Available via license: CC BY 4.0
Content may be subject to copyright.
Vol.:(0123456789)
Acta Politica
https://doi.org/10.1057/s41269-024-00358-2
ORIGINAL ARTICLE
Reconceptualizing “feeling represented”: Anew approach
tomeasure howfeelings ofpolitical representation are
constructed
AndreaVik1,2,3 · PieterdeWilde3,4
Accepted: 26 July 2024
© The Author(s) 2024
Abstract
In political representation, the concept of feeling represented has garnered signifi-
cant attention in recent years. However, the prevailing conceptualization of feeling
represented is limited in its scope and has effectively conflated a perception of being
represented with feeling represented. We argue that such a conflation undermines
our understanding of political representation, especially in the face of the construc-
tivist turn in representation. Constructivist scholars have theoretically elaborated
that representation results from a process in which a linkage is created between rep-
resented and representative. We argue that feeling represented is an essential part of
establishing this representative relationship. By drawing on the recent constructivist
turn in representation and applying key insights from psychology, we develop and
test a reconceptualization of feeling represented. Through an online survey, Ameri-
can prolific users (
N=605
) were asked a number of questions related to feeling
represented after witnessing representative claims by unelected representatives.
Through both exploratory and confirmatory factor analysis, we establish that feeling
represented is a multidimensional concept consisting of instrumental, expressive,
and aversive dimensions. Thus, feeling represented is a subjective state consisting
of clearly emotional components, and it can be measured as a response to unelected
representation. Acknowledging this goes beyond the current constructivist account
of representation and sets the stage for future research.
Keywords Feeling represented· Representation· The constructivist turn·
Psychology· Affect
Introduction
“A man (sic) is represented if he feels that he is, and not if he does not (Pit-
kin 1967,p. 9)”.
Extended author information available on the last page of the article
A.Vik, P.de Wilde
As argued by Pitkin in her seminal work (Pitkin 1967), political representa-
tion has an inherently psychological component (pp. 9–10: see opening quote).
While Pitkin acknowledges the importance of feeling represented from a social
psychology perspective, the thrust of her argument focuses on objective repre-
sentation instead. Subsequently, representation scholars have focused primarily
on objective representation as responsiveness, which dovetails with a more gen-
erally pervasive focus on objectivity and rationality in political science since the
rational voter perspective (Huddy etal. 2013). Existing conceptualizations of
‘feeling represented’ actually measure ‘perceptions of representation’, prompt-
ing citizens to reflect on whether they consider elected representatives to rep-
resent them in a substantive manner (Holmberg 2020; DeMulder 2022; Dvir-
Gvirsman etal. 2022). Yet, political psychology has shown us that many aspects
of political beliefs and behavior are at least partly built on emotional founda-
tions as opposed to only rational reflections on a substantive ideological match
(Marcus 2003; Redlawsk and Mattes 2022). As long as we do not take the ‘feel-
ing’ part of ‘feeling represented’ seriously, we may not fully comprehend how
political representation works.
Mapping and understanding feelings of representation is imperative because
even in countries with well-functioning institutions of representative democ-
racy and proportional electoral systems, many citizens are dissatisfied with the
way democracy works (Aarts and Thomassen 2008; Linde and Ekman 2003).
They may have voted and the candidates they have voted for may have gotten
into office, and yet they do not consider themselves to be represented in the sys-
tem. Being represented, apparently, does not automatically translate into feeling
represented, just as feeling represented does not necessarily translate into being
represented.
The lack of research on feeling represented is even more striking in light of the
recent constructivist turn in political representation (Disch 2011, 2015; Guasti
and Geissel 2019; Mansbridge 2003; Saward 2010, 2006). This constructivist
turn essentially reverses the relationship between authorization and representa-
tion in the standard approach of political representation, arguing that represen-
tation is the result of claim-making, which is a dynamic and evolving part of a
communication process that connects representatives as makers of claims to the
represented as audience and judges of such claims.
This article sets out to develop a novel measurement tool that acknowledges
the psychological nature of feeling represented. We consider this a subjective
state that occurs after witnessing representative claims. The need to develop this
measurement tool stems from a conceptualization of “feeling represented” as a
key mechanism between citizens’ exposure to representative claims and their
behavioral response in the form of acceptance or rejection of such claims. Peo-
ple witness a representative claim, feel or do not feel represented through it, and
subsequently express political beliefs and behavior. While we consider feeling
represented as a key mechanism between exposure to a claim and an observable
behavioral reaction in the form of acceptance or rejection, it is also likely that
many citizens do not translate their feelings into behavior. Hence, focusing on
observable acceptance or rejection of claims to determine the legitimacy and
Reconceptualizing “feeling represented”: Anew approach…
effect of representative claims risks overlooking the many citizens who witness
and process the claim but do not observably react. These citizens may still feel
represented and thus impact the legitimacy of democratic systems.
In the spirit of the constructivist turn in representation, we measure feelings
of representation as a result of exposure to claims made by unelected representa-
tives. These are activists, opinion makers, and other individuals who formulate
policy demands on behalf of certain constituencies in public, thus making rep-
resentative claims. It is a hard test because many may not consider these people
as representatives. If we succeed in developing a tool that can measure feelings
of representation as a result of exposure to representative claims by unelected
representatives, it is likely that these tools can be used to measure feelings of
representation as a result of witnessing claims by both elected and unelected
individuals and that these feelings are pervasive among citizens, thus forming a
key piece of the puzzle in understanding the legitimacy of our systems. We thus
proceed to ask the research question: “How can we measure whether and to what
extent exposure to representative claims by unelected representatives makes
people feel represented?”
To answer this research question, we draw on both research in representation
and psychology. Based on a large and diverse body of literature, we derive 25
different possible ways of asking people whether they feel represented. Follow-
ing exposure to constructed representative claims by unelected representatives
in the USA, we asked all of these 25 questions to 605 American respondents
in an online survey. Using exploratory factor analysis (EFA) followed by con-
firmatory factor analysis (CFA), we establish three dimensions of feeling repre-
sented; an instrumental dimension (factor 1) referring to agreement, an expres-
sive dimension (factor 2) referring to feeling a deep connection, and an aversive
dimension (factor 3) referring to feeling misrepresented and disconnected. We
demonstrate that feeling represented is a three-dimensional concept and we pro-
vide tools in the form of concrete survey questions to tap each of these three
dimensions. This opens up new avenues for research in an affective turn in the
study of political representation.
Theoretical framework
Reconceptualizing feeling represented: themissing link (in theconstruction
ofrepresentation)
Much of the literature on political representation denounces emotions as easily
manipulated by demagogues and as impeding rational thinking (Pitkin 1967). Build-
ing on a selective understanding of Pitkin, the standard account of political repre-
sentation sees representation as responsiveness to public opinion.1 Governments
1 Responsiveness is one of Pitkin’s contributions and one that is often referenced in the literature, but
this does not necessarily reflect her larger contributions to representation.
A.Vik, P.de Wilde
need to enact the will of the people or that of the median voter. Whether or not
people feel represented is not essential. The constructivist turn in political represen-
tation challenges this perspective (Disch 2011). Central to the constructivist turn is
the idea that representation is not merely the result of electoral politics but rather a
claim-making process (Saward 2010). In other words, we have gone from ‘acting in
the interest of the represented, in a manner responsive to them’ to a ‘dynamic claim-
making process’ (Saward 2010). As explained by Disch (2011), preferences are con-
stituted in interactive communication between (would-be) representatives and the
represented. Constituencies and public opinion are fluid, not fixed, where emotions
are part and parcel of the fluid process of their construction. “Representative rela-
tionships select... wants, desires, discontents, values, and judgments..” (Castiglione
and Warren 2019). Thus, this constructivist account of representation inherently
opens up for affective considerations even if they have not been empirically studied
much yet (Knops 2022).
We argue that before citizens can act out an acceptance or rejection of a rep-
resentative claim, they must first affectively and cognitively process the claim, as
all communication causes both cognitive and affective processing (Potter and Bolls
2012). We pick up how feeling represented is an essential part of this process-
ing, which aligns with recent research showing that feeling represented fluctuates
through political communication (Dvir-Gvirsman etal. 2022; Kaplan et al. 2023).
In conclusion, we lack a measurement of feeling represented that both acknowledges
the constructivist turn and research from political psychology.
The case ofunelected representation
Seeing representation as a dynamic process of claim-making and not, for example,
as a static fact of electoral politics can help us to make sense of major changes in the
daily politics of representation, such as the role played by unelected representatives.
This has also been called the expansionism of representation (Wolkenstein and Wra-
til 2020), and is considered one of three components many constructivist scholars
agree on (Wolkenstein 2021). When representation results from a process of claim-
making rather than elections, unelected representatives can be political representa-
tives as much as elected ones. After all, anyone can take the stage, claim to speak
on behalf of a larger group of people, and get a supportive following as a result. In
this way, the constructivist turn inherently invites us to study unelected representa-
tion. This practice of unelected representation may have strong implications for the
democratic functioning of political systems, as it circumnavigates key features of
liberal democracy (Tormey 2015; Montanaro 2017). Representative claims by une-
lected representatives can be very diverse and reach across the political spectrum, as
illustrated in Fig.1. In recent work, Saward (2016) states that our current crisis of
democracy suggests that the “time is ripe” to understand more about representative
claims made by unelected representatives. By limiting our study to the realm of une-
lected representation, we develop tools to measure feeling represented in an unor-
thodox environment, thus providing a hard test for our measurement instruments. It
simultaneously sheds light on the theoretical assumption in the constructivist turn
Reconceptualizing “feeling represented”: Anew approach…
that people may feel represented due to exposure to representative claims by une-
lected representatives.
