A six-factor model of brand personality and its predictive validity

The study examines applicability and usefulness of HEXACO-based model in the description of brand personality. Following contemporary theoretical developments in human personality research, Study 1 explored the latent personality structure of 120 brands using descriptors of six personality traits as defined in HEXACO model: Honesty-Humility, Emotionality, Extraversion, Agreeableness, Conscientiousness, and Openness. The results of exploratory factor analyses have supported HEXACO personality six-factor structure to a large extent. In Study 2 we addressed the question of predictive validity of HEXACO-based brand personality. Brand personality traits, but predominantly Honesty-Humility, accounted for substantial amount of variance in prediction of important aspects of consumer-brand relationship: attitude toward brand, perceived quality of a brand, and brand loyalty. The implications of applying HEXACO-based brand personality in marketing research are discussed.

• HEXACO personality model has been widely validated in human personality studies but it was never applied to brand personality research • The study explored construct and predictive validity of HEXACO model on a set of 120 diverse brands • The latent structure obtained on brands ratings to a large extent corresponded to HEXACO dimensions • The Honesty-Humility dimension have shown to be a significant predictor of relevant aspects of consumer-brand relationship.
Brand personality represents an important aspect of brand image (Kapferer, 2008;Keller, 2008).It is widely believed that brands that built up and established positive personality traits in consumers' perception have evident benefits of such an image (Freling & Forbes, 2005).There are at least two main streams in defining brand personality.The first one defines it as a set of human characteristics associated with a brand (Aaker, 1997), while the second one gives a narrower and therefore more precise definition by which brand personality is the unique set of human personality traits both applicable and relevant to brands (Azoulay & Kapferer, 2003).Nevertheless, authors agree that the personality traits associated with a brand, similar to those associated with an individual, tend to be relatively enduring and distinct (Aaker, 1997).The explanation of this phenomenon can be found in people's natural tendency to anthropomorphize non-human objects and to explain them in terms of their own experiences and conceptions (Moynihan, 1997).
Aaker's approach Aaker (1997) defines brand personality as a set of human characteristics associated with a brand.This broad definition implies that although there is a relationship between human and brand personality, some dimensions of human personality may not be mirrored in brands (Aaker, 1997).Therefore, she developed a scale based on Big five personality inventories, as well as on the scales most often used by marketing or advertising practitioners.Also, she included socio-demographic and physical characteristics, inner values and reflection of the typical consumer (e.g.young, good looking, family-oriented, etc.) (Azoulay & Kapferer, 2003).Aaker's (1997) seminal article identified five distinct brand personality dimensions: Sincerity, Excitement, Competence, Sophistication and Ruggedness.Studies that followed successfully used Aaker's scale and more or less confirmed its factor structure (Aaker, 1997, Aaker, 1999;Aaker et al., 2001;Kim et al., 2001).Research based on this approach replicated three brand personality factors that reflect Big five human personality traits.Namely, Sincerity consisted of both Agreeableness and Conscientiousness from the Big five model, Excitement encapsulated notions of sociability, energy and activity and highly resembled Extraversion, while Competence represented an amalgam of Conscientiousness and Extraversion.On the other hand, Sophistication and Ruggedness differed from any dimension of Big Five model (Aaker, 1997(Aaker, , 2000;;Geuens et al., 2009).

