Content uploaded by Ishan Kashyap Hazarika
Author content
All content in this area was uploaded by Ishan Kashyap Hazarika on Apr 27, 2022
Content may be subject to copyright.
Asian Review of Social Sciences
ISSN: 2249-6319 Vol. 9 No.1, pp. 27-33
© The Research Publication, www.trp.org.in
An Investigation of Beliefs, Information and the Halo Effect in Electoral
Decision Making
Ishan Kashyap Hazarika1 and Sourabh Rai2
1&2Student, Department of Economics, Hansraj College, University of Delhi, India
E-mail: kpishanh@gmail.com
Abstract - Rational ignorance suggests that voters largely
ignore a lot of information while voting due to the high cost of
attaining and processing the information. It is further
suggested that rational voters do not vote to affect election
results but to express opinions. It is thus likely that cognitive
biases shape electoral decision-making. The Halo effect, for
instance, extrapolates information in one domain to another
and helps voters avoid processing extra information. In this
paper, we investigate the conditions under which extra
information is processed or ignored, and first impressions are
generalised. We find, through a Randomised Control
Experiment, that new and weakly formed political beliefs also
have effects like strongly held political beliefs, on information
provided later. In particular, the study presented picture-
information about candidates, either accompanying or not
accompanying text-information. Additional text-information
did not significantly change voter-choice when the text
information reaffirmed picture-based preferences but did
significantly change voter-choice when it contradicted picture-
based preferences. These results are viewed from the
perspective of both the Identity-Protective Cognition Thesis
and the Halo effect, thus hinting that the two may be
connected, an insight that is largely missing in the previous
literature.
Keywords: Halo effect, Identity-Protective Cognition Thesis
(ICT), Rational ignorance, Randomized Control Experiments.
I. INTRODUCTION
The relevance of cognitive biases in explaining human
behaviour is being recognized in recent times in economics.
Similar biases, undoubtedly play an important role in other
manifestations of human behaviour too, such as political
decision-making. At the same time, the role of information
can also be scrutinized in the voter's decision problem. A
first glance immediately relates it to the idea of 'rational
ignorance', where the voters ignore relevant information
when making voting decisions as the cost of processing the
information is high and the expected benefits, due to the
usually large size of the electorate, low. The interplay
becomes even more interesting, as the same cost-benefit
argument is used by Downs (1957) to argue that rational
decision-makers would exhibit extremely low voter turnouts
(contrary to what is practically observed) in the famous
Voter's Paradox. This has prompted many to conclude that
voting is done not to influence election outcomes, but as a
tribal ritual to show support or other emotional benefits.
This brings us back to the domain of cognitive biases and
psychological effects from yet another direction.
In this paper, we investigate one such bias in the realm of
electoral decision-making by voters- the Halo effect. The
Halo effect can be defined as "the tendency of an opinion
created in one area to influence opinion in another area".
For example, views such as "a successful sportsperson
would also be good at relationships", or "a veteran is likely
to be a good leader too" etc. can be argued to follow from
the Halo effect. The tendency of the first impression having
a lasting impact can also be explained using this cognitive
bias.
A presence of this bias in electoral decision making is
already established. Several studies have found empirical
support for its presence using econometric as well as
psychometric techniques. The presence of the effect has also
been related to and studied vis-a-vis personality traits. In
this paper, we attempt to explore a hitherto largely
unexplored area of how the effect interacts with
information.
The study of cognitive biases and psychological effects as
an explanation of electoral behaviour in contrast to
information effect as usually studied in the rational choice
framework is common, but the study of both as influencing
each other and interacting is as of now relatively left out. In
this paper, we seek to at least partly to fill this gap in
research. Electoral behaviour can be studied from real-
world data as has been done many times. But to accomplish
the task, we use the method of Randomized Control
Experiments, as it serves well to isolate the Halo effect and
the effect of information on decision-making in the
surveyed population.
II. LITERATURE REVIEW
An early discussion of the Halo effect can be found in
Nisbett and Wilson (1977). Several studies have since then
extended and applied this idea. The effect of various
27
ARSS Vol.9 No.1 January-June 2020
attributes of a person on perceived political qualities, such
as competence, trustworthiness etc. has been studied.
