierential sense of duty. On the other hand,
our ndings suggest that if everyone felt as strongly attached to Europe as to
their country, 15 per cent would still have a weaker sense of duty in European
elections. Our rst hypothesis, according to which most people do not dif-
ferentiate the level of election when it comes to feeling a moral obligation to
vote or not, is conrmed. e second hypothesis, which states that dierential
attachment better predicts dierential duty than dierential perceptions of
power, is also supported. Finally, we also found that a dierentiated territorial
identication is more explicative of dierentiated duty among older people.
Comparing regional and national elections
We can now change the perspective and compare the regional and the national
levels. Note that we are able now to consider the Canadian provinces of Quebec
and Ontario as well as the Swiss cantons of Lucerne and Zurich but we do not
include France since the subnational unit that was examined in that country
was the municipality. We should also note that the four countries that we cover
are decentralised countries in which regional governments are quite important
and that two of the regions (Catalonia and Quebec) have strong nationalist
movements.
e rst question is whether many people have dierent feelings of duty
to vote depending on the level of the election. Table 3 shows the distribution
of dierential duty in the various regions. e numbers speak for themselves.
Everywhere about 90 per cent of the respondents indicate the same degree of
duty (or choice) in the national and regional elections. e proportion of people
who feel more or less duty bound to vote at one level surpasses 10 per cent only
in Zurich, where 11 per cent report a stronger sense of duty in national elections.
Next, we want to know whether one level of government is perceived to be
more inuential than the other. e modal response in each of the eight regions
is to see no dierence in the relative overall power of the two levels of govern-
ment (Table 4). at view is most widespread in Canada and less frequent in
Germany. Most of the time, about a quarter sees the national government as
more powerful and as many perceive the regional government as more inuen-
tial. e percentage of respondents choosing the national government as having
more leverage is highest in Lower Saxony while it is in Quebec that the greatest
number thinks that the regional government matters more. ese results conrm
recent ndings in the literature on second-order elections (Schakel and Jeery
2013); it is a mistake automatically to characterise regional governments as less
powerful, and citizens’ perceptions vary a lot, both across and within regions.
What about the strength of attachment to the region versus the country? In
ve of the eight regions (Lucerne, Zurich, Ontario, Madrid, and Lower Saxony)
Downloaded by [95.23.229.64] at 10:07 15 December 2015
WEST EUROPEAN POLITICS 13
the modal response is again no dierence. But the number of people who are
more attached to the region and to the country is about equal in the latter two
regions while many more have a stronger degree of attachment to the country
in Lucerne, Zurich, and Ontario. Finally, about half the respondents are more
strongly attached to their region in Catalonia, Bavaria, and Quebec, while only a
small minority identify more with their country. ere is thus more dierential
Table 5.Ordered logit estimations of differential duty: National vs. regional.
Standard errors in parentheses.
*p < 0.05;
**p < 0.001.
(1) Basic (2) Age* Impact
(3) Age*
Attachment
(4) Dif. impact*
Dif. attachment
Female 0.01 0.01 0.01 −0.01
(0.05) (0.05) (0.05) (0.05)
Age −0.88** −0.86** −0.97** −0.90**
(0.18) (0.18) (0.18) (0.18)
Education −0.16 −0.16 −0.15 −0.18
(0.15) (0.15) (0.15) (0.14)
Interest in politics 0.31** 0.31** 0.32** 0.29*
(0.09) (0.09) (0.09) (0.09)
Dif. impact (less/same/more
national)
0.37**
(0.04)
0.48**
(0.04)
0.37**
(0.04)
0.39**
(0.04)
Dif. attachment (less/same/
more national)
0.52**
(0.04)
0.52**
(0.04)
0.27*
(0.09)
0.52**
(0.04)
Level: national −0.05 −0.05 −0.05 −0.04
(0.06) (0.06) (0.06) (0.06)
Lucerne −0.01 −0.02 −0.02 −0.01
(0.11) (0.11) (0.11) (0.11)
Zurich 0.45** 0.44** 0.44** 0.47**
(0.10) (0.10) (0.10) (0.10)
Catalonia −0.72** −0.72** −0.70** −0.70**
(0.12) (0.12) (0.12) (0.12)
Madrid −0.08 −0.08 −0.08 −0.09
(0.11) (0.11) (0.11) (0.10)
Low Sax. 0.14 0.14 0.13 0.13
(0.10) (0.10) (0.10) (0.10)
Bavaria (Ref.) 0.00 0.00 0.00 0.00
(.) (.) (.) (.)