Exploring brand personality within Big five framework
Big Five model is a result of decades of personality research within lexical paradigm framework.The key assumption of the lexical approach is that the most socially relevant and salient personality characteristics are encoded in natural language, and thus it would be possible to derive basic personality dimensions simply by sampling language (Allport, 1937;Goldberg, 1993).After decades of research, personality psychologists had reached a consensus that the trait domain could be best described, at its broadest level of abstraction, by five personality dimensions: Extraversion, Agreeableness, Conscientiousness, Neuroticism and Openness (Costa & McCrae, 1992;Goldberg, 1993).
Contrarily to Aaker's approach and attempts to develop a scale for measuring brand personality, some authors use the descriptors of human personality to describe brand personality (Geuens et al, 2009).This approach differs from Aaker's in several ways.Firstly, it defines brand personality in a narrower way restricting it to human personality traits, in that way ensuring clearer measure of personality of a brand.Therefore, in this approach brand personality is measured with the same scales/descriptors that are used for measuring human personality.Consequently, the factors identified in brand personality studies are expected to be equivalent to five dimensions extracted from standard personality questionnaires, and therefore comparable to human personality structure.
However, only a few studies managed to replicate the factor structure identified in human personality when examining brand personality.Namely, one study extracted two factors, first representing the composite of Agreeableness and Extraversion, and the second which consisted of markers of both Extraversion and Openness (Caprara et al., 2001).More successful in replicating five-factor structure were Bosnjak et al (2007).They identified four dimensions, i.e., Drive that corresponded to Extraversion, Conscientiousness that highly resembled the same factor from the Big five model, Emotion that corresponded to Emotional stability and Superficiality that represented Agreeableness.To the authors' best knowledge, only two studies managed to replicate the Big Five structure on brands (Geuens et al., 2009;Milas & Mlačić, 2007).

Six-factor model of personality
Recently, some authors started questioning the idea that five broad dimensions can capture the domain of personality variation to the full extent (Ashton & Lee, 2001;Ashton et al., 2004).Alternative model, named HEXACO, proposed the existence of six high order dimensions of personality (Ashton & Lee, 2001;Ashton et al., 2004;Lee & Ashton, 2004;Ashton & Lee, 2007).HEXACO is based on the evidence that a similar set of six personality dimensions has been found in lexical studies of personality structure in diverse set of languages (Ashton, Lee, & Goldberg, 2004;Ashton et al., 2004;Ashton & Lee, 2007).This model contains six broad personality dimensions: Honesty-Humility, Emotionality, Extraversion, Agreeableness, Conscientiousness and Openness to experience.Three of the HEXACO factor scales, namely, Extraversion, Conscientiousness and Openness to experience are similar in content to the corresponding dimensions of the Big five model (Lee & Ashton, 2007).However, dimension Honesty-Humility does not have direct counterpart in the Big Five/Five-factor model.This dimension reflects tendency to avoid/ use manipulation of others for personal gains, temptations to follow/break rules, importance of material values and overall deceitfulness (Lee & Ashton, 2004).HEXACO Agreeableness includes traits such as gentleness, flexibility, and patience versus anger, hostility, and aggression, whereas Emotionality is marked by such lower-level traits as sentimentality, dependence and anxiety versus bravery and toughness, and are to a lesser extent similar to Agreeableness and Neuroticism from Big Five model (Lee & Ashton, 2007).For a detailed theoretical background of HEXACO model see Ashton and Lee (2001).

The aim of the present study
Current study aims to explore the applicability of six-factor personality model in brand personality research.Based on the available empirical evidence, several arguments favoring this model can be made.On the one hand, Milas and Mlačić (2007), and Geuens et al. (2009) successfully demonstrated that Big five personality dimensions, which closely correspond to the five dimensions of HEXACO model, can be used to describe brands.On the other hand, the "additional" HEXACO dimension of Honesty-Humility is not a new concept in brand personality research, although it was never recognized as such.Namely, Aaker's Sincerity dimension (down-to-earth, honest, wholesome, and cheerful) to some extent resembles Honesty-Humility trait.Additionally, Raffelt, Schmitt, and Meyer (2013), by applying Aaker's methodology, extracted the factor that reflected honesty and sincerity, which they named Naturalness.Moreover, in marketing literature honesty and reliability of a brand is often considered as an essential aspect of brand trust (Agustin & Singh, 2005;Bruhn, Schnebelen, & Schäfer, 2014;Chaudhuri & Holbrook, 2001;Lee & Jeong, 2014, Wang & Emurian, 2005).Finaly, regardless of the theoretical framework and operationalization used, previous scientific studies seldom aimed to establish predictive value of obtained brand personality traits.Thereupon, the purpose and the contribution of this article is to address two interrelated questions: (1) whether the brand personality can be described in terms of traits defined within the HEXACO model, and if so, (2) to provide evidence of predictive validity of such brand personality traits.