In particular, the manifestation of the Halo effect as the
“attractiveness stereotype” (Thorndike,1920), where
physical attractiveness may influence the impression of
other qualities, which we use in this paper has been studied
before. Several studies have found the existence of this
effect in political settings (Atkinson et al., 2009; Surawski
and Ossof,2006; Berggren et al.,2010; Lenz et al., 2011).
The existence of this effect is especially noticeable in
rating-data. (Kozlowski et al., 1986).
Research in political science has established the existence of
positive effects of attractiveness and perceived competence
on electoral success (Berggren et al., 2010; Mattes and
Milazzo,2014; Lawson et al.,2010; Little, Anthony, 2014).
Studies have been conducted to test this in very different
scenarios such as war and peace (Little et al. 2007; Little et
al. 2014). The effect is especially pronounced when voters
have low information (Lenz and Lawson, 2011). We shall
leverage this finding in our experimental design to ensure
the adequate presence of the Halo effect to study the
interaction with information.
Studies have shown that physical appearance affects not
only electoral outcomes but also other postings and rankings
in politically sensitive areas. Mazur et al. (1984) and
Murray (2014) have found that physical appearance can be
used to predict military rankings and promotions and
wartime duties respectively. The effect has also registered
presence in military coups and dictatorships (Bartelson,
1997; Avril, 1999)
It has been argued that look-based preferences are formed
unconsciously (Olson & Marshuetz, 2005; van Leeuwen &
Macrae, 2004).These preferences are found to be almost
uniform at least within a cultural group
(Cunningham1995).Some believe that these preferences
could even be universal (Berscheid & Walster, 1974;
Hatfield & Spreche,1986 and Langlois et al., 2000).Murray
(2014) in fact views the results from military data from an
evolutionary lens rather than a socio-cultural lens alone.
The study of information and electoral decision making has
also been undertaken from both theoretical and empirical
standpoints. An early theoretical exploration of political
decision-making, using a rational-choice framework is
found in Downs (1957). Downs argued that the decision to
vote in an election with a sizeable electorate is extremely
unlikely for rational decision-makers who only seek to
influence the election results. This point was also made by
Condorcet in 1793, as observed in Mclean and Hewitt
(1994). Responses have included the evoking of evolved
altruism (Fowler, 2016) to arguing that voting is rewarding
even if one's vote does not actually affect the outcome
because fulfilling "civic duty" yields utility (Riker and
Ordeshook, 1968). This interestingly leaves enough room
for a behavioural investigation into why are people voting,
and if people do not care if their vote actually affects the
outcome, on what basis do they chose which candidate to
vote?
This bears a direct relation to the idea of rational ignorance
discussed above. People may very well choose to ignore
relevant information on candidates and political affairs, and
rationally so. Voters then rely on heuristics and emotions to
decide their candidate of choice. This finally invites our
study in this paper, investigating the interplay of the
cognitive bias - the Halo Effect and information.
It is important to remember that these prejudices of
attractiveness in articulated preferences not only
automatically arise but also tend to remain in the light of
additional candidate information. Among other factors, the
above studies provided controls on partisanship and gender,
showing that the presence of certain characteristics in the
candidates often increases perceived competence and
electoral performance in varied settings (Palmer and
Peterson, 2015).
In this paper, we seek to uncover conditions under which
information plays a role or does not. In particular, we
provide people information either confirming or opposing
their already formed look-based perceptions and analyse the
differences.
III. EXPERIMENTAL DESIGN
In this paper, we use the method of randomized control
experiments to isolate the Halo effect, the effect of
information when it reinforces look-based preferences and
when it contradicts look-based preferences. In a
conventional RCE, the attributes of the population are either
known or controlled using a control group. We randomize
all attributes other than look based preferences and quantify
look based preferences in a controlled group. We then
compare results from two separate groups, in one of which
we provide preference reinforcing information and in the
other contradicting information. To minimize the effect of
the complexity of information, we provide crisp and point
wise information in both cases.
A. Survey Design
We designed three distinct survey forms for the three
groups. For the control group, we include only one picture
of each of two candidates. We explain to them in the form
that they are expected to vote for the post of the mayor of a
fictional city. They are asked to vote whom they find more
fit to be the mayor of the city-based solely on looks. Do
note, that we did not ask them to merely choose the one they
find more "attractive", but the one they find more fit for the
political post. We are cognizant of the fact that people may
use different physical features to infer performance in
different activities. Asking directly to vote for the mayor's
28
ARSS Vol.9 No.1 January-June 2020
Ishan Kashyap Hazarika and Sourabh Rai
post ensures that the effect of the looks on political choice
only is covered.