Why constructing political representation ishot: defining feelings andtheir role
inpolitics
As already highlighted when the term emotion was introduced to the English lan-
guage, “The exact meaning of the term emotion, it is difficult to state in any form of
words” (Brown 1836,p. 102). The term emotions is still highly debated in psychol-
ogy. What there is at least some agreement about is that emotional responses can be
elicited from stimuli such as political communication and can be expressed verbally,
behaviorally, and physiologically (Bradley and Lang 2000; Potter and Bolls 2012).
Emotions are considered to be rooted in basic biological and human motivation and
can motivate behavior and attitudes (Potter and Bolls 2012). Feelings are when the
emotional experience has been labeled in our brain, an emotional response we rec-
ognize to pertain to a certain feeling (Shouse 2005). Affect becomes an umbrella
term, that captures emotional responses, moods, and feelings and emotions, affect,
and feelings are thus all allied terms. We can categorize many emotional responses
with the help of the circumplex model of affect. This model suggests, through fac-
tor analysis of behavioral, self-report, and physiological data, a strong two-factor
solution composed of arousal and valence as two dimensions of emotional response
Fig. 1 Examples of Representative Claims by Unelected Representatives (Various Screenshots)
A.Vik, P.de Wilde
(Bolls etal. 2001). Arousal refers to feeling activated, whereas valence refers to the
feeling of pleasure or displeasure (Lang etal. 1999). These two factors are present
in many discrete emotional responses such as fear, anxiety, and anger—emotions
known to drive political attitudes and behavior (Druckman and McDermott 2008;
Valentino et al. 2011; Valentinoet al. 2009). In other words, we know that many
emotions are felt in terms of their valance and arousal.
Sixty years ago, Abelson (1963) created the “hot cognition” hypothesis, which
states that all sociopolitical phenomena are affectively charged concepts in our
brain. Lodge and Taber (2005) tested the hot cognition hypothesis and found evi-
dence illustrating the affective processing of political leaders, groups, and issues.
Interestingly, for more complex political phenomena, the affective processing was
faster for those with higher levels of political knowledge. Although cognitive and
affective processing are often traditionally thought of as two independent systems
(Zajonc 1980), in the last decade, much research has shown this is not the case. Both
thoughts and feelings are constructed in our brain (Barrett 2017), and research sug-
gests we can vote according to our policy preferences by only remembering how
we felt about political leaders (Coronel etal. 2012). Affective processing happens
in milliseconds after exposure to stimuli, and these initial affective processes often
color our attitude formation, a process known as affective contagion (Taber and
Lodge 2006; Erisen etal. 2014). But our prior attitudes also determine the emotional
response we experience (Barrett 2017), and even non-political physiological states
such as heart rate have even been found to influence leadership preferences (Tsakiris
etal. 2021). In short, we think about our feelings and feel things about our thoughts,
which is also true for political phenomena. In other words, our thoughts and feelings
are highly interconnected.
In fact, dual processing takes place in all communication, as we benefit from
cognitive processing, as well as emotional or affective processing (Potter and Bolls
2012). Bakker etal. (2021) have illustrated that simple policy demands cause physi-
ological emotional responses, that congruent policy demands lead to heightened
levels of physiological arousal, and that incongruent policy demands lead to nega-
tive affect. To highlight the democratic implications of such findings, Aarøe (2011)
have shown that emotional responses from political communication have been found
to influence policy support. Specifically, episodic frames, i.e., stories with an emo-
tional or human interest frame, cause a more intense emotional response, which in
turn strengthens the effects such frames have on policy support (Aarøe 2011). In
sum, policy demands and other political stimuli evoke emotional responses, and
emotions influence policy support and other political outcomes. Our feelings and
our political preferences are highly interwoven.
We tie these psychological considerations together with the constructivist turn
of political representation and create a circular model of claim-making with our
added mechanism “feeling represented” in Fig.2. Note that our goal is to provide a
measurement tool of feeling represented, not to identify the causes or consequences
of feeling represented. With our measurement tool, however, future research can
answer such questions.
Since every individual has a different set of previous sensations from which to
draw when interpreting and labeling their feeling (i.e., previous situations in which
Reconceptualizing “feeling represented”: Anew approach…
they felt represented by leaders), feelings are personal and biographical (Shouse
2005). We argue that feeling represented is somewhat labeled in citizens’ minds;
they know if they feel represented or not.
Considering the scope of this article, our focus is thus on conscious emotional
processes through self-reports. Feeling represented should, therefore, be a positive
and intense feeling. We would expect that individuals will experience varying levels
of self-reported arousal and valance in response to representative claims and that
such an emotional response may be a part of feeling represented. However, the pro-
cess of how citizens label the feeling of being represented will also have other affec-
tive components beyond the extent to which the claim makes them feel bad or good.
Furthermore, to feel represented and to feel this deeply does not necessarily indicate
an irrational process. Based on these considerations, feeling represented is recon-
ceptualized as a subjective feeling state of political representation, which is poten-
tially deeply felt and embodied, that arises after witnessing representative claims. To
operationalize this conceptualization, we draw both on an instrumental perspective
of feeling represented, which is the current conceptualization of the concept in the
political representation literature, and then introduce the new expressive perspective
in which we draw parallels to political psychology.
The current conceptualization offeeling represented
As highlighted by Pitkin (1967), feeling represented is in many ways a social-psy-
chological concept, yet she subsequently chose not to focus on feelings, stating “...I
will ask not what causes people to have a psychological feeling of being represented
but what reasons can be given for supposing someone or something is being rep-
resented” (p. 10). In her wake, current conceptualizations of “feeling represented“
concern mostly aspects of substantive representation. Substantive representation
Fig. 2 Illustration of how we have reconceptualized feeling represented to fit within the construction of
representation
A.Vik, P.de Wilde
refers to sharing issue positions and is rooted in an instrumental perspective in poli-
tics, which is what most of the current literature is concerned with so we deem it a
logical starting point to understand feeling represented.
Notably, Holmberg (2020) and DeMulder (2022) have recently put forward their
conceptualizations of feeling represented. Holmberg (2020) analyzes whether peo-
ple feel represented by a party and party leaders in a range of countries. He draws
on data from the CSES survey in the years 2001–2004. Questions asked were as fol-
lows: “Would you say that any of the parties in (country...) represent your points of
view reasonably well?” and “Irrespective of what you think about the parties, would
you say there is any party leader that represents your views reasonably well?” These
are dichotomous yes/no questions without ‘don’t know’ alternatives. Similarly,
DeMulder (2022) focus on “feelings of being substantively represented”, finding
that this is a multidimensional concept consisting of listening and knowing versus
acting and succeeding.
Both Holmberg (2020) and De Mulder (2022) conceptualizations conflate the
concepts of “being,” “perceptions,” and “feeling.” The use of the term feeling sug-
gests a subjective and emotional experience, whereas the term being implies a more
objective state of existence. Holmberg (2020) argues “Note that—strictly speak-
ing—the interview questions ask about perceptions of representation, not feelings of
representation.” DeMulder (2022) similarly argues that his understanding of feeling
represented may be understood in line with the notion of representation that Pit-
kin (1967) prioritizes: as feeling substantively represented. The basic assumption is
that people feel represented—subjectively—when they are objectively represented
in terms of policy preferences. This is built on considerations from democratic the-
ory stating that good political representation entails that opinions and preferences
held in society need to be reflected among elected officials in order for them to be
enacted into policy (Dahl 1971). These conceptualizations are, therefore, important
advancements in how we understand how citizens may understand if they are being
represented, but ignore a key component of constructivist perspectives on represen-
tation, which emphasizes that representation—including feeling represented—is
constructed through interaction between representatives and represented and thus
inherently subjective, which literature in political psychology also supports.
In addition to feeling substantively represented, we also highlight that it is prob-
able that citizens may feel descriptively represented. Where the English word ‘repre-
sentation’ does not directly allow for fine-grained differentiation, Pitkin (1967) drew
on the German words ‘vertreten’ (to act for) and ‘darstellen’ (to stand for) to con-
ceptualize the difference between substantive and descriptive representation, respec-
tively. Descriptive representation exists when elected parliaments and governments
reflect the microcosmos of society in terms of key demographics such as gender,
age, and ethnicity. Arguments that we need a quota for women in parliament build
on this descriptive perspective on political representation (Mansbridge 1999; Ruedin
2015). So, feeling descriptively represented should be related to whether citizens
feel that the representative is similar to them.
Reconceptualizing “feeling represented”: Anew approach…
Emphasizing thefeeling ofpolitical representation
In addition to the instrumental process that unfolds in which citizens consider
whether what representatives say and do match their preferences and expectations,
there is also an expressive process that unfolds in which citizens assess how they
identify with the maker of the message and how the message makes them feel.