Study 1: Brand personality structure
In order to address the first question, we have examined brand personality structure of highly familiar brands using a set of adjective descriptors for each of the HEXACO dimensions.

Method
In order to ensure that all brands were highly familiar to most people, in the first phase of the research a sample of 56 participants (27 male and 29 female) rated familiarity of 166 brands.For this initial pool of brands authors selected diverse set of brands from a variety of different categories (fast-moving consumer goods, media, automobiles, electronics, furniture, finance, etc.) of both international and well-established Serbian brands.Participants' task was to rate their familiarity with each of the brands using a five-point Likert-type scale (from 1-complitely unfamiliar, to 5-complitely familiar with).For the final set we have selected 120 well-established brands which were rated as highly familiar.The list of these brands is given in the Appendix.
Following HEXACO six-factor framework (Ashton & Lee, 2001;Ashton et al, 2004;Lee & Ashton, 2004;Ashton & Lee, 2007), we designed economic instrument that contains 24 markers (4 per dimension) of both pole of six personality dimensions.For Extraversion (sociable, withdrawn, passive, cheerful), Emotionality (tense, calm, anxious, relaxed), Openness (creative, traditional, unimaginative, curious), Agreeableness (inconsiderate, harsh, hardhearted, well-intentioned), and Conscientiousness (responsible, systematic, lazy, irresponsible) we have chosen proven markers of five factors from a more extensive list of attributes that was used for personality assessment by Lazarević and Knežević (2011) and Lazarević (2012).We have tried to avoid descriptors that are likely to lie in the interfactor space in order to get clearer factor structure.Moreover we made an effort to bypass conceptual differences between broad dimensions of Big Five and HEXACO model by selecting descriptors of common lower level traits of these models.Furthermore, based on the literature review we added four descriptors which aimed to reflect the essence of Honesty-Humility dimension (sincere, insidious, fair, generous).The decision to use adjectives instead of statements was made because they have broader applicability, thus are more suitable for capturing core brand attributes.
An independent sample of 120 participants (27 male and 93 female), ages ranging from 19 to 30 (M = 22.82, SD = 2.42) rated perceived personality of 120 brands on each of 24 adjectives, using 5-point Likert-type scale.Each brand was rated by ten different subjects.Brands were presented in random order along with their distinct logos.The participants were given the instruction to think of a brand as if it was a person, and to rate that person's characteristics.As we were interested in prevalent brand perceptions and qualities of a brand created by marketing strategies, not individual differences between participants' perceptions of given brands, and in order to equalize subjective perception of the brands, eliminate idiosyncratic ratings, and to enhance the reliability of obtained measures, ratings were aggregated, i.e. averaged across subjects, thus ensuring that only between-brand variance is analyzed.
For the analyses, the data were organized so that brands emulated subjects, adjectives for each personality dimension were presented in columns, and cells contained aggregated ratings.

