14
POLITICS, MEDIA AND THE

ELECTORAL ROLE OF PARTY
LEADERS

Anthony Mughan and Loes Aaldering

Introduction

The leaders of political parties are seen as key figures in the democratic
political process as they take primary responsibility for organizing their
parties? efforts to win elections and, if victorious, for governing the people
and the country they have been chosen to serve. Interestingly, however, the
conventional wisdom has been that, with the exception of presidential
candidates in the US, these same key political figures exert very little
influence on election outcomes given their limited impact on individuals?
vote choice. Essentially indistinguishable in the eyes of voters from the
party they represented, party leaders failed to influence the vote
independently of the strong, social cleavage-based partisan loyalties that
were the norm for much of the post-1945 period. Put differently, party
leaders in parliamentary systems of government were dismissed as at best
bit players in the larger election drama (Butler and Stokes 1969). More
recently, however, perceptions have changed dramatically and the study of
the electoral effects of party leaders is a growth area in the study of
democratic mass political behavior (Bean and Mughan 1989; Aarts, Blais
and Schmitt 2011; Bittner 2011).

Two particular developments are generally offered to explain the
emergence of parliamentary party leaders as electoral forces in their own
right. The first concerns a fundamental change in the electorate and the
second in the media environment in which elections take place. Let us take
the changing electorate first. Over the last several decades, voters? loyalties
to political parties have generally weakened as the social cleavages on
which those affiliations were based have become less salient. In addition,
many voters have become disillusioned with established parties?
performance in office (Franklin, Mackie and Valen 1992; Pharr and Putnam
2000). Commonly referred to as ?partisan dealignment,? this process
undermined the tendency to vote out of long-term habitual party loyalty and
made way for election-specific, or short-term, forces to exercise a greater
influence on the vote decision (Dalton and Wattenberg 2000). As originally
expounded in the context of US presidential elections, prominent among
these short-term forces were policy issues and candidates (with
parliamentary party leaders being the functionally equivalent competitors
for the position of chief executive) (Campbell et al. 1966). The door was
thus opened wider to election-specific influences on the vote decision and
party leaders stepped over the threshold (McAllister 2007).

The second common reason given for enhanced leader effects in
parliamentary elections is the transformed media environment in which
these contests now take place. In particular, television gradually displaced
newspapers and other communication forms as political parties? preferred
campaigning medium and voters? principal source of political information.
This shift partly reflected changing media consumption patterns in the
country at large and partly reflected the desire of increasingly catch-all
political parties to reach beyond their traditional support base to attract
votes. Not only did television allow them to reach unprecedented numbers
of voters, but it did so when they were in the comfort of their living rooms
with their partisan defenses relatively low. The cost, though, was that
parties had to adapt their campaigning strategies to the presentational
?logic? of their new communications medium of choice (Altheide and
Snow 1979). In particular, parties had to come to terms with television
being a medium of communication that is better suited to the projection of
personality than the discussion of complex issues. Thus, the foundations of
a new prominence role for party leaders in a television age were laid.

Leaving these foundations aside, the key question that these
developments pose for political scientists relates ultimately to the nature of
leader effects. What is it about party leaders that gives them ?added value?
in the eyes of voters and what determines the extent of this value? The
answers to these questions are not simple since electoral effects vary
according to the leaders themselves, as well as across space and time. It is
the purpose of this chapter to explore some of the major sources of this
variation.