Coleman (2005) is the closest to such a psychological account of feeling repre-
sented, and his work states that “Models of citizenship constructed around people’s
instrumental interest and normative duty to participate have tended to miss the expe-
riential and affective elements of what it means to be a civic actor, as opposed to a
political consumer or free-floating ego” (p. 198). In other words, Coleman (2005)
argues that a sense of connection is an overlooked key element of representation
and, as such, connectedness is essential to feeling represented. In a world increas-
ingly made up of loosely connected individuals and plural channels of communica-
tion, there can be many possible alternatives to the classic representative relationship
between a voter and their elected representative. Through qualitative and quantita-
tive approaches, Coleman (2005) defined representation as a four-dimensional con-
cept: closeness vs. distance; mutuality vs. detachment; coherence vs. exclusivity;
and empathy vs. aloofness (Coleman 2005).
Dvir-Gvirsman etal. (2022) build on Coleman (2005) and create a scale of feel-
ing represented. They argue that connection happens through online communica-
tion between politicians and citizens and that this connection can be measured with
items from substantive, descriptive, symbolic, and formalistic representation. Dvir-
Gvirsman etal. (2022) measure feeling represented subsequently to seeing online
communication from politicians. Their research illustrates that this concept is a state
that fluctuates through communication, lending much merit to understanding feeling
represented as an essential component in the construction of representation, which
we will get back to later. What is still missing is whether feeling represented is only
constituted from items that reflect substantive, descriptive, symbolic, and formalis-
tic representation. So far, the most compelling empirical analyses do not yet fully
embrace feeling represented as a psychological construct.
Urbinati (2006) states that in the constructivist turn, political representation is
considered to “facilitate the formation of political groups and identities” (p. 37).
We know that people identify—a form of psychological attachment—with differ-
ent parties, political leaders, and other political groups and that this matters greatly
for political mobilization (Huddy and Bankert 2017; Kunst etal. 2019; Swannetal.
2012). This form of identification should not be conflated with similarities in recog-
nizable traits [see StaehrHarder (2020)]. Although there is evidence that enduring
identification with a party or a political leader is strengthened by social affiliation to
gender, ethnicity, and religion (Greene 2004; Huddy and Bankert 2017), identifying
with a group, leader, or cause is a unique form of psychological attachment.
Our identity formations are situated, cognitive, and emotional (Epitropaki etal.
2017). Politicians, celebrities, and other leaders often communicate their own identi-
ties in order to get citizens to identify with them (Kreiss etal. 2020). Considering
that party identification is a key variable in explaining vote change, political engage-
ment, motivated reasoning, and the influence of political leaders (Dalton 2016), it
A.Vik, P.de Wilde
is arguably a good and common strategy to facilitate the formation of identities,
as suggested by Urbinati (2006). Specifically, in expressive partisanship theory,
Huddy etal. (2015) benefit from an expressive model of partisanship and find that
this model accounts better for citizens’ campaign involvement compared to instru-
mental measures such as strong stances on subjectively important policy issues,
the strength of ideological self-placement, or a measure of ideological identity.
The expressive multi-item partisan identity scale was built on considerations from
social identity theory (Huddy etal. 2015; Huddy and Bankert 2017). Social identity
theory states that social identity is a person’s knowledge or feeling that he or she
belongs to a social category or group, which forms the social part of their self-con-
cept (Stets and Burke 2000). Here, we draw parallels to political representation (as
does Disch (2021), in her Laclauian perspective on representation as mobilization).
In particular, we draw on the expressive partisanship theory to develop and empiri-
cally explore expressive components of feeling represented, i.e., a form of emotional
attachment to the representative and their claim.
Another identity theory is personal identification, which captures the extent to
which one defines oneself through another individual, such as one’s leader, whose
values and perspectives are perceived as aligned with one’s own (Epitropaki etal.
2017; Kunst etal. 2019). Identification with a leader or a group can become so
strong that it becomes a fused identity. Swann and Buhrmester (2015) explains that
identity fusion refers to a visceral feeling of oneness with a group or individual—
which is a psychological state that could be similar to feeling represented. It clearly
resonates with the importance of ‘connection’ Coleman (2005) ascribed to feel-
ing represented by others. Research on identity fusion has established that identity
fusion with a leader is distinct from identity fusion with a group and has distinct
consequences for our behavior (Kunst etal. 2019). People can both be fused with the
maker of the claim and the group that is the object of the claim. As such, fusion with
an individual, a representative, for instance, is conceptually different from the con-
structs of social identification, leader identification, or identity fusion with groups—
although they are all related (Kunst etal. 2019). Any of these identity fusion con-
cepts could be part of or influence the phenomenon of feeling represented.
As noted by Holmberg (2020), many do not feel represented in current politi-
cal systems. Coleman (2005) four dimensions of political representation always had
an inverse, finding that feelings of distance, detachment, exclusivity, and aloofness
also matter in representation. Such feelings of misrepresentation can be considered
the inverse of feeling represented and important to consider in the construction of
representation, which often is not successful. Further evidence that misrepresenta-
tion is an important component in the field of feeling represented is the documenta-
tion of the frequent occurrence of claims of misrepresentation in the public sphere
(i.e., “You do not speak for me”) (Guasti and Almeida 2019). Research suggests that
the features of media messages (i.e., ingroup exemplars) can influence intergroup
relations by altering the emotions people feel toward the outgroup (Atwell Seate
and Mastro 2015; AtwellSeate etal. 2018). So, this identity can be strengthened
or weakened through political communication, which is what we expect to be the
case for feeling represented (or misrepresented) as well. Representative claims may,
Reconceptualizing “feeling represented”: Anew approach…
therefore, not only lead citizens to feel represented but also to feel misrepresented,
an aversive state that denounces the representative claim.
In conclusion, feeling represented could be a form of emotional attachment with
the maker of the claim, with the cause or movement they speak for, or an emotional
attachment with the articulated constituency. It is likely that this feeling is both
instrumental (i.e., I agree with this claim) and expressive (i.e., I feel attached to this
maker/cause/group). An expressive model of political representation would likely
have different predictive power to, for instance, extreme pro-group behavior, such as
joining protests (Kunst etal. 2019; Swannetal. 2012; Swann and Buhrmester 2015).
Thus, opening up for expressive components of feeling represented can open up new
lines of inquiry into challenges we are seeing in political representation.
Method
Measurement instrument: feeling represented
To capture feeling represented, we have derived several items from concepts relating
to different aspects of feeling represented as highlighted in our theoretical frame-
work; this is illustrated in Table1. Interestingly, we see some overlap in these dif-
ferent schools of thought as many questions are also similar. For instance, connect-
edness by Coleman (2005) is related to identity fusion (I feel connected), social
identity (one of us), and so is descriptive representation (the representative is like
me). Such overlap is why we make no expectations as to whether these three con-
cepts are distinct dimensions or will load on one or more factors.
We created a total of 25 items. Since affective processing precedes cognitive pro-
cessing, the arousal and valence mannequins, as seen in Fig.3, were measured first.
Then, the respondents received the remaining 23 items for feeling represented in a
battery measured on a 5-point scale (strongly disagree—strongly agree) in random
order. This was done to ease the respondent’s task and make the items easily compa-
rable and suitable in the exploratory factor analysis.
The items were transformed to fit our context of unelected representatives or
political activists and were asked in a comparable matter, reducing time and inten-
sity for the respondent. Feeling represented could be a multifaceted concept with
both instrumental and expressive components, or it could be that there are several
dimensions to this concept. That is what we want to untangle by including the fol-
lowing items. As the claim introduces both an activist, a message, and a cause, it is
important to note that feeling represented can be a state that takes all of these into
account. Therefore, it is important to consider the subject of the item. For instance,
social identity requires an ingroup, which, in this case, includes the supporters of
both the cause and the activist. You can be fused with both the cause and the activ-
ist, and you can feel connected to both the cause and the activist. We tried to con-
sider this in the items and differentiate between the activist, cause, and message.
This means we could also find that various dimensions may be more related to
A.Vik, P.de Wilde
Table 1 Items feeling represented
Label Item wording Sources
fr.valance See figure below
fr.arousal See figure below The self-assessment manikin protocol developed by Bradley and Lang (1994)
fr.si1 When people criticize (representative x and his/her cause), it feels like a
personal insult
fr.si2 When I meet someone who supports the cause raised by (representative x) I
feel connected with this person
fr.si3 When people praise (representative x and his/her activism), it makes me feel good
fr.si4 When I speak about the cause raised by (representative X), I refer to it as my
cause
Social identity/expressive partisanship; Huddy etal. (2015) and Huddy and
Bankert (2017)
fr.i1 I identify with (representative X)
fr.if1 I am one with (representative x)
fr.if2 I feel immersed with (representative x)
fr.if3 I have a deep emotional bond with (representative x)
fr.if4 (Representative x) is me
fr.if5 I’ll do more for the cause raised by (representative x) than any other person
would do
fr.if6 Thinking about (representative x) and his/her cause makes me feel strong
fr.if7 I make (representative x) and his/her cause strong Leader Identification and fusion with leader; Gómez etal. (2020) and Kunst
etal. (2019)
fr.c1 I feel connected to (representative X)
fr.c2 (Representative X) doesn’t have a clue about the real world. (S)he says (s)he
does, but I feel it is just lip service
fr.c3 (Representative X) is a self-serving opportunist. (S)he is doing (activity Z)
for the sake of publicity
fr.c4 I feel distant and disconnected from (Representative X)
fr.c5 (Representative X) voices my opinion
Reconceptualizing “feeling represented”: Anew approach…
Table 1 (continued)
Label Item wording Sources
fr.c6 I empathize with (representative x) and their cause
fr.c7 This message makes me feel heard Political connectedness inspired by Coleman (2005)
fr.dr (Representative X) is similar to me
fr.sr1 I believe in the cause raised by this message, irrespective of who says it
fr.sr2 This message represents my points of view very well Descriptive and substantive representation: Holmberg (2020)
fr.fr I feel represented by (representative X) New item
A.Vik, P.de Wilde
feeling represented by the representative, feeling only represented by the message,
or only identifying with the group.