We did not quantify the information effect separately. This
was deemed unnecessary for our purpose as we seek to
understand what information adds to look based preferences
and not independent of it.
For the second group, we provide the same pictures again
but, in this form, they each accompany a distinct box of
information about the candidates' educational qualifications,
achievements and past performance.
Note that, with these two forms only, we are still not able to
assess whether the information reinforces look based
preferences or contradicts them.
For the third group, we again provide the same pictures and
the same boxes of information, but the boxes of information
are switched with each other.
By comparing the groups, we can assess the effect of
information in the manner outlined below.
B. Isolation of Effects
Let 𝐿1, 𝐿2 capture the percentages of the vote based solely
on looks. Regarding the information effects, we can
consider them to be independent of look-based preferences
or dependent on them. If we assume them to be independent
and let 𝐼
1,𝐼2 represent the change in vote share due to the
additional information, the following shall be the
interpretation of the results.
TABLE I INTERPRETATION OF GROUP-WISE DATA ASSUMING
INDEPENDENT INFORMATION EFFECTS
Group I
(Control)
Group II Group III
Candidate I
𝐿1
𝐿1+𝐼
1
𝐿1+𝐼2
Candidate
II
𝐿2
𝐿2+𝐼2
𝐿2+𝐼
1
To find 𝐼
1and 𝐼2 we simply subtract vote shares of
Candidate I in group I from group II and group I from group
III respectively to get
𝐼
1=(𝐿1+𝐼
1)− 𝐿1 (1)
𝐼2=(𝐿1+𝐼2)− 𝐿1 (2)
Similarly, for Candidate II we get the following equations
𝐼
1=(𝐿2+𝐼
1)− 𝐿2 (3)
𝐼2=(𝐿2+𝐼2)− 𝐿2 (4)
Because all the entries in the table are vote shares, therefore,
𝐿1 +𝐿2= 1 (5)
𝐿1+𝐼
1+𝐿2+𝐼2= 1 (6)
⇒ 𝐼
1=−𝐼2 (7)
Because of equation (7), if we find an 𝐼
1 significantly not
equal to 0, then 𝐼2 shall also be not equal to 0 with
significance. Furthermore, the signs are expected to be
opposite. Because of this, conditional on this hypothesis, if
we see a significant result in group II, we shall also see
significant results in group III and in opposite directions.
On the other hand, we may drop this assumption that
information effects are independent of look based
preferences. In this paper, from the numerous possibilities,
we hypothesize as a second hypothesis that the information
effect depends on whether the additional information
reinforces the look-based preferences or contradicts them.
Here then, 𝐼
1,𝐼2 are not unknown constants but unknown
functions from the set {0,1} to 𝑅, where 1 indicates that the
information reinforces the look based preferences and 0
indicates contradiction with look based preferences. If
𝑥 ∈ {0,1}, then the group data indicates the following:
TABLE II INTERPRETATION OF GROUP-WISE DATA ASSUMING
DEPENDENT INFORMATION EFFECTS
Group I
(Control) Group II Group III
Candidate I
𝐿1
𝐿1+𝐼
1(𝑥)
𝐿1+𝐼2(1
− 𝑥)
Candidate
II
𝐿2
𝐿2+𝐼2(𝑥)
𝐿2+𝐼
1(1
− 𝑥)
Here again, as the table entries are vote shares, we have,
𝐿1 +𝐿2= 1 (8)
𝐿1+𝐼
1(𝑥) + 𝐿2+𝐼2(𝑥) = 1 (9)
𝐿1+𝐼
1(1 − 𝑥) + 𝐿2+𝐼2(1 − 𝑥) = 1 (10)
Note that in this case 𝐼
1(𝑥) and 𝐼2(1 − 𝑥), which are the
information effect for candidate I in group II and III
respectively are independent of each other.
