“The impact of celebrity politicians perceptions on political party preferences”

Political parties usually try to understand the factors that affect a voter’s decision in elections. However, irrespective of voters’ preferences, studies rarely examine how voters’ attitudes toward celebrity politicians affected their party preferences, given the mediating effects of brand affinity and perceived attributes and the moderated mediating role of gender. Therefore, this study considers the effect of parties’ nominations of celebrity politicians on voters’ political party preference. Moreover, it investigates the causal relationship between those perceptions in an uprising nation against political parties. A representative national sample of one thousand two hundred sixty-nine (1269) Lebanese voters was administered via a cross-sectional survey in fifteen Lebanese districts. A stratified proportional random sampling technique was used. Findings showed that attitudes significantly affected political party preferences when nominat-ing celebrity politicians only through brand affinity on a 90% confidence level (proba-bility-value = 0.053 < 0.10) and perceived attributes on a 95% confidence level (0.039 < 0.05), evidencing a lack of a significant direct relationship (0.571 > 0.10). Voters’ gender conditional indirect effect was significant for females’ impact on brand affinity (0.025 < 0.05), whereas gender failed to determine voters’ indirect effect on perceived attributes (0.633 > 0.10). The results have shown that gender disparities in the brand’s emotional component could affect brand preferences.


INTRODUCTION
In democracies, the sustainability of political parties in power is explained by the electoral voting representation (Lindstaedt & Frantz, 2019).Democratic countries with proportionality laws and competitive multi-party contexts precede the closed-list systems and usually constrain voters' choices to one candidate or group.Parties' nominations of candidates become crucial (Rehmert, 2020).
Political parties could set effective communication strategies, such as nominating celebrity politicians (CPs) who are famous and from diverse backgrounds with political interests and ambitions but constrained by voters' and political parties' acceptance (Banerjee & Chaudhuri, 2020).However, implementing such a strategy could yield different outputs that would be either detrimental or beneficial (Brockington, 2014).Subsequently, it would be necessary for political parties and celebrities to understand their potential voters' rational and emotional statuses.
Prior research revealed associations between emotional factors and political brand preference (PBP), such as brand love (Banerjee & Chaudhuri, 2022), and rational factors, such as perceived attributes of celebrity politicians (PACP) (Banerjee & Chaudhuri, 2020;Bankert, 2020).However, although previous studies have recognized attitude as the PBP predictor (Reibstein et al., 1980), its impact on the latter in the presence of brand affinity (BA) and PACP has yet to be examined.Addressing this aspect might create opportunities for parties and celebrities.
The voters' acceptance of political parties' nominated CPs is challenging in an uprising nation against its political system.Lebanon, a multi-party country in the Middle East facing specific conflicts in politics and religion, was chosen for this study after recently witnessing a national uprising against political parties (Badaan et al., 2020).The CP strategy, rarely addressed in the literature for the Middle East, was previously experienced in the 2018 Lebanese parliamentary elections.It resulted in a few business people and one journalist being elected (National Democratic Institute, 2018).This study determines voters' attitudes toward nominating celebrities as party candidates for the upcoming elections in a highly disturbed context.