Ontario −0.06 −0.07 −0.07 −0.05
(0.11) (0.11) (0.11) (0.11)
Québec −0.90** −0.91** −0.88** −0.85**
(0.11) (0.11) (0.11) (0.11)
Time living in region −0.40* −0.41* −0.36* −0.41*
(0.17) (0.17) (0.17) (0.17)
Age * Dif. impact −0.24
(0.18)
Age * Dif. attachment 0.57**
(0.17)
Dif. impact* Dif. attachment −0.02
(0.04)
cut1_cons −4.24** −4.24** −4.27** −4.27**
(0.13) (0.13) (0.13) (0.13)
cut2_cons 2.27** 2.28** 2.26** 2.21**
(0.12) (0.12) (0.12) (0.12)
Pseudo R2 0.06 0.062 0.062 0.061
N18,230 18,230 1,8230 18,230
Downloaded by [95.23.229.64] at 10:07 15 December 2015
14 C. GALAIS AND A. BLAIS
attachment in general, and there are greater variations across regions. It is
also interesting to note that feelings of attachment are very similar in Bavaria,
Catalonia, and Quebec, even though there is no signicant separatist movement
in the rst case.
We wish to determine whether feeling a stronger attachment to the region
(country) or perceiving the regional (national) government as more powerful
leads to having a stronger sense of duty in regional (national) elections. Table
5 shows that older respondents and those who have been living in the region
for a longer period of time are more inclined to say that they have a stronger
duty in regional elections. e most interested in politics are more prone to
have a stronger duty in national elections. We also observe, without surprise,
that civic duty is weaker for national elections in Catalonia and Quebec than
in Bavaria (reference category).12
Dierential attachment and dierential impact have the expected eect on
dierential duty. e stronger one’s attachment to country and the greater the
perceived power of the national government, the greater is the tendency to feel a
stronger duty to vote in national elections and the lower the propensity to have
a stronger sense of duty in regional elections. ese attitudes and perceptions
do matter, and, as expected, dierential attachment appears to matter slightly
more than perceived dierential impact. Again, we performed a Wald test that
conrmed the inequality of coecients, and then checked the robustness of
our ndings by performing separate estimations in each of the eight regions.
As can be seen in Appendix 3, dierential attachment and dierential impact
have a positive and signicant coecient 14 times out of 16.13
As for the alternative models, we nd again only a signicant interaction
between age and attachment, meaning that older people are more inclined to
develop dierentiated levels of duty on the basis of their dierentiated iden-
tities (see column 3). When we look at the model displayed in column 4, we
can conclude that both dierential impact and dierential attachment exert
independent eects on dierential duty. e eect of perceived institutional
impact is not conditional on feelings of attachment.
Table 6 shows the predicted probabilities of being more, less or equally duti-
ful towards the national level (as compared to the regional level) depending on
the values of dierential impact and attachment. As with the previous compar-
ison between the European and the national level, these predicted probabilities
Table 6.Predicted probabilities of differential duty: National vs. European.
National vs regional
Dierence in attachment Dierence in impact of policies
Duty Less national Same More
national
Less national Same More
national
Less national 0.05 0.03 0.02 0.05 0.03 0.02
Same 0.91 0.91 0.88 0.91 0.90 0.89
More national 0.04 0.06 0.10 0.05 0.06 0.09
Downloaded by [95.23.229.64] at 10:07 15 December 2015
WEST EUROPEAN POLITICS 15
have been computed on the basis of the basic model presented in the rst
column of Table 5.
e impact of the perceived inuence of policies and attachment on dier-
ential sense of duty is relatively limited. e probability of feeling a stronger
sense of duty in national or regional elections increases by only 2 or 4 percentage
points when attachment or perceived impact is higher for that level. is is not
surprising, given that even though many people feel more strongly attached to
one level or the other and many also perceive dierences regarding the degree
of policy inuence, an overwhelming majority feel equally dutiful (or not) in
national and regional elections (see Table 4).