Other measures
In the questionnaire section, respondents were asked demographic questions such
as gender and age, as well as their level of political interest, ideological self-place-
ments, and which persons they had heard of our activists. In addition, the respond-
ents filled out batteries for “need for cognition” and “need for affect” (Arceneaux
and VanderWielen 2013). All measures are illustrated in Appendix A1.
Design andprocedure
We conducted an online survey to test our new measurement. The online survey was
conducted via Qualtrics and distributed by Prolific in February 2022. Prolific is a
crowdsourcing platform similar to MTurk but has been found to result in better data
quality and more diverse participants (Peer etal. 2017). The respondents received
0.8 dollars for their participation. The survey was published on prolific, and Ameri-
can prolific users could take the survey. Respondents were required to speak fluent
English, and the sample was set to be gender balanced. Respondents would have
to give their informed consent before being able to participate in the study, and the
study was completely anonymized to make sure no personal information was given.
The survey consisted of two parts. First, we asked the respondents to fill in a
questionnaire including questions about their gender, age, political interest, ideol-
ogy, need for cognition, need for affect, and which of the activists they knew of.
After this, each respondent was shown two random representative claims of the
six shown in Fig.4 and asked to answer each of the 25 items for feeling repre-
sented for both of them. We created a total of 6 representative claims that vary on
the claimant and the claim, which we elaborate on in the following section. This
procedure was done for three reasons: (1) to put the framework into a claim set-
ting as suggested by Saward (2010), and we know that political communication
influences citizens’ level of feeling represented (Dvir-Gvirsman etal. 2022). (2)
So that we can assess whether our measurement tool works across different types
of unelected representation. Note that this is not done to assess how these claims
Fig. 3 Self-assessment manikin protocol; arousal (top panel) and valance (bottom panel)
Reconceptualizing “feeling represented”: Anew approach…
predict levels of feeling represented but merely if the factor structure is consistent
across claims, ensuring that our new measurement instrument is valid. (3) When
respondents witness two claims, we can split the data and conduct both explora-
tory and confirmatory factor analysis, as conducting both on the exact same
observation biases results. Finally, the respondents were debriefed and asked to
make any comments they may have and actively send their responses. The total
duration of the survey was approximately 8minutes.
Thus, through our innovative approach, we can both explore possible factor struc-
tures and test them within the same data collection following the constructivist turn
in political representation, ensuring the robustness of the new measurement tool.
Representative claims
A representative claim can be considered a unique form of political communica-
tion or rhetoric. All representative claims contain a policy demand. Otherwise, they
would not qualify as claims (Koopmans and Statham 1999). In addition to that,
both the maker of the claim (i.e., the claimant) and the object of the claim (i.e., the
intended constituency, who the claimant is speaking for) are what we refer to as
essential components of representative claims (Gora and DeWilde 2019). The claim
is made as a tweet,2 as we know that unelected representatives use this platform and
that online communication influences citizens’ level of feeling represented (Dvir-
Gvirsman etal. 2022).
We choose claims that pertain to three political issues (1) climate change (pro-
climate action vs. anti-climate action), (2) the gender pay gap (pro-equal pay vs.
anti-equal pay), (3) foreign military/defense policy (pro-diplomacy vs. anti diplo-
macy). These three issues were chosen because according to new descriptive work
published by (Pew Research Center 2019), these issues range in partisan polariza-
tion, in which climate change is the most polarized and foreign policy the least. We
find unelected representatives in the USA who have made such policy demands in
Fig. 4 Representative claims by unselected representatives
2 This experiment was run in February 2022, before Twitter was made X.
A.Vik, P.de Wilde
their online communication.3 We decided to use real actors as the makers of the
claim and create the claims based on their real communication. This was to ensure
the external and, specifically, ecological validity of our design. We also argue that
using real actors is more suitable considering our measurements on leader identifi-
cation and identity fusion, as these measures would be difficult to measure for fic-
tional actors or cue-givers. The object was kept as the ambiguous “we,” “everyone,”
or “people” in all claims.
Participants
605 American prolific users took part in the study, consented, and sent their
responses. In our sample, 295 were male, 297 were female, 8 identified as nonbinary
or other, and 5 preferred not to say. The respondents ranged in age from 18 to 82
(mean = 37.25, SD = 13.61, median = 34). The sample was overall fairly interested
in politics (mean = 3.43, SD = 1.11, median = 4) and was slightly left-skewed in
ideology (mean = 3.99, SD = 2.58, median = 3).
Results
Since each respondent saw two claims, we divided our sample so that we had one
observation per respondent in each sub-sample, one for an exploratory factor analy-
sis (EFA) and one for the confirmatory factor analysis (CFA); we also performed a
multigroup confirmatory factor analysis (MG-CFA) on the whole sample. All analy-
ses were conducted in R [Macintosh Version 4.1.1; (R Core Team 2022)], and the
EFA benefited from the psych package [Version 1.7.8, (Revelle 2022)], whereas the
CFA was done with the lavaan package [Version 0.6.12; (Rosseel 2012)]. A repro-
ducible script and dataset is available in the OSF repository. There were no missing
data, and our items did not follow a normal distribution (see Fig. A2, in Appendix
A2), meaning we had to use estimators that could handle non-normally distributed
data.
Exploratory factor analysis (EFA)
Bartlett’s test of sphericity (Bartlett 1950) was used to ensure that the correlation
matrix was not random
(x2(300)
= 15953.21,
p
<
0.001
) and the KMO statistic
(Kaiser 1970) was required to be above a minimum of.70 (KMO = 0.98). Thus,
it was determined that the correlation matrix was appropriate for factor analysis.
After confirming that the correlation matrix was factorable (see Appendix A2 for
the correlation matrix), we proceeded with the EFA on a subset of our sample (one
3 To add the sender of the message does potentially bias the treatment effects, but since our goal is meth-
odological and not empirical combined with how the claimant is central in the theoretical conceptualiza-
tion of representative claims, we decided to include a variety of claim makers.
Reconceptualizing “feeling represented”: Anew approach…
observation per respondent,
n=605
). Exploratory factor analysis was selected
over PCA because the intent was to identify a latent factor structure (Fabrigar and
Wegener 2011). A PAF extraction method with initial commonalities estimated by
squared multiple correlations was employed because of its relative tolerance of non-
normality and demonstrated ability to recover weak factors (DeWinter and Dodou
2012). Following the general advice of an EFA, the eigenvalue and the visual scree
test were used to determine the appropriate number of factors to retain. We also took
into account theoretical considerations and parsimony. An obliminin rotation, spe-
cifically promax, was used due to the likelihood of correlated factors.
The eigenvalues and scree suggested that three or two factors should be retained.
The two different factor solutions were then examined for adequacy. For the three-
factor solution, factor 1 was initially saliently loaded with 13 items, factor 2 with
9, and factor 3 with 3. For the two-factor solution, factor 1 remained the same as in
the two-factor solution with 13 items, but factor 2 with 12 items. Unsurprisingly,
for both models, our two first factors may suffer from over-factoring (reliability was
above 0.95 for each factor). Thus, we made certain adjustments by removing items
that loaded on several factors to honor the simple-factor structure and items with
somewhat low loading (>0.5). In addition, we considered the proportion of vari-
ance for each item and theoretical parsimony. Then, we re-ran our 2 and 3-factor
EFA with 12 retained items. We have illustrated the EFA’s in Appendix A3. The
factor structures remained the same and with similar loadings when conducted with
different extraction methods, such as OLS.
To determine which model fits best, we benefit from certain fit indices (Finch
2020), specifically BIC (Bayesian Information Criterion), RMSEA (The Root Mean
Square Error of Approximation), TLI (Trucker-Lewis Index), and the chi-square
goodness of fit (chi-square/df < 3). The 2-factor model explained 70% of the cumu-
lative variance, with Factor 1 accounting for 41% and Factor 2 for 29%. Factor load-
ings ranged from
−0.76
to 0.94 for Factor 1 and from 0.58 to 0.99 for Factor 2. Over-
all, the 2-factor model demonstrated a moderate model fit [BIC 106.29, RMSEA
= 0.114 (90% CI = 0.104 0.125), TLI = 0.923, 6833.63(43)]. The 3-factor model
explained 75% of the cumulative variance, with Factor 1 explaining 31%, Factor 2
explaining 29%, and Factor 3 explaining 16%. Factor loadings ranged from 0.79 to
0.98 for Factor 1, from 0.58 to 0.98 for Factor 2, and from 0.52 to 0.74 for Factor 3.