LITERATURE REVIEW AND HYPOTHESES DEVELOPMENT
Considering the limited research dealing with the influence of CPs nominated by political parties on PBP, numerous antecedents of PBP and interrelationships were reviewed.This study addresses the different mechanisms behind voters' PBP relationships with ACP, PACP, and BA dimensions.
Electoral volatility is a common phenomenon where voters change their political party preferences in the elections, seeking better performance (Johann et al., 2018).Voters may distinguish and choose to satisfy their stable preferences above their instantaneous ones (Abel et al., 2021).In contrast, conventional parties try to win voters' preferences by transferring ideologies, interests, and preferences to outperform opposing parties (De Vries et al., 2021).In fact, voters aim to continuously upgrade their living conditions when supporting specific candidates against others, seeking social, economic, and political stability (Banerjee & Chaudhuri, 2022;Bankert, 2020 (Norris, 2004).Certain societies with high celebrities' political presence were found to be linked to uncorrupted and stable nations (Loader et al., 2016).
In contexts with high sectarian formation (Badaan et al., 2020), corruption, and economic crisis, voters may uphold their political identity, thus supporting their preferred traditional political party.Alternatively, they may rationalize their preferences per the party's performance and support different political parties aiming for reformation (Bisgaard, 2015;Burden & Klofstad, 2005).Nominating competent candidates can stimulate social progress (Cwalina & Falkowski, 2015) and voters' choice (Loader et al., 2016).Whereas nominating celebrities as political candidates can generate contradictory outcomes (Brockington, 2014).
Renowned as a PBP predictor (Reibstein et al., 1980), attitude is a psychological path that governs feelings and preferences, generating likes and dislikes (Veselý, 2021).Social psychologists recognize attitude with three interconnecting directions: beliefs related to cognition, feelings related to affect, and behavioral intentions related to conation (Reibstein et al., 1980).
Behavioral research assumes that dependency on preferences is crucial in purchasing decisions.Since brand preference is a behavioral tendency (Ebrahim et al., 2016), political parties pursue acceptance and favorability, and voters generate greater preference (Wheeler, 2013) Supporting nominees from parties that voters are already familiar with can prevent them from changing parties.Regardless of their intellectual or emotional assessments, many voters can support the candidates selected by their favored political party (Atchison, 2021).
Since brand advocates usually praise the brand, including all related feelings and beliefs (Kotler et al., 2019), political parties could select better candidates to advocate the brand's needs.While parties seek to promote candidates with long-term survival tenure and effect on public policy (Rehmert, 2020), candidates' attributes were proved to have a minor effect on preferences influenced by the leaders and national policies (Mulgan & Aimer, 2004).
Individual choice and preference reflect the attributes and values that are less impactful.Thus, predicting preference success relies on explicitly evaluating attributes importance to individuals.Some beneficial and physical product attributes could significantly influence ones preference judgments (Ebrahim et al., 2016;Singh et al., 2005).Moreover, in analytics, when preference is significant, additional features are retained, whereas fewer attributes are recalled when the preference is insignificant (Shi, 2022).Thus, if viewed positively, celebrity candidates' attributes could influence voters' perceptions (Street, 2019) and, ultimately, their preferences.