Conclusion
Until recently most research on elections has focused on national elections
but that focus has shied somewhat in the last few years, with a growing role
being played by supranational institutions on the one hand and the great push
for decentralisation in many countries on the other. is raises the question
of which elections citizens deem most and least important and whether these
perceptions aect voters’ attitudes, most especially whether they feel that they
have less of a duty to vote in less important elections.
is study makes an important contribution to that discussion on many
fronts. First, we examined citizens’ views about these questions since these
views in the end will shape their behaviour rather than the ‘objective’ reality.
Second, we compared these views across many countries but also across dier-
ent regions within these countries. ird, we looked at three levels of elections,
and thus compared both European and regional elections with national ones,
which is seldom done. Fourth, we analysed citizens’ basic orientation towards
the act of voting ‒ that is, whether they feel that they have less of a moral obli-
gation to vote in dierent types of elections. Fih, and nally, we addressed the
causes behind dierentiated levels of duty in dierent elections.
We tested two individual-level hypotheses drawn from the literature on the
turnout gap. e classical approach portrays national elections as ‘rst-order’
elections because there is more at stake. erefore, all other elections are ‘sec-
ond-order’, arousing weaker feelings of duty to vote. Yet this assumption is ques-
tionable given the transfer of authority to both supranational and subnational
institutions. Another approach suggests that socio-psychological factors, such
as dierences in attachment to territorial levels, may account for dierences
in the belief that voting is a duty. Our research addresses whether a dieren-
tial sense of duty follows from perceptions that a given level of government is
more or less relevant (in terms of policy impact), in line with the ‘second-order’
model, or from a weaker feeling of attachment to a given territorial community,
in line with a socio-psychological perspective.
Downloaded by [95.23.229.64] at 10:07 15 December 2015
16 C. GALAIS AND A. BLAIS
We found that the overwhelming majority of people feel an equally strong
or weak sense of duty to vote in national and regional elections, but that a sub-
stantial minority (about 20 per cent) feel dierently about European elections,
almost all of them expressing a weaker sense of duty to vote in European elec-
tions. We sought to determine whether this weaker sense of duty in European
elections was related to the perception that the European Union has less impact
on people’s well-being than the national government and/or to a lesser feeling
of attachment towards Europe than to their own country. We showed that both
attitudes come into play but that attachment has a greater eect than percep-
tions of policy impact.
Additional analyses suggest that there is a moderator eect of age on the rela-
tionship between attachment and duty: dierential attachment matters more
among the elderly. is nding suggests that younger people are less inclined to
develop dierential duty on the basis of dierential attachment to the political
community. is points to the fact that most young citizens develop inclusive
identities and, as a consequence, they are less prone to believe that they have
less of a duty to vote in some elections. On the other hand, our data do not
support the hypothesis that there is a negative interaction eect between impact
and attachment. e two variables have additive, independent eects on the
propensity to develop a dierentiated sense of duty.
Our research contributes both to the literature on the turnout gap, and to
that on the moral basis of voting. With respect to the rst research strand, we
identify a likely cause of citizens’ dierential voting behaviour across levels of
elections: that they feel they have a weaker duty to vote at one level. We also
examine some sources of that dierential feeling, which is aected by individ-
uals’ attachment to the various territorial communities and their perceptions
of the policy relevance of the dierent levels of government. With regard the
literature on the ‘D term’ or the moral basis of turnout, our results highlight that
most people hold a consistent sense of duty regardless of the level of election,
as this moral aspect is only weakly aected by contextual factors. Furthermore,
when individuals feel a stronger or weaker sense of civic duty towards some
elections, this is due more to aective factors than to rational evaluations of
what is at stake in that election. is nding points to a relationship between
community identication and sense of duty and to the possibility that these
orientations are developed early and stay pretty stable over the life cycle.