Leader effects

If only by virtue of their position of institutional leadership, party leaders
can influence any number of political outcomes. They are usually, for
example, the principal driving force behind government formation as well
as policy proposals and outcomes. In specifically electoral terms, however,
there are two major effects that they can have on voters. On the one hand,
there is a reinforcement effect whereby, through their personality or actions,
they strengthen partisan loyalties to the party that they lead in the election.
On the other hand, there is a defection effect. This can be defined as the
?added value, in electoral terms, that a specific ? candidate is able to bring
his/her party or coalition through the effectiveness of his/her public image
as appraised at that specific time? (Barisione 2009: 474). That is, through
their personality or actions, a party leader persuades partisans of other
parties to leave the party for which they usually vote (or for which they
voted in the last election) to cast their ballot for the party he/she leads.
Given that the conventional wisdom is that election campaigns, and the
specific medium of newspapers in particular, reinforce political attitudes
and behaviors rather than change them (Berelson, Lazarsfeld and McPhee
1954; Butler and Stokes 1969), this chapter focuses primarily on the
dynamics of defection. Our starting point is that this dynamic is now
common, but it is not uniform in strength across all leaders, at all times or
in all places. Rather, we argue, the magnitude of leader effects varies with
the personalities of the leaders themselves, their institutional environment
and media coverage of them.

Personality

The unanimous consensus is that it is the psychological variable of
personality that draws voters to party leaders. The public values certain
personality characteristics in its political leaders and these can be
sufficiently attractive to persuade voters to deviate from habitual voting
choices and cast their ballot for another party. Personality itself is
conceptualized either as overall affect for the leader or as a set of character
traits that, for voters at least, suit her/him to the position of chief executive
and leader of the government. In the former case, affect is measured by
voters giving each party leader a score on a thermometer scale ranging from
0 to 10 and usually anchored by ?strongly dislike? and ?strongly like?
respectively. In the latter case, voters ascribe to party leaders? individual
character traits, like competence, deemed desirable in a chief executive.

This emphasis on leader personality has led to their coming to enjoy an
increased electoral influence over time, a process labeled the
?personalization? (or sometimes even ?presidentialization?) of politics. This
?personalization? thesis has sometimes been misinterpreted as implying that
the electoral impact of party leaders increases with each passing election.
But while there is some evidence of such an upward trend in leader effects
over certain periods of time in Australia (Hayes and McAllister 1997),
Britain (Mughan 2000) and the United States (Wattenberg 1994), other
studies have challenged this conclusion and claim to show a decrease or no
change in their magnitude either within single countries over time (Clarke
et al. 2004; Gidengil and Blais 2007) or cross-nationally (Aardal and Binder
2011). Even when the attractiveness of leaders to voters is measured in the
same way, this disagreement is only to be expected. For a start, there is
variation in the popularity of party leaders both relative to each other and
over time for the same individual. Such variance means that, depending on
the larger context of specific elections, the ability of a leader to attract
defectors from another party can go down as well as up. It may be, for
example, that other electoral forces, like the state of the economy or
involvement in a foreign war, can come to the fore and overshadow party
leaders in one election more so than was the case previously.

When it comes to the character traits that attract voters to the leaders of
parties other than the one they consider to be their own, no definitive list of
such traits has been identified so that different studies rely on different traits
to measure the attractiveness of leaders to voters. However, one
comprehensive attempt to compile a set of traits from the literature and to

test for their presence in Dutch newspaper articles has come up with the
following list: political craftsmanship (including competence), vigorousness
(including strong leadership), integrity, communicative skills and
consistency (Aaldering and Vliegenthart 2015). When it comes to the
question of the electoral impact of specific character traits, however, there
are several areas of disagreement in the literature. One such area pits those
who take the view that voters look for the same uniform set of traits in all
party leaders (Miller, Wattenberg and Malanchuk 1986) against those taking
the position that the traits that are important for voters can vary over leaders
and for the same leader over time (Hayes 2005). A second area of
disagreement concerns which traits matter most to voters who fall prey to
the influence of a leader of a party other than the one that usually
commands their loyalty. One side advocates for the primacy of
performance-related traits as competence and reliability (Miller, Wattenberg
and Malanchuk 1986) and others claim to show the greater persuasive
power of character-related traits, like integrity and empathy (Bittner 2011).
Finally, there is the question of the relative potency of perceived positive or
negative leader traits in encouraging voting defection among party loyalists.
Those who opt for the primacy of negative evaluations base their argument
on prospect theory in social psychology, which holds that voters respond
more strongly to negative impressions of political parties and their leaders
than to positive ones (Klein 1991). Other studies, in contrast, show
empirically that the pull-factors in a party leader?s image are more
influential for voters than the push-factors (see, for example, Aarts and
Blais 2011).