It can be shown that 𝐼
1(𝑥) = −𝐼2(𝑥) and 𝐼
1(1 − 𝑥) =
−𝐼2(1 − 𝑥), but this is not only mathematically trivial,
given that the vote shares add up to 1, but are also
tautologically true regardless of the hypothesis chosen and
thus empirically unfalsifiable. In this sense, this hypothesis
may appear operationally meaningless. This serves as a
framework to interpret results only. A finding from the
result can be used to augment this hypothesis and thus,
29
ARSS Vol.9 No.1 January-June 2020
An Investigation of Beliefs, Information and the Halo Effect in Electoral Decision Making
render it operationally meaningful. We seek to do this too in
this paper.
The key takeaway for our empirical purpose here from this
hypothesis is that significant results in group II do not imply
significant results in group III.
We can measure 𝐼
1(𝑥) by subtracting Candidate I’s vote
share in group I from that in group II, and 𝐼2(1− 𝑥) by
subtracting the vote share in group I from group III.
𝐼
1(1 − 𝑥) And 𝐼2(𝑥) can then be deduced simply by using
the negative relations outlined in the previous paragraph.
C. Sample and Sample Size
We had access to a list of students from various colleges of
the University of Delhi. We further gained access to
contacts of students and working professionals from various
parts of India. From this list, we selected 252 people at
random. To each of these 252 people, we sent one and only
one of the three forms at random. Thus, who among the 252
people received which form was completely random. We
had a total of 146 responses.
Tackling Selection Bias: The questions were not revealed
to the respondents before they agreed to respond. Thus, they
could not have decided to respond based on which form
they received. Their decisions were completely independent
and thus, the bias of self- selection was not possible. Again,
there could have been differences between the
characteristics of people who chose to respond and those
who did not, but this bias did not penetrate our results
because we did not compare the results of the people who
selected against those who did not. We only compared sub-
groups of the responding population.
Within the group of people who did choose to respond, the
allocation of the forms was completely random and
independent of the decisions made by them. This has
increased our confidence in the fact that the results are
largely unbiased.
IV. RESULTS
A. Descriptive Statistics
We treat each vote as a Bernoulli random variable assigning
a value of 1 for a vote to Candidate I and 0 for a vote to
Candidate II. Note, however, that these variables are
different from the preference reinforcing-contradicting
variable defined before, and represent only voting choices,
not group preferences. The mean of each group is simply
the vote share of Candidate I. The statistics for Candidate II
can be easily deduced.
The following are the descriptive statistics for the three
groups:
Group I (Control Group)
TABLE III DESCRIPTIVE STATISTICS FOR GROUP I
Sample Mean 0.75
Sample Standard Deviation 0.43
Total Number of votes 57
Candidate I, with 75% of the 57 votes is the winner. We
infer that looks suggest Candidate I to be a better choice
than Candidate II.
Group II
TABLE IV DESCRIPTIVE STATISTICS FOR GROUP II
Sample Mean 0.63
Sample Standard Deviation 0.50
Total Number of votes 32
Candidate I am the winner again with 63% of the votes.
Although the margin appears to have reduced, we shall test
for its significance later.
Group III
TABLE V DESCRIPTIVE STATISTICS FOR GROUP III
Sample Mean 0.54
Sample Standard
Deviation
0.50
Total Number of votes 52
Candidate I am the winner for this group too, with an even
lower margin. We shall test for the significance of this drop
too later.
Although we have not yet tested for the significance, we can
comment on the prospects of the information effects being
independent of the Halo effect even now.
This hypothesis predicted that the signs of differences in
victory margins of group II and III from the control group
would be opposite. But, because the margins are lower for
both group II and III, this hypothesis can be defended only
if the drops in both the groups are statistically not
significant.
V. INTERPRETATION OF RESULTS
To estimate the information effects, we find the difference
of the mean of group I from the means of group II and
group III, respectively. We also check for significance in
both the cases, with a level of significance of 10%. We
assumed equal variances as the groups are randomly divided
from a single population.