The model demonstrated an improved model fit [BIC =
−
156.28, RMSEA = 0.033
(90% CI = 0.017, 0.048), TLI = 0.993, 6833.63(33)]. Overall, the 3-factor model
provided a superior fit to the data compared to the 2-factor model, as evidenced by a
lower BIC value and better fit indices, indicating better model parsimony and good-
ness of fit. The factors and their items are illustrated in Table2, and we have illus-
trated their distributions in Fig.5.
Theoretically, we also have a basis that can support the three-factor structure.
Factor 1 can be considered the Instrumental dimension of feeling represented, which
signifies a pragmatic alignment and functional relevance between the representa-
tive’s message and one’s own viewpoints, encompassing belief in the cause, empathy
with the representative, and the representative’s role as a conduit for personal opin-
ions, illustrating a utilitarian resonance where the message’s alignment takes prec-
edence over the messenger. Factor 2 can be considered the Expressive dimension of
A.Vik, P.de Wilde
Table 2 The factors and their respective items
Factor Definition Items
1 Instrumental (
𝛼=0.96
) agreement and resonance sr 2: this message represents my points of view very well
sr1: I believe in the cause raised by this message, irrespective of who says it
c6: I empathize with (representative x) and their cause
c5: (Representative X) voices my opinion
2 Expressive (
𝛼=0.92
) emotional connection and identification if6: Thinking about (representative x) and his/her cause makes me feel strong
if1: I am one with (representative x)
si3: When people praise (representative x and his/her activism) it makes me feel good
c1: I feel connected to (representative X)
si4: When I speak about the cause raised by representative X, I refer to it as my cause
3 Aversive (
𝛼=0.81
) skepticism, suspicion, and emotional disconnection c2: (Representative X) doesn’t have a clue about the real world. (S)he says (s)he does,
but I feel it is just lip service
c3: (Representative X) is a self-serving opportunist. (S)he is doing (activity Z) for the
sake of publicity
c4: I feel distant and disconnected to (Representative X)
Reconceptualizing “feeling represented”: Anew approach…
feeling represented, which represents an emotional connection and personal iden-
tification with the representative and their cause, reflecting sentiments of empow-
erment through alignment, unity with the representative, emotional responses to
criticism, deep connectedness, and a sense of personal ownership in the advocated
cause. Factor 3 can be considered the aversive component of feeling represented,
and it encapsulates a collection of skeptical perspectives and emotional detachment
toward the representative, characterized by reservations about their authenticity, sus-
picions of self-interest, and a sense of emotional remoteness.
How commonplace are these feelings of representation? Looking at the distribu-
tions of our constructs in Fig.5, feeling expressively represented skews toward the
low end of the distribution, meaning it is more common not to feel very strongly
expressively represented. Feeling instrumentally represented has a U-shaped distri-
bution, which means that respondents typically either agree or disagree with the pol-
icy issues of the claims. Feeling not represented, as captured in the aversive dimen-
sion, has a more normal to flat distribution, meaning more of our respondents felt
aversive toward these representative claims. These differences in distributions have
important implications for analytical procedures in future research and also open up
important discussions regarding the magnitude of feelings of representation.
Confirmatory factor analysis
Following the EFA, we conducted several CFAs to confirm the findings from our
EFA. Since CFA offers more model flexibility, we can compare different models,
also with covariances added to determine the best-performing model (Brown 2015).
Based on the EFA, we decided to test the following 4 models in our CFA on a dif-
ferent subset of our sample (
n=605
). Model 1 was the two-factor model. Model 2
is a variation of Model 1 with error covariances added. Model 3 was the three-factor
model our EFAs indicated would be the best fit. Model 4 is a variation of Model 3
with error covariances added. We used the WLSMW (mean and variance-corrected
Fig. 5 The distribution of the three factors of feeling represented
A.Vik, P.de Wilde
weighted least squares) estimator to account for the fact that none of our variables
are normally distributed and are measured on an ordinal level. Goodness-of-fit of
all the CFA models was examined with The root mean square error of approxi-
mation (RMSEA) <0.06 (90 % CI <0.06), the standardized root mean square
residual (SRMR) <0.08, the Comparative Fit Index (CFI), and the Trucker-Lewis
Index (TLI)
≥
0.95 (Kline 2015). In Table3, the results of the series of CFAs are
illustrated.
Model 1 showed a non-acceptable fit due to the slightly high RMSEA (
=0.1
).
Model 2 showed a remarkably improved fit after adding error covariances to Model
1. Model 3 achieved an adequate fit with almost all measures within acceptabil-
ity and RMSEA on the verge of acceptability (
=0.06
). Model 4 also showed an
improved fit after adding error covariances to Model 3 (RMSEA = 0.05). Thus, we
conclude that Model 4, with our three-factor solution with error variances, performs
the best. Model 4 is illustrated with a path diagram in Fig.6 below.
Table 3 Confirmatory factor
analysis models rmsea cfi tli srmr
Model 1 (2 factor) 0.10 1.00 1.00 0.05
Model 2 (2 factor + error covariance) 0.07 1.00 1.00 0.04
Model 3 (3 factor) 0.06 1.00 1.00 0.04
Model 4 (3 factor + error covariance) 0.05 1.00 1.00 0.03
Fig. 6 Path diagram CFA Model 4, the 3-factor solution with error covariance. f1 Instrumental dimen-
sion, f2 expressive dimension, f3 aversive dimension. Standardized Factor Loadings (dotted line indicates
fixed loading). All factor loadings are significant at the
p>
0.001 level
Reconceptualizing “feeling represented”: Anew approach…
Model assessment: determining thereliability andvalidity ofCFA
To establish if we have obtained a reliable factor solution, we look closely at indica-
tor reliability, scale reliability, construct validity, and discriminant validity (Brown
2015). As a measure of indicator reliability, we look at squared factor loadings,
which indicate how much of the variance in our indicators can be explained by our
latent variables. All of our items have a squared factor loading above 0.4, indicating
that we have indicator reliability. For scale reliability, we look at Cronbach alpha,
Raykov’s reliability, which is typically considered as more accurate (Raykov 1997).
All of our scales have high Cronbach’s alpha of above 0.8 (Factor 1 =
𝛼0, 96
, Factor
2 =
𝛼0, 92
, Factor 3 =
𝛼0, 81
). Factor 1 (RCC = 0.75) and 2 (RCC = 0.78) also have
good Raykov’s reliability, whether factor 3’s reliability was non-acceptable (RCC
= 0.55). For convergent validity, we look at the average variance extracted (AVE),
which refers to the fact that our latent variable should explain at least an average of
50% of its associated indicators. Factor 1 obtained an AVE of 0.89, factor 2, 0.83,
and factor 3, 0.61, indicating that we have convergent validity for all our factors.
Finally, discriminant validity refers to that our latent variables are distinct, so when
the AVE is larger than the squared correlation, which is true for both factor 1 and
factor 2, but not for factor 3. In conclusion, factor 3 might be a less reliable and valid
measurement of feeling represented, whereas factors 1 and 2 perform very well.
Multigroup CFA
Multigroup analyses were conducted on our best-performing model with our 3-fac-
tor solution (CFA:model 4). First, we conduct CFAs separately for each representa-
tive claim, again with the WLSMV estimator. This is illustrated in Table4. There
seems to be a good model fit regardless of the claim; however, claims 1 and 5 have
slightly worse model fits. These are the claims by Leonardo Dicaprio and Ange-
lina Jolie, the UNREPs with celebrity status beyond their activism. This could mean
that feeling represented is a concept that works differently in response to different
claims, but to put this to the test, we conduct a multigroup CFA.
To further see whether feeling represented can be measured from different claims,
we used an unconstrained–constrained approach (see The lavaan Project (2022)) on
our full sample (n = 1210) to establish if we have measurement invariance; the sta-
tistical property of a measurement that indicates that the same underlying construct
Table 4 Multigroup CFA:
model fit for each claim rmsea cfi tli srmr
Claim 1 0.10 0.99 0.99 0.08
Claim 2 0.00 1.00 1.00 0.03
Claim 3 0.08 1.00 0.99 0.07
Claim 4 0.00 1.00 1.00 0.03
Claim 5 0.09 0.99 0.99 0.08
Claim 6 0.00 1.00 1.00 0.02
A.Vik, P.de Wilde
is being measured across groups (Bialosiewicz etal. 2013). First, an unconstrained
model (i.e., configural) was run, which allowed parameters to vary freely. This anal-
ysis was followed by a fully constrained model, where the parameters were con-
strained to be equivalent between the groups (in our case, the claims) (Hirschfeld
and VonBrachel 2014). The unconstrained and constrained models were then com-
pared using Chi-square difference testing to examine the presence of moderation in
the overall models (Hirschfeld and VonBrachel 2014). Chi-square is sensitive to
sample size, so we also compare fit indices and
Δ
CFI (which should be <.01) to
establish if we have structural equivalence (Chen 2007). Our MG CFA is illustrated
in Table5.