But commonly being based on studies of single elections at single points
in time, these contrasting conclusions tend to ignore change over time in
either leader images or the impact of specific leader traits on voters. If only
for this reason, they probably represent false dichotomies that disguise a
more complex reality. Put differently, the characteristics desired in a leader
likely change with the circumstances of individual election contests. Being
competent might be a trait valued in a context taking place in hard times,
but being caring could well trump competence when times have improved.
In the bitterly fought 1983 British general election, for instance, the
perception that the prime minister, Margaret Thatcher, and opposition
leader, Michael Foot, were caring mattered little for voting patterns. By
contrast, the same perception of Prime Minister Thatcher in the less

polarized 1987 contest played a significant and substantial role in
persuading Labour identifiers to vote Conservative (Mughan 2000: 67?68).
It is probably no coincidence that contributing to this change were Mrs.
Thatcher?s great efforts between the two contests to moderate her public
image so as to appear less strident, more understanding and more
sympathetic to voters. Among these efforts were changes in dress and hair
style as well as voice lessons to lower her natural pitch level and moderate
what was described as a ?grating, relentless monotone that drove half the
nation into paroxysms of irritation? (Young 1991: 429).

Political institutions

If the party leaders themselves are an important source of variation in the
magnitude of leader effects in democratic elections, so too is the
institutional architecture within which those elections take place. This
architecture is, of course, highly complex and many parts of it may offer
opportunities for party leaders to influence voters. Compulsory voting, for
example, is likely to bring to the polls a considerable number of voters who
lack deep party loyalties or interest in the election. Under such conditions,
one might expect the impact of factors such as the personality of the party
leaders to have a greater influence on voting behavior. There is any number
of such possible institutional influences so limitations of time and space
lead us to focus on three of the most widely recognized of them: regime
type, electoral/party systems and political parties.

Regime type
Institutions matter for leader effects and perhaps the most important
institutional difference conditioning them is form of government,
presidential or parliamentary (Ohr and Oscarsson 2011). Two dimensions of
difference would seem to be particularly relevant here. First, in presidential
systems, the party leader stands for the position of chief executive
separately from the rest of the party?s office seekers and voters choose
directly between the competing candidates in a nationwide vote. In
parliamentary systems, by contrast, the road to the position of chief
executive (prime minister, chancellor, or whatever) is indirect. The would-
be chief executives stand for election in just one single- or multi-member

district and as one among many representatives of a political party seeking
office; voters do not get the chance to vote directly for their next head of
government. Instead, the choice is made after the election is over and the
distribution of parliamentary seats determined. The leader of the party best
able to command a majority is elevated to the position of chief executive by
the parliamentary majority. In other words, leader effects are likely to be
stronger in presidential than parliamentary systems because voters choose
an individual, who is nonetheless the most prominent representative of a
party, in the former, and a party in the latter (Wattenberg 1991).

Second, presidents are elected for a fixed period, can claim to have
received a personal mandate and need not always carry their party along if
they are to govern effectively. Prime ministers, in contrast, need to maintain
majority support in the legislature to govern, which means that they are less
autonomous and less easily held responsible for the actions or inactions of
the government they head. Blame is relatively easily shifted to the likes of
incompetent cabinet colleagues, recalcitrant coalition partners or an
uncooperative legislature. Relatively isolated, presidents are less able to
shift blame in this way and are more readily held accountable by voters for
political outcomes, thereby functioning under normal circumstances as a
bigger influence on the vote for themselves or their party. The notion of
?presidentialization? captures the argument that prime ministers have
become like presidents in terms of now having the possibility of enjoying
an electoral influence in their own right and of being far more important
cues for voters than other senior figures in their party (Mughan 2000;
Poguntke and Webb 2007).