30
ARSS Vol.9 No.1 January-June 2020
Ishan Kashyap Hazarika and Sourabh Rai
TABLE VI T-TEST - GROUP I V/S GROUP II
Group I Group II
Mean 0.75 0.63
Variance 0.19 0.24
Observations 57 32
Pooled Variance 0.22
Hypothesized Mean
Difference
0
Degrees of freedom 87
t Statistics 1.29
P(T<=t) one-tail 0.10
t Critical one-tail 1.30
P(T<=t) two-tail 0.20
t Critical two-tail 1.66
TABLE VII T-TEST - GROUP I V/S GROUP III
Group I Group III
Mean 0.75 0.54
Variance 0.19 0.25
Observations 57.00 52.00
Pooled Variance 0.22
Hypothesized
Mean Difference
0.00
Degrees of
freedom
107.00
t Statistics 2.40
––P(T<=t) one-tail 0.01
t Critical one-tail 1.29
P(T<=t) two-tail 0.02
t Critical two-tail 1.66
The observed difference in means of group I and group II is
0.12. With a pooled sample variance of 0.22, the difference
of means is not significantly different from 0.
The observed difference in the mean of group I and group
III is 0.21. With a pooled sample variance of 0.22, the
difference of mean is significantly different from 0.
The direction of Information Effect: We observe by
comparing group I and group III that the mean of group III
is not only significantly different from the mean of group I
but also significantly less than it, as observed from the one-
tailed t-test with the same level of significance of 10%. We
thus believe that the information contradicted the
preferences based on looks in this case.
VI. A DISCUSSION ON THE HYPOTHESES
We commented earlier that the hypothesis that the
information effects are independent of the look-based
preferences predicts, 𝐼
1=−𝐼2. We further commented later
on that we shall expect significant results for both group II
and group III or in none of them but not in one group but
not the other. Contrary to the expectation we found a
significantly negative information effect for group III while
the information effect in group II is not significantly
different from 0. For clarity, we directly test the prediction
of this hypothesis. We observe that according to this
hypothesis the effect of information on Candidate I in group
II is equal to the negative of the effect in group III.
Therefore, adding these is predicted to yield zero. We get
the estimates for each from tables 6 and 7 respectively. The
respective estimates are 0.12 and 0.21. The estimated
variances are 0.011 and 0.008. Because the degrees of
freedom are substantially higher than 30 in each case, we
assume normality. Accordingly, the sum of these variables
would also be normal. The following are the estimated
values:
TABLE VIII TEST OF SIGNIFICANCE ASSUMING I1 + I2 = 0
Mean
0.33
Variance 0.019
Z -value 2.39
P-value 0.017
The estimated mean is significantly different from 0 at a
significance level of 10%. Thus, the hypothesis is rejected.
This brings us to the second hypothesis. The following are
the estimated values for this hypothesis:
TABLE IX ESTIMATES OF VARIABLES ASSUMING DEPENDENT
INFORMATION EFFECTS
𝐼
1(1)
-0.12 Insignificant
𝐼
1(0)
-0.21 Significant
𝐼2(1)
0.12 Insignificant
𝐼2(0)
0.21 Significant
We observe that the information effect is significantly non-
zero when the information contradicts the look-based
preferences while they are not significantly non-zero when
they reaffirm look-based preferences. This may invite a
discussion on how people choose to use their mental faculty
to understand the world around them. An idea that seems
relevant here is the Identity-Protective Cognition Thesis
(ICT). A paper supporting this thesis, against an alternate
hypothesis, for instance, found that people use their skills of
numeracy selectively based on whether the numerical data
presented reaffirms or contradicts the beliefs they already
had. When the numerical data reaffirmed their political
beliefs at a superficial level, even people with high
numeracy skills chose not to use their skills for proper
inspection of the data. On the other hand, when the data
contradicted their beliefs even at a superficial level, people
did choose to investigate further (Kahan et al, 2013).
31
ARSS Vol.9 No.1 January-June 2020
An Investigation of Beliefs, Information and the Halo Effect in Electoral Decision Making
In this paper, numerical data has not been prepared, but
pictorial and verbal data, two points may be noted from the
results of this paper, concerning the ICT:
1. The result that the information effect was insignificant
in group II where the verbal information reinforced the
look-based preferences may indicate that respondents
chose not to thoroughly examine the verbal data, once
it superficially supported the look-based preferences.
This may indicate that the ICT holds even for verbal
data following pictorial data. This may also lead to the
question of whether some forms of data are prioritized
higher than others when it comes to the intuitive
analysis of data. This paper, however, does not test
this claim, or even the ICT directly. These are mere
comments regarding the possibilities in light of the
data analysed here and may serve to inspire future
research.