As can be seen in the table, the test for weak invariance (configural versus metric)
yields both a significant chi-square test and a small
Δ
CFI, whereas for strong invari-
ances (metric versus scalar), we still have a small
Δ
CFI but the chi-square test is not
significant. In conclusion, we may interpret these results that we have some invari-
ance but further investigations into how feeling represented works in response to dif-
ferent claims are needed to understand if we have a measurement that measures the
same construct across different representative claims.
Discussion andconclusion
This paper begins by making a theoretical contribution. We define “feeling repre-
sented” as a subjective feeling state that emerges from witnessing representative
claims, thus placing this construct as a missing link in the constructivist turn in
political representation. We identify additional shortcomings in existing methods
for measuring feeling represented, which primarily rely on Piktin’s concept of sub-
stantive representation and overlook the socio-psychological aspect of representa-
tion. We remedy these shortcomings by introducing a psychological viewpoint to
equip the construct of feeling represented fully. Through an elaborate methodologi-
cal effort we devise a new measurement tool to measure whether people feel repre-
sented as a result of witnessing representative claims.
Surveying a wide spread of relevant literature, we developed 25 different ways of
asking people whether they feel represented. We provided a sample of 605 American
citizens with a set of representative claims by unelected representatives and asked
them these questions after exposure. From these data, we conducted exploratory,
Table 5 Multigroup CFA: invariance testing
a Metric vs configural
b Scalar vs metric
𝜒2
df p-value rmsea cfi tli srmr
Δ𝜒2
Δdf
p
Configural 179.29 300.00 1.00 0.00 1.00 1.01 0.04 – – –
Metric 311.88 345.00 0.90 0.00 1.00 1.00 0.06 132.594a 45 0
Scalar 371.19 390.00 0.75 0.00 1.00 1.00 0.06 59.311b 45 0.075
Reconceptualizing “feeling represented”: Anew approach…
confirmatory, and multigroup confirmatory factor analyses. We establish that feeling
represented is a three-dimensional concept consisting of 12 items, all three consist-
ing of cognitive and affective aspects. We label these as an instrumental dimension
(factor 1), an expressive dimension (factor 2), and an aversive dimension (factor 3).
The instrumental dimension captures the extent to which respondents agree with the
policy demand voiced in the claim and whether they like and empathize with the
maker. The expressive dimension captures the respondent’s emotional attachment
to the maker of the claim and the claim itself. Finally, the aversion dimension cap-
tures whether the respondent feels misrepresented and finds the claim and claimant
untrustworthy and in sincere.
Our finding that feeling represented is a multidimensional concept consisting of
both instrumental and expressive processes matters greatly for the research agenda.
This article’s aim has been methodological: to generate a measurement tool to
empower subsequent empirical research on feeling represented. The different dimen-
sions are likely to be influenced by individual differences, have different behavioral
outcomes, and have different democratic implications.
First of all, it opens up the question of whether individual characteristics of
respondents affect our three dimensions of feeling represented in different ways. Are
some people more likely than others to strongly agree with claims than others, or to
have an aversive reaction? If so, how can we explain these differences? Second, vari-
ations in feeling represented along our three-dimensional space may, in turn, have
different effects on political behavior. The expressive dimension is similar to iden-
tity fusion, which we know is important for extreme pro-group behavior (Gómez
etal. 2020; Kunst etal. 2019). This dimension of feeling represented could poten-
tially explain more of how citizens get involved in various political causes, includ-
ing social movements. Thirdly, the three different dimensions of feeling represented
may have different implications for democratic legitimacy. The aversive dimension,
in particular, may be related to citizens’ trust in representatives, the system, and
their political perspectives. Importantly, some feelings of (mis-)representation may
not translate into observable behavior while nevertheless carrying consequences for
democratic legitimacy. It is thus important to investigate feelings of representation
directly rather than to assume it is linearly related to more easily measured political
behavior such as voting. Future research can hopefully use the tools developed in
this article to address such questions.
Furthermore, building on the constructivist turn in political representation,
our study lends credence to the idea that people may feel represented (or misrep-
resented) due to exposure to claims made by unelected representatives. Empirical
variation along our three dimensions of feeling represented in our study testifies to
this. This implies that the network of connections between representatives and rep-
resented affects citizens’ trust in politics, and the legitimacy of liberal representative
democracy extends beyond the electoral arena. Unelected representatives and their
behavior may well affect the legitimacy of the system, just like elected politicians
do. This invites research on whether, how, and to what extent unelected representa-
tives affect citizens’ trust in politics and satisfaction with the way democracy works
through their claims.
A.Vik, P.de Wilde
The instrumental, expressive, and aversive dimensions of feeling represented cor-
relate strongly but have different distributions. Thus, feeling represented is evidently
a complex psychological state that requires substantial cognitive and affective effort
from citizens. The expressive dimension in particular, a dimension that arguably is
a strong form of feeling represented, was skewed toward the low end of the distri-
bution. This raises critical questions regarding the empiricism of the constructivist
turn: How strongly and frequently do people feel represented as a result of exposure
to representative claims? Is feeling strongly (mis)represented a condition for repre-
sentative claims to become (de)legitimized in this two-way communicative process?
We urge future research to take such questions into consideration.
In the event that research only has the opportunity to ask one single question
to tap feelings of representation, and opt for the intuitive and generic question “I
feel represented by (representative X)?”, we advise the results to be interpreted as
containing both instrumental and expressive components. In our study, this particu-
lar item loaded on both the instrumental and expressive factors, indicating again
that feeling represented is a complex state that elicits both cognitive and affective
processes.
Our study carries clear limitations, as any other study. We measured feeling rep-
resented with an American case study, with American claim makers, making it dif-
ficult to generalize our findings beyond the American context. It could be that coun-
tries with higher generalized and political trust and different cultural values have
citizens who react differently to representative claims. Furthermore, our claims con-
cern the issues of climate change, gender equality, and foreign policy. This was done
to keep the study manageable and because the issues differed in their levels of polar-
ization (Pew Research Center 2019). Looking at more issues can further our under-
standing of whether people feel represented across issue domains. We also measured
feeling represented at a single point in time but do not know how long such feelings
last or how much they potentially fluctuate.
These limitations notwithstanding, it is important to mention that the factor
structure remained similar across all of our claims, meaning that these items should
work to measure feelings of representation in response to a variety of representa-
tive claims with different issue positions and claimants. Future research could rep-
licate our study in other countries with large and diverse samples, compare claims
by unelected representatives with claims by elected representatives, and address
other issues and issue positions. As such, we view this measurement tool as an
important stepping stone in a larger research agenda: the affective construction of
representation.
We conclude that feeling represented is part of the construction of representation,
has an instrumental, expressive, and aversive dimension, and can occur after citizens
witness representative claims from unelected representatives. This lends credence to
the constructivist notion that representation is the product of a dynamic process of
claim-making and opens up a range of interesting and important research avenues
informed by this psychological perspective on political representation.
Supplementary Information The online version contains supplementary material available at https:// doi.
org/ 10. 1057/ s41269- 024- 00358-2.
Reconceptualizing “feeling represented”: Anew approach…
Acknowledgments We thank the anonymous reviewers and journal editors for their excellent comments
and suggestions on our manuscript. All remaining mistakes are, of course, our responsibility alone. Previ-
ous versions of the manuscript were presented at the ECPR joint session in 2021 (project idea), the ISPP
annual conference in 2022, and the ECPR annual conference in 2022 (first draft). We thank Lene Aarøe,
Vin Arcenaux, and Oliver Treib for their helpful suggestions on earlier versions of the manuscript. We
gratefully acknowledge the financial support provided by the Norwegian Research Council under Grant
No. 287861.
Data availability Data and code is available at the Open Science Framework, with the following link:
https://osf.io/d46tz/
Declarations
Conflict of interest On behalf of all authors, the corresponding author states that there is no conflict of
interest.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License,
which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long
as you give appropriate credit to the original author(s) and the source, provide a link to the Creative
Commons licence, and indicate if changes were made. The images or other third party material in this
article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line
to the material. If material is not included in the article’s Creative Commons licence and your intended
use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permis-
sion directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/
licenses/by/4.0/.
References
Aarøe, L. 2011. Investigating frame strength: The case of episodic and thematic frames. Political Com-
munication 28 (2): 207–226. https:// doi. org/ 10. 1080/ 10584 609. 2011. 568041.
Aarts, K., and J. Thomassen. 2008. Satisfaction with democracy: Do institutions matter? Electoral Stud-
ies 27 (1): 5–18. https:// doi. org/ 10. 1016/j. elect stud. 2007. 11. 005.
Abelson, R. 1963. Computer simulation of “hot” cognition. In Computer simulation of personality, eds.
S. Tomkins and D. Messick, 277–298.
Arceneaux, K., and R.J. Vander Wielen. 2013. The effects of need for cognition and need for affect on
partisan evaluations. Political Psychology. https:// doi. org/ 10. 1111/j. 1467- 9221. 2012. 00925.x.
Atwell Seate, A., R. Ma, H.-Y. Chien, and D. Mastro. 2018. Cultivating intergroup emotions: An inter-
group threat theory approach. Mass Communication and Society 21 (2): 178–197. https:// doi. org/ 10.