Electoral/party systems
Institutional differences within parliamentary regimes can also influence
whether the magnitude of leader effects is more or less similar to that found
in presidential regimes. Majoritarian electoral systems tend to produce two-
party systems and single-party governments, whereas proportional systems
promote multiple parties and coalition governments. Leader effects are
stronger in the former precisely because the leaders of single-party
governments can be held more readily accountable for the successes or
failures of their government; blame cannot be diverted onto others so easily

and, in addition, there is no need to share the credit for political successes
(Curtice and Hunjan 2011).

A second characteristic of party systems potentially relevant to the
magnitude of leader effects is the ideological character of the parties
themselves. Where parties are deeply rooted in social cleavages and
ideologically distinctive, there is little likelihood that relatively transient
influences like party leaders will disrupt habitual voting loyalties. This is
probably the main reason these leaders were not taken seriously as electoral
forces prior to the onset of partisan dealignment. Starting in the last part of
the twentieth century, however, political parties in most Western
democracies became more alike in terms of their political ideologies and
goals for society (Franklin, Mackie and Valen 1992). This transition from
?mass-based? to ?catch-all party? systems weakened habitual voting
loyalties and encouraged the emergence of party leaders as potentially
potent electoral forces in their own right (Costa Lobo 2014). The converse
is also true, of course. The greater the ideological divergence between
parties, the less room there is for shortterm influences on the vote, like party
leaders (Holmberg and Oscarsson 2011).

Parties still matter
The focus on party leaders implies a diminished role for the institution of
the political party in shaping the electoral calculus of voters, but its decline
should not be overstated. Partisan dealignment notwithstanding, the
majority of voters continue to demonstrate a long-term commitment to a
particular party and this identification still has a very strong influence on
their vote (Bittner 2011). Along with leader evaluations, for example, party
identification lies at the core of the recently floated valence model of voting
(Clarke et al. 2004). But party is also an important short-term influence on
the vote and election-specific evaluations of parties have non-trivial
implications for the magnitude of leader effects. The norm is for individual
party leaders to be treated as separate entities in the leader effects literature,
but it turns out that electoral impact is substantially stronger when they are
conceptualized not as stand-alone stimuli, but as objects that voters evaluate
relative to the parties on offer in the election. Of course, leaders affect the
vote when they are treated as stand-alone stimuli, but defection is
substantially more likely when voters like the leader of another party in that

election more than they like their own party; the greater this gap, the greater
the likelihood of defection at the polls to the favored leader?s party
(Mughan 2015). To be sure, leaders matter, but the fact that political parties
remain a continuing multifaceted reference point for voters should not be
lost from sight in discussions of leader effects.

Media

As the principal information source for voters, the mass communications
media lie at the heart of the electoral calculus of voters. Initially, the
dominant medium of communication between governors and governed was
the newspaper and early research found that it tended to reinforce long-
standing political predispositions and behaviors rather than to change them.
The reason was a combination of voters exposing themselves only to
newspapers that echoed their existing political biases and their having
defensive psychological mechanisms, like selective recall, that shielded
these biases against discordant information (Berelson, Lazarsfeld and
McPhee 1954; Butler and Stokes 1969). With the primary intent of
reinforcing their electoral support rather than changing it in a world where
partisan loyalties ran deep, political parties mostly controlled their media
environment.

In more recent years, this control has been seriously eroded by the
emergence of television as voters? preeminent source of political
information and, consequently, as an important influence on their attitudes
and behaviors. The essential problem for political parties generally is that
they cannot control this medium and state-licensed broadcasters are
commonly required by law to be impartial between them in their political
coverage (Gunther and Mughan 2000). Thus, not being able to control the
new medium, parties had to adapt to its logic as the first step in harnessing
its power to their own electoral ends. This adaptation entailed above all a
shift from a pattern of political communication dominated by a party logic
in which political parties played the major role in determining what is
politically newsworthy to a media logic in which a wholly different style of
political communication built around personalized coverage, visualization,
simplification, negative coverage, horse race coverage and framing politics
as conflict is the norm (Altheide and Snow 1979; Mazzoleni 1987). For the
purposes of this chapter, the most important consequence of this switch in