2. The second point speculates the power of first
impressions. In Kahan et al. (2013), strongly held and
long formed beliefs about politics and society
confronted new numerical data. If the ICT holds, even
in this paper involving pictorial and verbal data, as
contemplated by the authors, then it is astonishing that
beliefs and preferences about the candidates formed
merely by looking at one picture each of the
candidates can induce such strong effects, to prompt
respondents to ignore new data in certain cases, as if
the look based preferences are sacrosanct beliefs. The
refusal to investigate further data thoroughly even
though the data presented earlier was scarce and
pictorial, may indicate that the issue is grave and
widespread.
Seeing these results in the light of the Halo effect, on the
other hand, provides us with an altogether new approach to
explain the data, but given the striking analogy (or even
homology in our view) to the ICT results, to also explain
other ICT results from a new approach. Van Leeuwen &
Macrae (2004) have argued that the Halo effect reflects the
inability of the voters to differentiate between different
characteristics being evaluated, although sometimes
subconsciously. In this light, when the additional
information reinforces look- based preferences, it simply
makes the two sets of information consistent, without
adding much new. On the other hand, when the
contradiction occurs, people integrate the two sets of
information without a proper grasp of which set of
information to use for which purpose.
VII. CONCLUSION
The interaction of the Halo effect and additional
information, at least in the political domain is complex.
How additional information is treated or analysed depends
crucially on beliefs held earlier. While this has been studied
before especially in cases where the pre-existing beliefs
were formed in a complex manner and were deeply
embedded in society, in this paper we allowed the formation
of completely fresh beliefs. Furthermore, the beliefs were
formed based on very scarce data- only one photograph of
each candidate. It appears that even these feebly formed
beliefs prompt respondents to ignore further data in certain
cases.
Whether the respondents’ choices are affected by the new
information depended crucially on whether the new
information reaffirmed or contradicted their look-based
beliefs. In group II, where reaffirmation took place, the
information effect was not significant. In group III on the
other hand, where contradiction took place, the information
effect was significant. The idea of rational ignorance
suggested that a lot of information will be ignored by the
voters. In some studies, the idea of selective processing of
information was introduced where strongly held beliefs
confronted new data. In the current paper, on the other hand,
new and feebly formed beliefs confronted new data.
Interestingly, even the new and feebly formed beliefs based
solely on one picture each of the candidates prompted a
similar effect as the strongly held beliefs.
This suggests that the reliance on the first impression is
strong and this effect may relate to the repulsion to process
new information, suggested by rational ignorance. In that
sense, in combination with the impressive strength
displayed by the feebly formed look based beliefs, it
appears as if political decisions rarely reflect societal
concerns. The electoral results may largely represent vague
instincts .This paper helps us understand a way of forming
these instincts, and also provides the astounding finding that
not only is new information ignored in presence of strongly
held beliefs but even in the presence of very weakly formed
beliefs, based on very scarce data. This shockingly
highlights the extent of ignorance, to be possibly high
enough to render the electoral process very shaky.
REFERENCES
[1] Atkinson, Matthew, D., Ryan, D., Enos and Seth J. Hill. (2009).
Candidate Faces and Election Outcomes: Is the Face-Vote
Correlation Caused by Candidate Selection? Quarterly Journal of
Political Science, 4 (3), 229–249.
[2] Avril, P. (1999). From glory to disgrace: The Haitian army, 1804-
1994. Universal Publishers.
[3] Bartelson, J. (1997). Making Exceptions: Some Remarks on the
Concept of Coup d' etat and Its History. Political Theory 25 (3), 323–
346.
[4] Berggren, N., Jordahl., H. and Poutvaara, P. (2017). The Right Look:
Conservative Politicians Look Better and Voters Reward It. Journal
of Public Economics. 146, 79–86.
[5] Berscheid, E. and Walster, E. (1974). Physical attractiveness. In
Berkowitz, L. (Ed.), Advances in experimental social psychology, 7,
157-215.
[6] Cunningham, M. R., Roberts, A. R., Barbee, A. P., Druen, P. B., Wu,
C. H. (1995). Their ideas of beauty are, on the whole, the same as
ours: Consistency and variability in the cross-cultural perception of
female physical attractiveness. Journal of Personality and Social
Psychology, 68, 261-279.