Results
To explore the brand personality structure, we applied exploratory factor analysis (EFA) using Maximum likelihood extraction method along with Promax rotation.The Kaiser-Meyer-Olkin (KMO) sampling-adequacy measure (.89) and Bartlett's test of sphericity (χ 2 (276) = 3010.542,p <.001) were applied to the data prior to factor extraction to ensure that the characteristics of the data set were suitable for factor analysis.The scree plot demarcated a slope change after the sixth factor.In line with theoretical expectations the number of factors in the analysis was fixed to six.A six-factor solution (table 1), which accounted for 79.35% of variance, closely corresponds to dimensions defined within HEXACO model 3 : Agreeableness (5 items), Consciousness (4 items), Openness (4 items), Honesty-Humility (3 items), Emotionality (5 items), and Extraversion (2 items).Overall, pattern matrix showed few or no cross-loadings.All items showed primary loadings on the expected factors, except from adjective insidious, which was initially an indicator of Honesty-Humility, but had extremely high loading on the first factor, and cheerful, which was originally a marker of positive affectivity within Extraversion, but loaded higher on Emotionality factor.On the other hand, adjective sociable had equally distributed marginal loadings, and thus could not be considered as a distinct marker of any single trait in particular.
PSIHOLOGIJA, 2017, Vol.50(2), 141-155 Descriptive statistics for six brand personality traits are given in table 2. The average correlation between factors was moderate (r = .334),yet some of the extracted factors have shown unexpectedly high between-factor correlations, e.g.Honesty, and Agreeableness, Agreeableness and Emotionality, Openness and Extraversion (table 3).

Discussion
The exploration of brand personality structure has shown that majority of adjectives tend to group around latent dimensions which correspond to the theoretical expectations of HEXACO model.Thereupon, Conscientious brand is the one that is perceived as highly responsible, systematic, and hard-working.Creative and inquisitive brands are perceived as open to new experiences and novel things, while the conventional, traditional and unimaginative brands are considered as conservative.The brands that project tension, anxiety, fearfulness and nervousness are viewed as Emotional, i.e. emotionally unstable.
In terms of brand personality, Extraversion is perceived as slightly narrower trait than in humans, at least in this study.In other words, Extraversion brings together descriptors that reflect energy level and activity of a brand (passive and withdrawn), but not necessarily its social aspect (sociable/ outgoing).Agreeableness seems to be the most prominent brand personality trait.Brand agreeableness subsumes characteristics such as being inconsiderate, and insidious.Furthermore, brands that are perceived as not being well intentioned tend to be perceived as aggressive and harsh.Finally, Honesty-Humility trait subsumes brand's sincerity, fairness, and generousness.
It is important to note that the inter-correlations between factors are higher than those typically observed in human personality studies (Costa & McCrae, 1992, John & Srivastava, 1999;Lee & Ashton, 2016).This indicates that people tend to perceive brand personality more holistically, rather than through distinct set of independent traits.Other explanation of such correlations might lie in the nature of the object of assessment, i.e. due to the lack of information about the brand, in contrast to people, one has a natural tendency of giving more stereotypical and simplified descriptions.Yet the direction of obtained correlations is in line with theoretical expectations and empirical evidence, mimicking relations between dimensions of personality frequently presented in human personality assessment (Costa & McCrae, 1992, John & Srivastava, 1999;Lee & Ashton, 2016).

Study 2: Predictive validity of brand personality
In the second study we examined the predictive validity of HEXACObased brand personality.As criterion variables we have chosen an important aspects of consumer-brand relationship (Aaker, 1991;Fournier, 1998;Loureiro, 2012): (1) attitude toward brand -predisposition to respond in a favorable or unfavorable manner to a particular brand (2) perceived quality of a brand -the customer's perception of the overall quality or superiority of a product or service with respect to its intended purpose, relative to competition brands, and (3) loyalty to brand -tendency of buying/using products of a certain brand over time rather than switching between competition brands.In that way we wanted to grasp relatively wide consumer-brand relationship, and to test and evaluate predictive value of obtained measures of brand personality traits.