Notes
1. e authors include perceived regional government impact and regional identity
as the two main individual-level independent variables in their study of the
determinants of turnout in regional elections.
Downloaded by [95.23.229.64] at 10:07 15 December 2015
WEST EUROPEAN POLITICS 17
2. Indeed it has been shown that citizens who identify with their political
community register and vote more than those who do not (Booth and Seligson
2009), possibly because they have a greater sense of civic duty (Campbell 2006;
Raney and Berdhal 2009).
3. If sense of duty is developed early in life it should be more strongly shaped by
identication with one’s political community than by perceptions of relative
impact of government decisions, which are usually developed later in life. In
other words, if duty is a consequence of political socialisation; dierential
duty should be better explained by dierential identity than by dierential
perceptions of political power.
4. e data have been weighted so that the sample ts the socio-demographic
composition of the population. When analysing the merged database (all
elections, regions, and countries), an additional weight has been created so
that each region counts the same.
5. In Bavaria, a ve-wave panel was conducted, starting before the regional election
in September 2013 and ending aer the European election of 2014. We only
use the rst wave, which tapped respondents’ views about the three election
levels being studied.
6. e questions were not asked consecutively. Questions about impact came
rst, followed – aer a break ‒ by attachment and – again, aer two unrelated
questions ‒ duty. Note that the order of the levels (European, national, regional)
varies across questions.
7. More explicitly, the independent variables are: (1) national attachment versus
European attachment, (2) national attachment versus regional attachment,
(3) national impact versus European impact, and (4) national impact versus
regional impact. Note that we treat these variables as ordinal with equally
spaced scores, then test for linear associations with the dependent variable.
Estimating dierent coecients for the dierent categories of these independent
variables would yield less parsimonious models. Moreover, in order to make
all the dierentiated variables comparable and parallel, the dependent variable
would require the same treatment, which in turn would demand multinomial
logistic estimation, whose results are also more complex to interpret.
8. We performed additional analyses with continuous measures for the duty
(between ‒3 and +3), the attachment and the impact variables (between ‒10
and +10). ese alternative OLS estimations yielded very similar results. We
decided to stick to the collapsed codication because it allowed a simpler display
of the descriptive patterns, as well as of the simulations.
9. All non-dichotomous variables (age, education, time living in the country/
region) have been recoded to range between 0 and 1.
10. Note that we may be underestimating dierences. Disengaged respondents may
have rushed to the end of the survey ‘straight-lining’ (Herzog and Bachman
1981; Malhotra 2008), therefore producing articial consistency. We have
checked how many individuals gave exactly the same scores for each level of
duty, attachment, and impact. e results for the four regions (the two German
and the two Spanish) where questions for the three levels (European, national,
regional) were available yielded an average of 6 per cent, giving the same scores
with respect to duty, attachment, and impact. We conclude that we may be
slightly underestimating dierential duty but that this should not seriously aect
the main ndings.
Downloaded by [95.23.229.64] at 10:07 15 December 2015
18 C. GALAIS AND A. BLAIS
11. Note that this might be pointing either to an age or to a cohort eect. We may
conclude that younger people are more rational. But it is also possible that older
Spaniards, French, or Germans, who came into contact with the idea of the
European Union once their values and political attitudes had crystallised, did
not develop a strong sense of belonging towards the newest level of government.
Our data indeed show a larger gap in civic duty between European and national
elections; as well as in attachment, for this older cohort. Unfortunately, we do
not have the appropriate (panel) data to explore this relationship further.
12. It is also noteworthy that alternative models with xed eects by region yielded
almost identical results.
13. e sole exceptions are Catalonia and Madrid, where the dierential impact
variable does not reach statistical signicance (though having the expected
positive sign).
Disclosure statement
No potential conict of interest was reported by the authors.
Funding
is work was supported by the Social Sciences and Humanities Research Council of
Canada under the Insight Grant programme [grant number 435-2014-0077].