the pattern of political communication was an enhanced role for party
leaders in the media coverage of politics in general and election campaigns
in particular. The reasons for this ?personalization? of television coverage
of politics are simple. The majority of political coverage takes place
through news programs and television newscasts that lend themselves to
short, snappy sound bites and the projection of personality more than to the
outline and discussion of complex political issues. Additionally, television
needs a (familiar) visual image to cover political news stories (McAllister
2007). Moreover, the more authoritative the person behind the visual image,
the better the television so that party leaders became preferred over other
senior political figures. Inevitably, the party leader became the
spokesperson and public face of the party. In the process, the content of
media coverage of election campaigns shifted. Less attention was paid to
parties? proposals, performance and policy plans and more to their leaders,
but with a focus less on their political credentials, issue positions and
promises and more on such non-political characteristics as family, personal
appearance, life-style, upbringing and religion (Langer 2007).

The upshot of these changes in the media landscape is that the party
choice of the many voters has come to be influenced by televised leader
images. The strongest direct evidence for such a media effect comes from
the United States where sophisticated experiments have demonstrated not
only that presidential candidates have an electoral impact through voters?
exposure to them on television, but they have also identified the
psychological mechanisms through which this impact makes itself felt.
These mechanisms are labeled priming and framing. Priming entails that
?news content suggests to news audiences that they ought to use specific
issues as benchmarks for evaluating the performance of leaders and
governments? (Scheufele and Tewksbury 2007: 11). Framing, by contrast,
involves television in presenting political and social problems in such a way
as to place blame on different sets of actors and institutions. For example,
unemployment can be framed in the media as a product of presidential
policies or as the result of economic developments. In the case of the
former, voters could well lay the blame for unemployment at the president?s
door and decide not to vote for him (Iyengar and Kinder 1987; Iyengar
1992). Although comparable systematic evidence regarding the role of
priming and framing on leader effects in parliamentary elections is lacking,
there are indications that here too television plays an important role in

shaping voters? perceptions of politicians and the credibility of their
messages. One quasi-experimental study of BBC and ITV news broadcasts
in Britain, for example, found that leader effects were not dependent on the
informational content in the news reports. Instead, if a party succeeds in
convincing voters that the television news is biased in favor of a rival, then
this perception of bias will result in a stronger electoral impact for the
leader whose party is deemed the victim of this bias (Mughan 1996).

The importance of television for leader effects is nowhere more apparent
than in the evolution of the institution of the televised leader debate.
Starting with the famous 1960 confrontation in the US between John
Kennedy and Richard Nixon, debates have been a continuous and integral
part of presidential election campaigns there since 1976. As well, the
contagion has spread. A clear indication that political parties now give a
more prominent electioneering role to their leader is that debates between
party leaders have now taken root in 14 other democracies, most of them
parliamentary systems without the popular election of a president
(Argentina, Australia, Brazil, Canada, France, Germany, Ireland, Kenya,
Malta, Mexico, the Netherlands, New Zealand, the United Kingdom and
Uruguay). There is also convincing evidence that these debates are
important and consequential for voters. For example, they help voters to
position parties on policy issues more accurately (Van der Meer, Walter and
Aelst 2016). Additionally, they influence how heavily voters rely on their
perception of leaders? personalities in their evaluations of them (Druckman
2003). There is even some preliminary evidence to suggest that televised
debates might constitute the most influential television broadcast form for
voters, albeit perhaps only in a system where paid political advertising is
illegal. Despite 2015 being only the second British general election to
include leader debates, a survey of over 3,000 people found that 38 percent
of them claimed to be influenced by them as compared to 23 percent who
were influenced by TV news coverage and 10 percent by party political
broadcasts (BBC News 2015). This pattern of responses is, of course, no
more than suggestive, not least because the ?influence? it measures does not
indicate whether, and in what measure, these debates reinforce or
undermine long-standing attitudes and behaviors. Nonetheless, it does
suggest that the interaction between party leaders and the media has become
a potent electoral force that we are only beginning to understand.