1080/ 15205 436. 2017. 13812 62.
AtwellSeate, A., and D. Mastro. 2015. Communication monographs media’s influence on immigration
attitudes: An intergroup threat theory approach. https:// doi. org/ 10. 1080/ 03637 751. 2015. 10684 33.
https:// www. tandf online. com/ action/ journ alInf ormat ion? journ alCode= rcmm20
Bakker, B., G. Schumacher, and M. Rooduijn. 2021. Hot politics? Affective responses to political rhet-
oric. American Political Science Review 115 (1): 150–164. https:// doi. org/ 10. 1017/ S0003 05542
00005 19.
Barrett, L.F. (2017). How emotions are made: The secret life of the brain.
Bartlett, M.S. 1950. Tests of significance in factor analysis. British Journal of Statistical Psychology 3
(2): 77–85. https:// doi. org/ 10. 1111/j. 2044- 8317. 1950. tb002 85.x.
Bialosiewicz, S., K. Murphy, and T. Berry. 2013. An introduction to measurement invariance testing:
Resource packet for participants do our measures measure up? The critical role of measurement
invariance (Tech. Rep.). Washington: American Evaluation Association.
Bolls, P.D., A. Lang, and R.F. Potter. 2001. The effects of message valence and listener arousal on atten-
tion, memory, and facial muscular responses to radio advertisements. Communication Research.
https:// doi. org/ 10. 1177/ 00936 50010 28005 003.
A.Vik, P.de Wilde
Bradley, M., and P. Lang. 2000. Measuring emotion: Behavior, feeling, and physiology. In Cognitive Neu-
roscience of Emotion, 49–59.
Bradley, M.M., and P.J. Lang. 1994. Measuring emotion: The self-assessment manikin and the seman-
tic differential. Journal of Behavior Therapy and Experimental Psychiatry. https:// doi. org/ 10. 1016/
0005- 7916(94) 90063-9.
Brown, T. 1836. Lectures on the philosophy of the human mind, Vol. 1 of 3, vol.1. https:// www. guten
berg. org/ ebooks/ 43116/ pg431 16- images. html. utf8
Brown, T.A. 2015. Confirmatory factor analysis for applied research. Guilford Publications.
Castiglione, D., and M.E. Warren. 2019. Rethinking democratic representation: Eight theoretical issues
and a postscript1. In The constructivist turn in political representation, 21–47. Edinburgh Univer-
sity Press. https:// doi. org/ 10. 3366/ edinb urgh/ 97814 74442 602. 003. 0002
Chen, F.F. 2007. Sensitivity of goodness of fit indexes to lack of measurement invariance. Structural
Equation Modeling: A Multidisciplinary Journal 14 (3): 464–504. https:// doi. org/ 10. 1080/ 10705
51070 13018 34.
Coleman, S. 2005. The lonely citizen: Indirect representation in an age of networks. Political Communi-
cation. https:// doi. org/ 10. 1080/ 10584 60059 09331 97.
Coronel, J.C., M.C. Duff, D.E. Warren, K.D. Federmeier, B.D. Gonsalves, D. Tranel, and N.J. Cohen.
2012. Remembering and voting: Theory and evidence from amnesic patients. American Journal of
Political. https:// www. ncbi. nlm. nih. gov/ pmc/ artic les/ PMC39 17545/https:// doi. org/ 10. 1111/J. 1540-
5907. 2012. 00608.X.
Dahl, R.A. 1971. Polyarchy: Participation and opposition. Yale University Press.
Dalton, R.J. 2016. Party identification and its implications. In Oxford research encyclopedia of politics.
https:// doi. org/ 10. 1093/ acref ore/ 97801 90228 637. 013. 72.
DeMulder, A. 2022. Making sense of citizens’ sense of being represented. A novel conceptualisation
and measure of feeling represented. Representation. https:// www. tandf online. com/ action/ journ alInf
ormat ion? journ alCode= rrep20https:// doi. org/ 10. 1080/ 00344 893. 2022. 20956 62.
DeWinter, J.C.F. and Dodou, D. 2012. Factor recovery by principal axis factoring and maximum likeli-
hood factor analysis as a function of factor pattern and sample size. Journal of Applied Satistics 39
(4): 695–710. https:// www. tandf online. com/ action/ journ alInf ormat ion? journ alCode= cjas20https://
doi. org/ 10. 1080/ 02664 763. 2011. 610445.
Disch, L. 2021. Making constituencies: Representation as mobilization in mass democracy. The Univer-
sity of Chicago Press.
Disch, L. 2011. Toward a mobilization conception of democratic representation. American Political Sci-
ence Review. https:// doi. org/ 10. 1017/ S0003 05541 00006 02.
Disch, L. 2015. The “Constructivist turn’’ in democratic representation: A normative dead-end? Constel-
lations. https:// doi. org/ 10. 1111/ 1467- 8675. 12201.
Druckman, J.N., and R. McDermott. 2008. Emotion and the framing of risky choice. Political Behavior.
https:// doi. org/ 10. 1007/ s11109- 008- 9056-y.
Dvir-Gvirsman, S., K. Tsuriel, T. Sheafer, S. Shenhav, A. Zoizner, L. Lavi, etal. 2022. Mediated repre-
sentation in the age of social media: How connection with politicians contributes to citizens’ feel-
ings of representation. Evidence from a longitudinal study. Political Communication 39 (6): 779–
800. https:// doi. org/ 10. 1080/ 10584 609. 2022. 21243 35.
Epitropaki, O., R. Kark, C. Mainemelis, and R.G. Lord. 2017. Leadership and followership identity pro-
cesses: A multilevel review. https:// doi. org/ 10. 1016/j. leaqua. 2016. 10. 003.
Erisen, C., M. Lodge, and C.S. Taber. 2014. Affective contagion in effortful political thinking. Political
Psychology 35 (2): 187–206. https:// doi. org/ 10. 1111/J. 1467- 9221. 2012. 00937.X.
Fabrigar, L.R., and D.T. Wegener. 2011. Exploratory factor analysis. Oxford University Press.
Finch, W.H. 2020. Using fit statistic differences to determine the optimal number of factors to retain in
an exploratory factor analysis. Educational and Psychological Measurement 80 (2): 217. https:// doi.
org/ 10. 1177/ 00131 64419 865769. https:// www. ncbi. nlm. nih. gov/ pmc/ artic les/ PMC70 47263/.
Gómez, Ã., J. Chinchilla, A. Vázquez, L. López-Rodríguez, B. Paredes, and M. Martínez. 2020. Recent
advances, misconceptions, untested assumptions, and future research agenda for identity fusion the-
ory. Social and Personality Psychology Compass. https:// doi. org/ 10. 1111/ spc3. 12531.
Gora, A. and P. De Wilde. 2019. Handbook for representative claims analysis (Tech. Rep.). NTNU.
https:// www. recon nect- europe. eu
Greene, S. 2004. Social identity theory and party identification. Social Science Quarterly 85 (1):
136–153.
Reconceptualizing “feeling represented”: Anew approach…
Guasti, P., and D.R.D. Almeida. 2019. Claims of misrepresentation: A comparison of Germany and Bra-
zil. Politics and Governance 7 (3): 152–164. https:// doi. org/ 10. 17645/ pag. v7i3. 2143.
Guasti, P. and B. Geissel. 2019. Rethinking representation: Representative claims in global perspective.
Politics and Governance. https:// doi. org/ 10. 17645/ pag. v7i3. 2464
Hirschfeld, G., and R. Von Brachel. 2014. Improving multiple-group confirmatory factor analysis in R-A
tutorial in measurement invariance with continuous and ordinal indicators. Practical Assessment,
Research, and Evaluation. 19: 7. https:// doi. org/ 10. 7275/ qazy- 2946.
Holmberg, S. 2020. Feeling represented. In The oxford handbook of political representation in liberal
democracies. https:// doi. org/ 10. 1093/ oxfor dhb/ 97801 98825 081. 013. 21.
Huddy, L., D. Sears, and J. Levy. 2013. The Oxford handbook of political psychology, 2nd ed. Oxford
University Press.
Huddy, L., and A. Bankert. 2017. Political partisanship as a social identity. In Oxford research encyclope-
dia of politics. https:// doi. org/ 10. 1093/ acref ore/ 97801 90228 637. 013. 250.
Huddy, L., L. Mason, and L. Aarøe. 2015. Expressive partisanship: Campaign involvement, political
emotion, and partisan identity. American Political Science Review. https:// doi. org/ 10. 1017/ S0003
05541 40006 04.
Kaiser, H.F. 1970. A second generation little jiffy. Psychometrika 35: 401–415. https:// doi. org/ 10. 1007/
BF022 91817
Kaplan, Y.R., T. Sheafer, I. Waismel-Manor, and S.R. Shenhav. 2023. People’s sense of political repre-
sentation and national stories: The case of Israel. International Political Science Review. https:// doi.
org/ 10. 1177/ 01925 12123 11855 76.
Kline, R.B. 2015. Principles and practice of structural equation modeling, 2nd ed. Guilford Publications.