Notes on contributors
Carol Galai s is currently a postdoctoral fellow at the Universitat Oberta de Catalunya
(UOC). She obtained her PhD in political science from the Universitat Pompeu Fabra
and has held postdoctoral fellowships at the Universitat Autonoma de Barcelona and
Université de Montréal. Her research interests focus on political socialisation, civic duty,
political engagement attitudes and local political participation. [carolgalais@gmail.com]
André Blais is Professor in the Department of Political Science at the University of
Montreal. He is a member of the Centre for the Study of Democratic Citizenship, holds
a Research Chair in Electoral Studies, and is the principal co-investigator of ‘Making
Electoral Democracy Work’ (www.electoraldemocracy.com). His research interests
are elections and voting, electoral systems, turnout, and public opinion. [andre.blais@
umontreal.ca]
References
Berg, L. (2007). Multi-Level Europeans. e Inuence of Territorial Attachments on
Political Trust and Welfare Attitudes. Göteborg: University of Gothenburg.
Blais, A. (2000).To Vote or Not to Vote? e Merits and Limits of Rational Choice eory.
University of Pittsburgh Press.
Blais, A. (2010). ‘Making Electoral Democracy Work’, Electoral Studies, 29:1, 169–170.
Downloaded by [95.23.229.64] at 10:07 15 December 2015
WEST EUROPEAN POLITICS 19
Blais, André and Christopher H. Achen (2015). ‘Taking Civic Duty Seriously: Political
eory and Voter Turnout’. Working paper, available at http://www.princeton.edu/
csdp/events/Achen031110/Achen031110.pdf.
Booth, J.A., and M.A. Seligson (2009). e Legitimacy Puzzle in Latin America: Political
Support and Democracy in Eight Nations. Cambridge: Cambridge University Press.
Bruter, Michael (2005). Citizens of Europe? e Emergence of a Mass European Identity.
New York, NY: Palgrave McMillan.
Campbell, D.E. (2006). Why We Vote: How Schools and Communities Shape Our Civic
Life. Princeton, NJ: Princeton University Press.
Castano, Emanuele (2004). ‘European Identity: A Social-Psychological Perspective’, in
Richard K. Herrmann, omas Risse, and Marilynn B. Brewer (eds.).Transnational
Identities: Becoming European in the EU. Oxford: Rowman & Littleeld Publishers,
40–58.
Colomer, J.M., and A. Padró-Solanet (1992). ‘Espacio político Ideológico y temas de
campaña’ (El ejemplo de las elecciones autonómicas de Cataluña de 1992) [Political-
Ideological Space and Campaign Issues (the Example of the 1992 Catalan Regional
Elections)], Revista de estudios políticos, 78, 131–60.
Craig, P. (2010).e Lisbon Treaty: Law, Politics, and Treaty Reform. Oxford: Oxford
University Press.
Dennis, Jack, ed. (1973). Socialization to Politics: A Reader. New York, NY: Wiley.
Dworkin, Ronald (1978). Taking Rights Seriously, Vol. 136. Cambridge, MA: Harvard
University Press.
Franklin, M. (2001). ‘e Dynamics of Electoral Competition’, in Laurence Leduc,
Richard Niemi, and Pippa Norris (eds.), Comparing Democracies 2: Elections and
Voting in Global Perspective. ousand Oaks, CA: Sage, 216–35.
Graham, Jesse, et al. (2011). ‘Mapping the Moral Domain’, Journal of Personality and
Social Psychology, 101: 2, 366–85.
Haidt, J. (2003). ‘e Moral Emotions’. in R.J. Davidson, K.R. Scherer, and H.H.
Goldsmith (eds.). Handbook of Aective Sciences. Oxford: Oxford University Press,
852–70.
Henderson, A., and N. McEwen (2015). ‘Regions as Primary Political Communities:
A Multi-Level Comparative Analysis of Turnout in Regional Elections’, Publius: e
Journal of Federalism, 45:2, 189–215.
Herzog, A.R., and J.G. Bachman (1981). ‘Eects of Questionnaire Length on Response
Quality’, Public Opinion Quarterly, 45:4, 549–59.