Method
For each brand, we have calculated six brand personality scores based on the results obtained in factor analysis 4 .In order to evaluate validity of brand personality, a different sample of participants reported their attitude toward brands, perceived quality of brands and rated their loyalty to brands.For attitude toward brand, 55 participants (21 male and 34 female) were presented with the same set of 120 brands, for which brand personality assessments were previously collected.Participants' task was to rate each brand on a scale from 1 (extremely negative) to 5 (extremely positive), on the bases of their overall brand perception.In order to estimate prevalent attitude toward each brand, ratings were averaged across all participants.
For the assessment of perceived quality we have recruited a new sample of 30 participants for each brand.A selection criterion was personal experience with a particular brand, i.e. a positive answer to the following question: Have you ever used ______ [name of 4 Values on each dimension were recoded so that higher values are congruent with dimensions' labels.
PSIHOLOGIJA, 2017, Vol.50(2), 141-155 a brand]? in order to ensure ratings based on personal experience with the brand.Participants rated statements 5 about perceived quality of a brand, on a scale from 1 (absolutely disagree) to 5 (completely agree).In order to get prevalent and more reliable measure of perceived quality of a brand, ratings were averaged across participants and statements.Loyalty measures were collected in a similar manner.Thirty participants that indicated having personal experience with the brand, rated statements about brand loyalty 6 .Again, all ratings were averaged across both statements and participants, in order to obtain a single loyalty measure for each brand.As some brands are less accessible than others (e.g.Rolex), complete data for perceived quality and loyalty were obtained for 72 most commonly used brands.
In order to assess validity of six-factor brand personality, we conducted several regression analysis -each time brand personality traits were entered as predictors, while attitude toward brand, perceived quality and loyalty were set as criterion variables.

Results
Brand personality accounted for nearly 60% of the variance of attitude toward brand (R =.758, F (6,113) = 25.483,p <.01).Although, attitude toward brand was related to five out of six traits, positive attitude toward brand was exclusively related to high Honesty, Openness, as well as low Emotionality of a brand (table 4).Furthermore, brand personality traits have shown to predict percived quality of a brand (R = .715,F (6,65) = 11.315,p <.01), yet only Honesty and Openness emerged as significant predictors.Finally, brand personality traits have shown to succesfuly predict brand loyalty (R =.607, F (6,65) = 6.310, p <.01).Even though almost all personality traits have shown significant zero-order correlations with loyalty, the only significant independent predictor was the trait of Honesty-Humility. 5 1) The quality of this brand is high; 2) The quality of this brand is higher than the quality of similar brands on the market; 3) It is unlikely that this brand will fail customers' expectations regarding quality.6 1) When I use this kind of service/product, I always choose this brand; 2) This brand will be my first choice, if I use this service/product; 3) If this brand is available, I would not use/buy any other.

Discussion
In study 2 we demonstrated that HEXACO-based brand personality has not only satisfactory latent structure, but also predicts important aspects of consumers' relationship with a brand.As seen from zero-order correlations almost all brand personality traits reflect in attitude toward brand, perceived quality of a brand and loyalty to brand.Attitude toward brand has shown to be most closely linked to brand personality.People form a positive attitude toward brands that seem honest, open to new experiences, and emotionally stable.This profile fits the concept of "positive personality", since it combines socially desirable and valued personal characteristics which are perceived to be projected by the brand.On the other hand, although perceived quality and loyalty are accompanied by the perception of a brand being less emotional, more open to experience, conscientious, and agreeable, these traits have little or no predictive power in presented models.When it comes to perceived quality and loyalty the most important brand personality trait that plays a substantial role is Honesty-Humility.It seems that all positive attributes of a brand that are important for ones perception of quality and loyalty to brand are summarized in the perception of brand's honesty and sincerity, thus brand which put the emphasis on these attributes via different communication strategies may benefit from more positive brand perception.