[7] Downs, A. (1957). An economic theory of democracy. Harper and
Row.
[8] Fowler, J. (2006). Altruism and Turnout. The Journal of Politics. 68
(3), 674-683.
32
ARSS Vol.9 No.1 January-June 2020
Ishan Kashyap Hazarika and Sourabh Rai
[9] Hatfield, E., Sprecher, S. (1986). Mirror, mirror: The importance of
looks in everyday life. State University of New York Press.
[10] Kahan, M., Peters, E., Dawson, E. and Slovic, P. (2013). Motivated
Numeracy and Enlightened Self-Government.Behavioural Public
Policy. 1, 54-86. http://dx.doi.org/10.2139/ssrn.2319992.
[11] Kozlowski, S. W., Kirsch, M. P., Chao, G. T. (1986). Job knowledge,
rate familiarity, conceptual similarity and halo error: An exploration.
Journal of Applied Psychology. 71, 45-49.
[12] Langlois, J. H., Lisa K., Adam J. R., Andrea L., Monica H. and
Monica S. (2000). Maxims or Myths of Beauty? A Meta-Analytic and
Theoretical Review. Psychological Bulletin. 126, 390–423.
[13] Lawson, C., Lenz, G. S., Baker, A. and Myers, M. (2010). Looking
Like a Winner: Candidate Appearance and Electoral Success in New
Democracies. World Politics. 62, 561–593.
[14] Lenz, G. S. and Lawson, C. (2011). Looking the Part: Television
Leads Less Informed Citizens to Vote Based on Candidates
Appearance. American Journal of Political Science. 55 (3), 574–589.
[15] Little, A. C., Burriss, R. P., Jones, B. C., Roberts, S. C. (2007). Facial
appearance affects voting decisions. Evolution & Human Behavior.
28, 18-27.
[16] Little, A. C., Roberts, S. C., Jones, B. C., De Bruine, L. M.
(2012). The perception of attractiveness and trustworthiness in male
faces affects hypothetical voting decisions differently in wartime and
peacetime scenarios. The Quarterly Journal of Experimental
Psychology. 65, 2018-2032.
[17] Little, A.C. (2014) Facial appearance and leader choice in different
contexts: Evidence for task contingent selection based on implicit and
learned face-behaviour/face-ability associations. The Leadership
Quarterly. Vol. 25, pp. 865–874.
[18] Mazur, A., Mazur, J. and Keating, C. (1984). Military Rank Attainment
of a West Point Class: Effects of Cadets’ Physical Features. American
Journal of Sociology. 90, 125–150.
[19] Mclean, I. and Hewitt, F. (1994). Condorcet. Edward Elgar Publishing.
245-246.
[20] Murray, G. R. (2014). Evolutionary preferences for physical
formidability in leaders. Association for Politics and the Life Sciences.
33, 33–53.
[21] Nisbett, R. E., & Wilson, T. D. (1977). The halo effect: Evidence for
unconscious alteration of judgments. Journal of Personality and Social
Psychology. 35 (4) 250-256. https://doi.org/10.1037/0022-
3514.35.4.250.
[22] Olson, I. R. and Marshuetz. C. (2005). Facial attractiveness is
appraised at a glance. Emotion. 5, 498-502.
[23] Palmer, C. and Peterson, R. (2015). Halo Effects and the Attractiveness
Premium in Perceptions of Political Expertise. American Politics
Research. 44.
[24] Surawski, M. and Ossoff, E. (2006). The Effects of Physical and Vocal
Attractiveness on Impression Formation of Politicians. Current
psychology. 25, 15-27.
[25] Thorndike, E. L. (1920). A constant error in psychological ratings.
Journal of Applied Psychology, 4, 25-29
[26] Van-Leeuwen, M. L. And Macrae, C. N. (2004). Is beautiful always
good? Implicit benefits of facial attractiveness. Social Cognition.
22, 637-649.
[27] William H., Ordeshook, P.C. (1968). A Theory of the Calculus of
Voting. American Political Science Review, 62, 25-42.
https://doi.org/10.1017/S000305540011562X
33
ARSS Vol.9 No.1 January-June 2020
An Investigation of Beliefs, Information and the Halo Effect in Electoral Decision Making