Homann-Martinot, V. (1992). ‘La participation aux élections municipales dans les
villes françaises’ [Participation in Municipal Elections in French Cities], Revue
française de science politique, 4:1, 3–35.
Hooghe, L., and G. Marks (2001). Multi-Level Governance and European Integration.
Boulder, CO: Rowman & Littleeld.
Keating, Michael (2001). Plurinational Democracy Stateless. Nations in Post-Sovereignty
Era. Oxford: oxford University Press.
Klandermans, B., J.M. Sabucedo, and M. Rodriguez (2004). ‘Inclusiveness of
Identication among Farmers in the Netherlands and Galicia (Spain)’, European
Journal of Social Psychology, 34:3, 279–95.
Lewis-Beck, M.S., H. Norpoth, and W. Jacoby (2009). e American Voter Revisited.
Ann Arbor, MI: University of Michigan Press.
Liñeira, R., and J.M. Vallès (2014). ‘Abstención diferencial en Cataluña y en la Comunidad
de Madrid: explicación sociopolítica de un fenómeno urbano’ [Dierential Abstention
in Catalonia and the Community of Madrid: A Socio-political Explanation of an
Urban Phenomenon], Revista Española de Investigaciones Sociológicas, 146:1, 69–91.
Downloaded by [95.23.229.64] at 10:07 15 December 2015
20 C. GALAIS AND A. BLAIS
Malhotra, N. (2008). ‘Completion Time and Response Order Eects in Web Surveys’,
Public Opinion Quarterly, 72:5, 914–934. doi:10.1093/poq/nfn050.
McManus-Czubińska, C., W.L. Miller, R. Markowski, and J. Wasilewski (2003).
‘Understanding Dual Identities in Poland’, Political Studies, 51, 121–143. doi:
http://dx.doi.org/10.1111/1467-9248.00416.
Medrano, J.D., and P. Gutiérrez (2001). ‘Nested Identities: National and European
Identity in Spain’, Ethnic and Racial Studies, 24:5, 753–78.
Mols, F., and M. Weber (2013). ‘Laying Sound Foundations for Social Identity eory-
Inspired European Union Attitude Research: Beyond Attachment and Deeply Rooted
Identities’, JCMS. Journal of Common Market Studies, 51:3, 505–21.
Montero, J.R., and Joan y Font (1991). ‘El voto dual: lealtad y transferencia de voto
en las elecciones autonómicas’ [Dual Voting: Loyalty and Vote Transference in the
Regional Elections], in VV.AA. Equip De Sociologia Electoral (eds.), L’electorat català
a les eleccions autonómiques de 1988: opinions, actituds i comportaments. Estudis
Electorals/10. Barcelona: Fundación Jaume Boll, 183–214.
Morlan, R.L. (1984). ‘Municipal Vs. National Election Voter Turnout: Europe and the
United States’, Political Science Quarterly, 99:3, 457–70.
Piris, J.C. (2010). e Lisbon Treaty: A Legal and Political Analysis. Cambridge:
Cambridge University Press.
Prinz, J.J. (2008). ‘Is Morality Innate?’ in W. Sinnott-Armstrong (ed.), Moral Psycholog y,
vol. 1: Evolution of Morals. Cambridge, MA: MIT Press, 367–406.
Rallings, C., and M. rasher (2007). ‘e Turnout ‘Gap’and the Costs of Voting–A
Comparison of Participation at the 2001 General and 2002 Local Elections in
England’, Public Choice, 131:3–4, 333–344.
Raney, T., and L. Berdahl (2009). ‘Birds of a Feather? Citizenship Norms, Group Identity,
and Political Participation in Western Canada’, Canadian Journal of Political Science,
42:1, 187–209.
Rawls, John (1971). A eory of Justice. Cambridge, MA: Harvard University Press.
Reif, K., and H. Schmitt (1980). ‘Nine Second‐Order National Elections–A Conceptual
Framework for the Analysis of European Election Results’, European Journal of
Political Research, 8:1, 3–44.