The attitudes and feelings that affect the bond between a voter and a candidate are a component of purchasing behavior, which is more of a set of particular mental tactics than a specific psychological activity (Cwalina & Falkowski, 2015).The affective perception is the foundation of brand affinity, an attraction-related feeling, and the desire to develop and maintain partnerships (Oberecker & Diamantopoulos, 2011).In politics, affinity is described as a favorable opinion or a positive affection for political candidates that can influence the voter-candidate relationship (Dolan, 2008).
Individuals are more likely to adopt favorable affective evaluations of things or people when good associations support them with other things or brands (Wänke et al., 1998).Individuals link certain emotions, feelings, and attributes with brands to distinguish them from competitive brands (Baines et al., 2019).They interact with multiple brands and experience feelings of love, affection, and belonging.However, the values of celebrities in politics are viewed differently (Mendick et al., 2018), and voters' attitudes can affect their attributes' perceptions and affinity toward CPs, influencing their party preferences.
Rational choice considers that people vote by self-interest and cost-benefit analysis, just like selecting a good or service.An individual's ability determines the interaction of emotion with reason, in which emotion is considered an attribute of motivation and a component of decision-making within the rational choice paradigm.Emotions are thought to be linked to motivation and cognition, with enthusiasm or apathy, hate, anger, or fear, determining the level of motivation and influencing electoral choice.The cognitive map of decisions can result from solid emotions, stimulating cognition and learning.
Although the balance between emotion and rea-son differs, it is interesting to note how this balance affects preferences (Dean & Croft, 2009).
The gender gap literature promotes significant differences in opinion within diverse fields, acknowledging its ability to shape a brand's choice and the public's understanding of politics (Sperling, 2014).According to Gidengil et al. (2003), gender differences exist in policy preferences and political beliefs.Women are expected to develop higher emotions than men (Kite & Whitley, 2016).Feminist philosophers assert that men are linked to reason while women are to feelings (Pavco-Giaccia et al., 2019).However, regardless of women's emotional suggestions, both genders employ rational thinking levels, particularly concerning unfavorable action rationality outcomes (Kerr, 2021).
The literature review showed that ACP, PACP, BA, and gender are important factors of influence on PBP.
Although ACP and PBP are recognized as having a two-way relationship, it is suggested that PACP and BA have significant mediating roles that may explain ACP's influence on voters' PBP.Additionally, the literature evidenced that women diverge from men regarding emotional feelings and rational reasoning.
Therefore, this study aims to inspect the impact of ACP on PBP, including the PACP, BA, and gender effects in a disturbed, corrupted, and rebelled context.The impacts can be seen directly, indirectly, and conditionally.
Based on the literature review, the study suggests five hypotheses, including direct, indirect, and conditional effects.Figure 1 displays the proposed research model: H1: Lebanese voters' attitudes toward celebrity politicians positively affect their political brand preference.
H2: Voters' attitudes toward celebrity politicians positively affect voters' political brand preference by the mediation of brand affinity.
H3: Voters' attitudes toward celebrity politicians positively affect voters' political brand preference by the mediation of perceived attributes of celebrity politicians.
H4: Gender moderates the relationship between attitudes toward celebrity politicians and political brand preference so that the mediating effect of brand affinity would be more significant for females.
H5: Gender moderates the relationship between attitudes toward celebrity politicians and political brand preference so that the mediating effect of perceived attributes of celebrity politicians would be more significant for males.