Knops, L. 2022. Towards an affective turn in theories of representation: The case of indignation. Rep-
resentation. https:// doi. org/ 10. 1080/ 00344 893. 2022. 20910 11. https:// www. tandf online. com/ action/
journ alInf ormat ion? journ alCode= rrep20
Koopmans, R., and P. Statham. 1999. Political claims analysis: Integrating protest event and political
discourse approaches. Mobilization: An International Quarterly 4 (2): 203–221. https:// doi. org/ 10.
17813/ maiq.4. 2. d7593 37060 7l6756.
Kreiss, D., R.G. Lawrence, and S.C. McGregor. 2020. Political identity ownership: Symbolic contests to
represent members of the public. Social Media + Society 6 (2): 205630512092649. https:// doi. org/
10. 1177/ 20563 05120 926495.
Kunst, J.R., J.F. Dovidio, and L. Thomsen. 2019. Fusion with political leaders predicts willingness to
persecute immigrants and political opponents. Nature Human Behaviour. https:// doi. org/ 10. 1038/
s41562- 019- 0708-1.
Lang, A., P. Bolls, R.F. Potter, and K. Kawahara. 1999. The effects of production pacing and arousing
content on the information processing of television messages. Journal of Broadcasting and Elec-
tronic Media. https:// doi. org/ 10. 1080/ 08838 15990 93645 04.
Linde, J., and J. Ekman. 2003. Satisfaction with democracy: A note on a frequently used indicator in
comparative politics. European Journal of Political Research 4 (2): 391–408. https:// doi. org/ 10.
1111/ 1475- 6765. 00089.
Lodge, M., and C.S. Taber. 2005. The automaticity of affect for political leaders, groups, and issues: An
experimental test of the hot cognition hypothesis. Political Psychology 26 (3): 455–482.
Mansbridge, J. 1999. Should blacks represent blacks and women represent women? A contingent “yes’’.
Journal of Politics. https:// doi. org/ 10. 2307/ 26478 21.
Mansbridge, J. 2003. Rethinking representation. American Political Science Review. https:// doi. org/ 10.
1017/ S0003 05540 30008 56.
Marcus, G.E. 2003. Emotions in politics. Annual Review of Political Science 3: 221–250. https:// doi. org/
10. 1146/ annur ev. polis ci.3. 1. 221.
Montanaro, L. 2017. Who elected Oxfam?: A democratic defense of self-appointed representatives.
https:// doi. org/ 10. 1017/ 97811 08297 721.
Peer, E., L. Brandimarte, S. Samat, and A. Acquisti. 2017. Beyond the Turk: Alternative platforms for
crowdsourcing behavioral research. Journal of Experimental Social Psychology 70: 153–163.
https:// doi. org/ 10. 1016/J. JESP. 2017. 01. 006.
Pew Research Center. 2019. In a politically polarized era, sharp divides in both partisan coalitions.
https:// www. pewre search. org/ polit ics/ 2019/ 12/ 17/ in-a- polit ically- polar ized- era- sharp- divid es- in-
both- parti san- coali tions/.
Pitkin, H. 1967. The concept of representation. University of California Press.
A.Vik, P.de Wilde
Potter, R.F., and P.D. Bolls. 2012. Psychophysiological measurement and meaning: Cognitive and emo-
tional processing of media. Routledge. https:// doi. org/ 10. 4324/ 97802 03181 027.
R Core Team. 2022. R: A language and environment for statistical computing. Foundation for Statistical
Computing, Vienna, Austria. https:// www.R- proje ct. org/
Raykov, T. 1997. Estimation of composite reliability for congeneric measures. Applied Psychological
Measurement 21 (2): 173–184. https:// doi. org/ 10. 1177/ 01466 21697 02120 06.
Redlawsk, D.P., and K. Mattes. 2022. Emotions and politics. The Cambridge Handbook of Political
Psychology. https:// doi. org/ 10. 1017/ 97811 08779 104. 010. https:// www. cambr idge. org/ core/ books/
cambr idge- handb ook- of- polit ical- psych ology/ emoti ons- and- polit ics/ EB5DE AEB56 6658E F5C52
1A763 A53E5 0E.
Revelle, W. 2022. Package “psych” (Tech. Rep.). https:// perso nality- proje ct. org/r/ psych/.
Rosseel, Y. (2012). lavaan: An R package for structural equation modeling. Journal of Statistical Soft-
ware. https:// doi. org/ 10. 18637/ jss. v048. i02.
Ruedin, D. 2015. Why aren’t they there? ECPR Press.
Saward, M. 2006. The representative claim. Contemporary Political Theory. https:// doi. org/ 10. 1057/ palgr
ave. cpt. 93002 34.
Saward, M. 2010. The representative claim. Oxford University Press. https:// doi. org/ 10. 1093/ acprof: oso/
97801 99579 389. 001. 0001.
Saward, M. 2016. Fragments of equality in representative politics. Critical Review of International Social
and Political Philosophy 19 (3): 245–262. https:// doi. org/ 10. 1080/ 13698 230. 2016. 11448 53.
Shouse, E. 2005. Feeling, emotion, affect. M/C Journal. https:// doi. org/ 10. 5204/ mcj. 2443.
Staehr Harder, M.M. 2020. Pitkin’s second way: Freeing representation theory from identity. Representa-
tion 56: 11–12. https:// doi. org/ 10. 1080/ 00344 893. 2019. 16368 53.
Stets, J.E., and P.J. Burke. 2000. Identity theory and social identity theory. Social Psychology Quarterly.
https:// doi. org/ 10. 2307/ 26958 70.
Swann, W.B., and M.D. Buhrmester. 2015. Identity fusion. Current Directions in Psychological Science.
https:// doi. org/ 10. 1177/ 09637 21414 551363.
Swann, W.B., J. Jetten, Ã. Gómez, H. Whitehouse, and B. Bastian. 2012. When group membership gets
personal: A theory of identity fusion. Psychological Review. https:// doi. org/ 10. 1037/ a0028 589.
Taber, C.S., and M. Lodge. 2006. Motivated skepticism in the evaluation of political beliefs. American
Journal of Political Science 50 (3): 755–769.
The lavaan Project. (2022). Tutorial lavaan: Latent variable analysis. https:// lavaan. ugent. be/ tutor ial/
groups. html.
Tormey, S. 2015. The end of representative politics. Cambridge: Polity Press.
Tsakiris, M., N. Vehar, and R. Tucciarelli. 2021. Visceral politics: A theoretical and empirical proof
of concept. Philosophical Transactions of the Royal Society B: Biological Sciences 376 (1822):
20200142. https:// doi. org/ 10. 1098/ rstb. 2020. 0142.
Urbinati, N. 2006. Political representation as a democratic process. In Redescriptions: The yearbook of
political thought and conceptual history, 18–40.
Valentino, N.A., K. Gregorowicz, and E.W. Groenendyk. 2009. Efficacy, emotions and the habit of par-
ticipation. Political Behavior. https:// doi. org/ 10. 1007/ s11109- 008- 9076-7.
Valentino, N.A., T. Brader, E.W. Groenendyk, K. Gregorowicz, and V.L. Hutchings. 2011. Election
night’s alright for fighting: The role of emotions in political participation. Journal of Politics.
https:// doi. org/ 10. 1017/ S0022 38161 00009 39.
Wolkenstein, F., and C. Wratil. 2020. Multidimensional representation. American Journal of Political
Science 65 (4): 862–876. https:// doi. org/ 10. 1111/ ajps. 12563.
Wolkenstein, F. 2021. Revisiting the constructivist turn in political representation.https:// doi. org/ 10.
1177/ 14748 85121 10559 51.
Zajonc, R.B. 1980. Feeling and thinking: Preferences need no inferences. American Psychologist 35 (2):
151–175. https:// doi. org/ 10. 1037/ 0003- 066X. 35.2. 151.
Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps
and institutional affiliations.
Reconceptualizing “feeling represented”: Anew approach…
Andrea Vik is a postdoctoral research fellow at the Centre for the Politics of Feelings, Royal Holloway,
University of London. She holds a Ph.D in political science from the Norwegian University of Science
and Technology (NTNU) and a M.Sc in communication science from the University of Amsterdam.
Andrea’s research interest lie at the intersection of political psychology, political communication, and
political representation.
Pieter de Wilde holds the chair in European Politics & Society at the University of Groningen, Faculty of
Arts. Before joining the University of Groningen in January 2024, he was Professor in Political Science
(2019-2023), Professor in European Studies (2018-2019) and Associate Professor in European Studies
(2016-2018) at the Norwegian University of Science and Technology (NTNU) in Trondheim, Norway.
His research interests concern comparative politics, democratic theory, European integration, globaliza-
tion, and political sociology.
Authors and Aliations
AndreaVik1,2,3 · PieterdeWilde3,4
* Andrea Vik
andrea.vik@rhul.ac.uk
1 Department ofPsychology, Royal Holloway, University ofLondon, London, UnitedKingdom
2 Centre forthePolitics ofFeelings, School ofAdvanced Study, University ofLondon, London,
UnitedKingdom
3 Department ofSociology andPolitical Science, Norwegian University ofScience
andTechnology (NTNU), Trondheim, Norway
4 Faculty ofArts, University ofGroningen, Groningen, TheNetherlands