Riba, C. (2000). ‘Voto dual y abstención diferencial. Un estudio sobre el comportamiento
electoral en Cataluña’ [Dual Vote and Dierential Abstention. A Study of Voting
Behavior in Catalonia], Revista Española de Investigaciones Sociológicas, 91, 59–88.
doi:http://dx.doi.org/10.2307/40184275.
Riera, P. (2011). ‘Dierential Abstention in the Basque Country and Catalonia’, Revista
De Estudios Políticos, 154, 139–73.
Riera, P. (2012). ‘La abstención diferencial en la España de las autonomías. Pautas
signicativas y mecanismos explicativos’ [Dierential Abstention in the Spain of
Autonomies. Signicant Patterns and Explanatory Mechanisms], Revista Internacional
de Sociología, 70:3, 615–42.
Riker, W.H., and P.C. Ordeshook (1968). ‘A eory of the Calculus of Voting’, American
Political Science Review, 62:1, 25–42.
Rosenstone, S.J., and J.M. Hansen (1993). Mobilization, Participation, and Democracy
in America. Longman Classics Edition. Longman Publishing Group.
Santana-Leitner, Andrés (2008). Expresividad, cálculo y movilización en la decisión de
votar [Expressiveness, Calculus and Mobilization in the Decision to Vote]. Madrid:
Centro de Estudios Avanzados en Ciencias Sociales, Instituto Juan March.
Sears, D.O., C.P. Hensler, and L.K. Speer (1979). ‘Whites' Opposition to “busing”: Self-
interest or Symbolic Politics?’, American Political Science Review, 73:2, 369–84.
Downloaded by [95.23.229.64] at 10:07 15 December 2015
WEST EUROPEAN POLITICS 21
Schakel, Arjan H., and Charlie Jeery (2013). ‘Are Regional Elections Really ‘Second-
Order’ Elections?’ Regional Studies, 47:3, 323–41.
Shively, W.P. (1992). ‘From Dierential Abstention to Conversion: A Change in Electoral
Change, 1864–1988’, American Journal of Political Science, 309–30.
Tajfel, Henri (1978). Dierentiation between Social Groups: Studies in the Social
Psychology of Intergroup Relations. London: Academic Press.
Terry, Deborah J., Michael A. Hogg, and Katherine, M. White (1999). ‘e eory of
Planned Behaviour: Self-Identity, Social Identity and Group Norms’, British Journal
of Social Psychology, 38:3, 225–44.
Turner, J.C., P.J. Oakes, S.A. Haslam, and C. McGarty (1994). ‘Self and Collective:
Cognition and Social Context’, Personality and Social Psychology Bulletin, 20, 454.
Vallés, J.M. (2009). ‘La abstención ‘diferencial’: una nota sobre los casos de Cataluña y
de la Comunidad Autónoma de Madrid’ [e ‘dierential’ Abstention: A Note on the
Cases of Catalonia and the Autonomous Community of Madrid], Revista Española
de Ciencia Política, 21, 93–105.
Van der Eijk, Cees, and Mark Franklin, eds. (1996). Choosing Europe? e European
Electorate and National Politics in the Face of Union. Ann Arbor: University of
Michigan Press, 28
Van Deth, J. W., S. Abendschön, and M. Vollmar (2011). ‘Children and Politics: An
Empirical Reassessment of Early Political Socialization’, Political Psychology, 32:1,
147–74.
Appendix 1. The list of surveys
No. Region Country Type of election Time N
1 Lucerne Switzerland National (Fed.) Oct. 2011 1,108
2 Lucerne Switzerland Regional (Cantonal) April 2011 1,972
3 Zurich Switzerland National (Fed.) Oct. 2011 1,057
4 Zurich Switzerland Regional April 2011 1,912
5 IDF France National June 2012 966
6Provence France National June 2012 983
7 Catalonia Spain National Nov. 2011 951
8 Catalonia Spain Regional Nov. 2012 993
9 Madrid Spain National Nov. 2011 976
10 Lower Saxony Germany National (Fed) Sept. 2013 975
11 Lower Saxony Germany Regional (State) Jan. 2013 983
12 Bavaria Germany Regional Sept. 2013 5,906
13 Ontario Canada Regional Oct. 2011 1,309
14 Quebec Canada Regional Sept. 2012 990
15 Provence France European May 2014 1,039
16 IDF France European May 2014 975
17 Lower Saxony Germany European May 2014 978
18 Catalonia Spain European May 2014 985
19 Madrid Spain European May 2014 974
20 IDF France Local March 2014 1,208
21 Provence France Local March 2014 725
Total 27,965
IDF: Île de France.