RESULTS
The extracted communalities, the significance of the Bartlett test of sphericity, and the Kaiser-Meyer-Olkin measured sampling adequacy and item loadings.According to Gaskin (2020), all constructs' skewness and kurtosis evidenced the normal data distribution as per thresholds of -/+1 and -/+3, respectively.The Variance Inflation Factor examined the multicollinearity assumption with the highest value of 1.007, a value less than the 3.3 threshold.Seven items from PACP and six items from ACP were removed from the analysis.As a result, in the Exploratory Factor Analysis (EFA), the EFA convergent validity was evidenced by items loadings exceeding the value of 0.50, with indicators average value above 0.60, and the EFA discriminant validity by the absence of cross-loadings of item indicators.All constructs inter-correlation had a value inferior to 0.85, and Cronbach's alpha exceeded the value of 0.50, ensuring the EFA reliability.All coefficients exceeded the 0.50 threshold with a value of 0.856 for PACP, 0.725 for BA, 0.916 for PBP, and 0.820 for ACP, evidencing the variables' reliability (Gaskin, 2020).
This  The displayed results in Table 3 showed that the hypothesized positive direct effect between ACP and PBP was statistically negative and non-significant at the conventional 0.05 level with a standardized regression weight, β = -0.016,and a probability-value = 0.571 > 0.05, rejecting H1.
The indirect mediating effects were tested using the suggested bootstrapping method with a 95% confidence interval and 2000 re-samples.The outlining percentiles lower and upper bound bootstrap mediating results are significant if not including a zero value at a confidence interval of 95% or 90% (Gaskin, 2020).The hypothesized positive indirect mediating effect displayed in Table 4 shows that the impact of BA on the ACP and PBP relationship was statistically significant at the conventional 0.1 (90%) level (β = 0.011, p-value = 0.053 > 0.05 and < 0.10).In addition, the mediating effect via PACP was positively significant on the conventional 0.05 (95%) level (β = 0.009, p-value = 0.039 < 0.05).All lower and upper levels had no zero value between them, thus, supporting H2 on the 90% and H3 on the 95% confidence level.
Using the moderated mediation plugin (Gaskin, 2020), the gender moderating impact on the ACP and PBP mediation via BA was tested.Both groups tested and compared the gender multi-group moderating effect on the ACP relationship with PBP.For females, the parameter estimates evidenced a positive significant mediating effect of BA between ACP and PBP relationship (β = 0.027, p-value = 0.006).At the same time, it failed to evidence a mediation significance for males (β = 0.000, p-value = 0.929).Using the moderated mediation plugin (Gaskin, 2020), the gender moderating impact on the ACP and PBP mediation via BA was tested.The findings confirmed the significance of the hypothesized moderated mediation effect of gender (lower and upper values did not contain a zero, p-value = 0.025).Furthermore, it has yielded better-standardized estimates for the female group than the males.Thus, H4 is supported.Moreover, the mediation via PACP failed to yield significance for females (β = 0.012, p-value = 0.108) and males (β = 0.007, p-value = 0.130).
It indicated a lack of gender mediation effect on PACP, thus, rejecting H5.
The effect size (f²) of the single-group dependent variable PBP was checked.It accounted for a value of 0.072, with no impact of ACP, and a small impact of PACP and BA, with values of 0.0291 and 0.0453, respectively.In examining the moderating impact of gender's multi-group analysis, the effect size of the male group accounted for a small effect value of 0.068, in which ACP had zero impact and PACP and BA had minor effects of 0.0258 and 0.044, respectively.Moreover, the female multi-group effect size accounted for an approximate ACP null value of 0.0022 and minor effects of PACP and BA with values of 0.0325 and 0.0444, respectively.Thus, they showed weak effects as per Aiken et al.'s (1991) thresholds in examining each variable impact on the dependent variable.
One theorized direct path was found non-significant across the single-group and the multi-group analyses.Consequently, the model's strength is detected by performing a post hoc power analysis per Daniel Soper's valuation.The PBP construct with three predictors, an observed coefficient of determination R² of 0.07, and a sample size of 1269 participants tested on a 95% confidence level yielded an observed statistical power value of 1, more significant than the 0.8 thresholds.It indicates a 100% chance of finding if a significant path between ACP and PBP exists, concluding that the study's model has enough power to confirm all paths' significance.