A good starting point might be the improved specification of the
characteristics of voters who are susceptible to mediatized leader images
when making their vote decision. To be sure, something is known about the
psychological processes, like priming and bias perception, that help to
translate leader evaluations into an actual voting decision, but there is little
agreement on the political and sociological characteristics of those
succumbing to the pull of the party leaders. The conventional wisdom used
to be that political leaders are most central to the decision making of
unsophisticated voters: ?The common ? indeed universal ? view has been
that voting choices based on policy concerns are superior to those based on
party loyalty or candidate images. Only the former represents clearly
sophisticated behavior? (Carmines and Stimson 1980: 79). Some empirical
evidence is consistent with this view, showing that leader effects are indeed
strongest among less educated or politically sophisticated voters (Gidengil
2011). The explanation of this finding is usually that leader effects can be
considered as heuristics or shortcuts for voters who lack the reasoning
capacity and/or political knowledge to make a decision based on ?policy
concerns.? However, other scholars find precisely the opposite and show
that leader effects are actually strongest for more highly educated and
politically sophisticated voters (Bittner 2011; Lachat 2014). Their
explanation is that leader effects are not just shortcuts. Rather, the
evaluation of a leader is based on the voter?s assessment of how well the
leaders performed, or would perform, in office so that the ?process of leader
evaluation is complex ? perhaps more so than we think? (Bittner 2011: 54).
Closer analysis of the increasingly popular televised leader debates should
help to adjudicate between such conflicting conclusions.

Future research

One of the more consequential electoral developments in recent decades has
been the generalized growth of the importance of party leaders as electoral
forces in their own right. From an initial consensus that they used to matter
little to not at all in parliamentary elections in particular, the argument has
been made that they can even be the difference between victory and defeat
for their party in closely fought elections (Bean and Mughan 1989, but see
King 2002). To argue for leaders? independent electoral impact, however, is
not to imply that this impact is uniform across them, time or space. Rather,

the central theme of this chapter is that the magnitude of leader effects is
conditional on a number of factors and, among them, we have examined
briefly the personalities of the leaders themselves, the political institutions
in which leaders operate, the parties in the election and the media. But this
?conditionality research? is only in its infancy; much remains to be done to
specify the conditions shaping the magnitude of leader effects and we
would like to conclude this chapter with some suggestions for future areas
of research. Reflecting much of what has already been written on the
subject of leader effects, we will focus on the personalities of the party
leaders and the media, both old and new.

Dating from at least Max Weber?s writings on charismatic leadership, the
importance of personality for the political relations between leaders and led
has been recognized and this tradition has defined leader effects studies to
this point. Voters respond to party leaders as individuals they like or dislike,
respect or disrespect, and so on. But if leaders have increasingly become the
public face of the party, then the question inevitably arises as to whether
voters? reactions to them have a political as well as personality content.
Especially the longer they stay at the head of their party, do leaders? images
become defined by their party?s traditional ideological stance or its current
policy positions as much as (or even more than) by their personalities? A
case in point is if voters do indeed defect because, in the context of a
specific election, they like the leader of another party more than they like
their own party (Mughan 2015), what can be the basis of their choice to
defect? After all, if political parties do not have personalities, it can?t be
personality that is being compared in the voter?s mind. Voter preferences
must be based, at least partly, on other criteria and surely political criteria
can be expected to loom large here.1 Thus, future research should
?politicize? leader effects instead of just continuing to ?psychologize? them.
This is not to say that the latter are not important to voters, but the
possibility should be entertained that both politics and psychology play a
role in conditioning leader effects so that the interesting question becomes
under what conditions one becomes more powerful than the other in
aff