General discussion and implications
This study examined applicability and usefulness of six-factor model in brand personality research.To the best of authors' knowledge this is the first attempt to implement HEXACO framework to marketing research of brand perception.The main rationale for such an attempt stems from the notion that HEXACO is relatively novel but already is well established model in human personality research, and therefore provides extensive conceptual and theoretical framework for description of personality (Ashton & Lee, 2001;Ashton et al, 2004;Lee & Ashton, 2004;Ashton & Lee, 2007).Moreover, in terms of its predictive value for a variety of psychological criteria this model demonstrates several important advantages over the Big five framework (Ashton & Lee, 2007), and in our opinion can be considered as the most comprehensive taxonomy of relevant personality traits to date.
The results of the first study demonstrated that when attributes, which are usually used for the description of human personality are applied to brands, they group to form expected higher order personality dimensions, thus, suggesting that brand personality can be described in terms of personality traits proposed by the HEXACO model.The second study demonstrated that, six-factor brand personality, and especially Honesty-Humility dimension, has predictive power for some of the important aspects of the consumer-brand relationship.
Some of the previous studies demonstrated that Big five model is applicable to brand (Bosnjak et al., 2007;Geuens et al., 2009;Milas & Mlacic, 2007).We believe that HEXACO model of brand personality should not be perceived as essentially different from the Big five, but can be considered as its extension.The importance and incremental value of Honesty-Humility dimension is well demonstrated in human personality literature (Ashton & Lee, 2001;Ashton et al, 2004;Lee & Ashton, 2004;Ashton & Lee, 2007).The marketing research literature also recognizes the importance of Honesty-Humility dimension, on the one side as an aspect of brand personality (see Aaker, 1997, Aaker, 1999;Aaker et al., 2001;Raffelt et al, 2013), and on the other as an independent factor that influences consumers' brand-related behavior.Namely, Aakers' approach highlights the Sincerity dimension, which is important factor of mutual understanding between consumer and the brand (Aaker, Fournier, & Brasel, 2004).
On the other hand, honesty is often considered to be a part of brand trust, which is defined as a consumer's confident beliefs that s/he can rely on the brand to deliver promised services or products (Agustin & Singh, 2005).A number of authors highlight honesty as an essential aspect of brand trust (Bruhn et al., 2014;Chaudhuri & Holbrook, 2001;Lee & Jeong, 2014, Wang & Emurian, 2005).Brand trust is the necessary catalyst for establishing and maintaining longterm brand-consumer relationship, and it facilitates consumer satisfaction with, and commitment to a brand (Chaudhuri & Holbrook, 2001).This can explain high prediction of Honesty-Humility dimension of brand loyalty, as well as its predictive power of perceived quality of brands' products.Furthermore, brand trust determines the affective relationship to brand (Chaudhuri & Holbrook, 2001), which leads to positive attitude toward brand.Finally, brand trust can be viewed as a function of perceived quality of a brand and brand loyalty whose antecedents probably lie in the perception of brand as honest.
Taken together, the present study indicates that the application of HEXACO model in brand personality research and practice can be useful, since it is more comprehensive than five-factor approach, and thus enables capturing important aspects of brand personality that are closely related to brand trust.

Limitations and further directions
Since this is a pilot study of the applicability of HEXACO model to the research of brand personality, it faces several limitations.First, the sample of brands is relatively restricted in terms of size.Nevertheless, the sample of brands is carefully selected to fit the demands of the research questions and is fairly broad regarding product categories used.For the first study we have chosen highly familiar brands of different categories and in the second the additional criterion was the personal experience with the brand.Further research needs to examine the replicability of these findings in extended set of brands, as well as within different brand categories.Second, limitation of the current study lies in the properties of the instruments used.Namely, we have chosen to use short scales, since they are more welcomed by participants and less time consuming than extensive ones.Although the instrument used in this research has shown satisfactory psychometric properties, and reproduced clear factor structure, it can be argued that the content of some personality factors is not fully covered by the limited set of descriptors.This is certainly true, but since this study did not aim to produce comprehensive instrument for measuring brand personality, but rather to test the applicability of HEXACO model to this line of research, the usage of narrow and well established markers of the traits can be considered appropriate.Finally, further research should seek to address the question of cross-cultural generalizability of the results obtained in current study, since previous studies have shown that brand personality structure, unlike human personality, can be culture dependent (Bosnjak et al., 2007;Geuens et al., 2009).

Table 4
Prediction of attitude toward brand, perceived quality of a brand, and brand loyalty