DISCUSSION
Research has been continuously revising antecedents of voters' PBP.This study aimed to discover patterns of voters' reasoning, emotions, and perceptions regarding CPs when nominated by parties.In addition, the study addresses a political strategy in a compet-  Since Lebanese voters were constrained to one political party or candidate, the insignificant relationship highlights the lack of motivation to support the party's nominated celebrities.However, even if voters changed their PBP, they would shift to an alternative party due to the Lebanese electoral law, but the relationship could remain insignificant.This finding highlights the citizens' high dissatisfaction with the party's history.The weak negative relationship indicates that voters would be able to cope with specific celebrity candidates through certain factors that might make the relationship significant.This result has been unexpected yet quite reasonable given the disturbing context of the study.
Political candidates' elections are tied to political party preferences, influencing voters' decisions regarding CPs.However, little was known about the role of ACP, PACP, BA, and PBP.Accordingly, the study findings indicated that BA and PACP fully and positively mediate ACP and PBP linkage.Furthermore, voters were found to support multi-attribute competent and liked celebrity candidates that can improve their favorable attitudes.Congruent with research on cognition, affect, and conation, this study promotes the ACP psychological path that directs feelings, perceptions, and preferences (Kotler et al., 2019).
BA construct has limited studies within the political marketing context.The current study findings demonstrate that BA significantly mediates the relationship between ACP and PBP.A positive mediating effect of BA indicated higher affinity levels, resulting in higher PBP.This result is congruent with the literature suggesting that people develop affective evaluations and emotional ties to specific entities, thus, impacting their preferences (Rambocas et al., 2014).ACP failed to directly influence PBP, denoting BA's importance in fully mediating this relationship.
PACP evaluation differs between voters' level of attribute importance, which was found significant in this study.Research shows, in congruence with the literature, that perceptions of CPs' attributes, if positively viewed and assessed, can ultimately affect their PBP (Rehmert, 2020).If supported, the increased favorability of beneficial outcomes enhances voters' PBP.This outcome shows voters' rationale in assessing candidates before their brand preference decisions.This finding is congruent with voters' preferred profession type of celebrity intellectuals and business candidates, which evidences their rational judgments.
The multi-group model showed differences between the gender groups and the BA and PACP mediating roles that have lost real significance in the PACP role and the male's moderation significance in the BA mediating role.Gender differences were confirmed in the female group, evidencing the feelings association in more significant affinity mediation (Kite & Whitley, 2016).The findings did not conform to the PACP linkage to reason and rationality, affirming the gender types' moderation in shaping brand choice.This study's findings evidence that females develop higher feeling associations in brand preference evaluations.This result is in line with the gender emotion differences in asserting that males and females behave in dissimilar predicted ways based on the situation and their social identities.The positive effect of rational choice theory is in line with Kerr's (2021) suggestions that people make rational decisions based on preferences.The significant mediating effect of PACP strengthened this theory.
In addition, the dimension of gender differences has proved the rational thinking of both males and females, highlighting differences only in the females' BA emotional liking.These findings contribute to the literature on gender differences in voters' political candidates' perceptions, specifically the CPs.
The discussion on positively influencing PBP in nominating celebrities as parliament candidates in a corrupted-sectarian-multi-party context needs to be more consistent, opening paths for additional research.Given the results of this study, the trend of celebrities' interest in politics makes these findings relevant for celebrities aiming for politics, political parties aiming for celebrities' and voters' support, and political marketers considering influence factors.It is recommended that political marketers select favorable, competent, honest, and liked celebrities to increase voters' positive PBP development.
Although the study findings revealed a lack of favorable attitudes toward political parties' preferences, the result highlighted voters' stress in choosing promising candidates for better living, indicating voters' willingness to support liked and competent celebrity candidates.
Celebrities would be transparent with the public about their education, experience, background, financial status, willingness, and ability to reform when aiming to join politics.Moreover, celebrities are encouraged to present achievable reform plans aligned with the political party through marketing campaigns, news releases, and social media to engage voters better.Finally, political parties would consider nominating liked celebrity politicians in districts with more significant female presence to maximize party preferences.

CONCLUSION
This study evaluates the effect of political parties nominating celebrity politicians on voters' preferences in Lebanon after the national uprising.Results showed that celebrities as political candidates could significantly influence voters' acceptance in the Lebanese parliament elections.As such, political parties have a critical role in nominating appropriate celebrities.However, certain parties might select celebrities and support them behind the scenes in a so-called independent party.As such, they will covertly sustain their official power.Future studies might address such conditions.
The findings of this study revealed that in crisis, attitudes toward celebrity politicians could positively affect party preferences with liked celebrities who possess beneficial attributes.In addition, findings on gender impacts have challenged the literature in failing to account for any moderation role regarding the perceived celebrity politicians' attributes rationale, evidencing females' moderation role in the affective evaluations.Therefore, an exhaustive understanding of these political brand preference predictors would assist in effectively targeting Lebanese voters and influencing their party preferences.

Table 1 .
Correlation analysis study addressed the Common Method Bias (CMB) by statistical methods following Gaskin (2020).First, Harman's single factor test evidenced the absence of bias in data with 15.181% explained variance, a value inferior to the 50% threshold.

Table 5 .
Gender moderated mediation effects

Table 6 .
Gender regression weights