Downloaded by [95.23.229.64] at 10:07 15 December 2015
22 C. GALAIS AND A. BLAIS
Appendix 2. Question wording
Duty
Is voting a duty or a choice at the European level of government? Q44C
1 = Duty; 2 = Choice; 9 = Don’t know
(if ‘duty’) How strongly do you feel that voting is a duty at the: European level of
government? Q45D: 1 = Not very strongly; 2 = Somewhat strongly; 3 = Very strongly;
9 = Don’t know
Is voting a duty or a choice at the national level of government? Q44B
1 = Duty; 2 = Choice; 9 = Don’t know
(if ‘duty’) How strongly do you feel that voting is a duty at the national level of gov-
ernment? Q45B: 1 = Not very strongly; 2 = Somewhat strongly; 3 = Very strongly;
9 = Don’t know
Is voting a duty or a choice at the regional level of government?
1 = Duty; 2 = Choice; 9 = Don’t know
(if ‘duty’) How strongly do you feel that voting is a duty at the regional level of gov-
ernment? Q45A: 1 = Not very strongly; 2 = Somewhat strongly; 3 = Very strongly;
9 = Don’t know
Impact
How much inuence do policies of the following level of government have on the
well-being of you and your family: European Union?
0 = Very small impact; 10 = Very big impact; 99 = Don’t know
How much inuence do the policies of the following level of government have on the
well-being of you and your family: national government?
0 = Very small impact; 10 = Very big impact; 99 = Don’t know
How much inuence do the policies of the following level of government have on the
well-being of you and your family: regional government?
0 = Very small impact; 10 = Very big impact; 99 = Don’t know
Attachment
How attached are you to the European Union?
0 = Not attached at all; 10 = Very strongly attached; 99 = Don’t know
How attached are you to Canada/France/Switzerland/Spain/Germany?
0 = Not attached at all; 10 = Very strongly attached; 99 = Don’t know
How attached are you to (name of the province/region)?
0 = Not attached at all; 10 = Very strongly attached; 99 = Don’t know
Downloaded by [95.23.229.64] at 10:07 15 December 2015
WEST EUROPEAN POLITICS 23
Appendix 3. Eects of dierential attachment and impact by
region
(a) National versus European
IDF Provence Catalonia Madrid Low Sax Bavaria
Attachment 0.44*(0.21) 0.83**(0.12) 0.5**(0.08) 0.64**(0.14) 0.33**(0.08) 0.16**(0.06)
Impact 0.28**(0.09) −0.04(0.09) 0.39**(0.09) 0.20*(0.09) 0.36**(0.08) 0.26**(0.06)
Standard errors in parentheses. *p < 0.01; **p < 0.001. Cell entries are ordered logit coefficients. The
estimation accounts for the controls shown in Table 2 (sex, age, etc.).
(b) National versus Regional
Lucerne Zurich Catalonia Madrid Low Sax Bavaria Ontario Quebec
Attachment 0.41**
(0.11)
0.26**
(0.1)
1.16**
(0.12)
0.58**
(0.17)
0.54**
(0.1)
0.19**
(0.07)
1.0**
(0.3)
0.54**
(0.19)
Impact 0.37**
(0.10)
0.58**
(0.1)
0.11
(0.09)
0.12
(0.14)
0.23*
(0.09)
0.34**
(0.07)
1.1**
(0.26)
0.56**
(0.21)
Standard errors in parentheses. *p < 0.01; **p < 0.001. Cell entries are ordered logit coefficients. The esti-
mation accounts for the controls shown in Table 5 (sex, age, etc.).
Downloaded by [95.23.229.64] at 10:07 15 December 2015
