The Study on Brand Personality of Automobile Advertisements and the Communication of Brand Personality

(1)

Chiao Da Management Review Vol. 33 No.2, 2013 pp.139 -180

汽車廣告之品牌個性分析與品牌個性

傳播之研究

The Study on Brand Personality of Automobile

Advertisements and the Communication of Brand

Personality

蔡佳靜 1 Chia-Ching Tsai 國立雲林科技大學企管系

Depar恤lentofBusiness Administration, National Yunlin University ofScience and Technology

何佳慕 Chia-Chen Ho 義守大學企業管理研究所

Graduate School ofBusiness Administration, I-Shou University

摘要:本研究將 2006 至 2008 年十月台灣前三大汽車雜誌(一手車訊、超越車訊和汽車購買指南)的汽車平面廣告，以內容分析法與 Aaker (1997) 的品牌個性尺度為分類準則，探討汽車廣告業主採用何種品牌個性來塑造品牌。本研究以內容分析發現汽車業者主要以教養、刺激和勝任的品牌個性來塑造品牌。本研究亦以實驗法，針對教養、刺激和勝任的品牌個性，探討不同的品牌個性應搭配何種程度外表吸引力的模特兒進行傳播。研究結果發現品牌個性與模特兒外表吸引力對傳播效果皆有顯著的影響。且以廣告態度而言，不論品牌個性為高刺激或低刺激、高勝任或低勝任、高教養或低教養，均通合與高度外表吸引力的模特兒搭配。以品牌態度而言，當品牌個性為高刺激時，高度與過度外表吸引力的模特兒無顯著差異;當品牌個性為低刺激時，適度外表吸引力模特兒邊際顯著優於高度外表吸引力的模特兒，其餘不論品牌個性為高勝任或低勝任、高教養或低教養，高度與過度外表吸引力的

1 Corresponding author: Department of Business Administration, Nationa1 Yunlin University of Science and Techno1ogy, Yunlin, Taiwan, E-mai1: tsai胡叫a剖ic伽hc犯ch峙@yun川m叫t叫e郎叫叫ch.e叫.e吋du叫I

(2)

』一

140 The Stu砂 onBrand Personality 01 Automobile

Advertisements and the Communication 01 Brand Personality

模特兒的品牌態度皆無顯著差異。

關鍵詞:內容分析;品牌個性;外表吸引力;配過假設;傳播效果

Abstract : This study uses content ana1ysis to investigate the brand personalities of automobi1e advertisements according to brand personality sca1e (sinceri紗，

excitement, competence, sophistication, and ruggedness) of Aaker (1997). The advertisemerits' were selected 企omthe top three automobile magazines in Taiwan (i.e., Car News, Taiwan Motor, and Car Guide) 企om January 2006 to October 2008. This study found that the dimensions of sophisticαtion， excitement, and compe仿nce are more 企equent than sincerity and ruggedness in brand personalities. In addition, the study uses experimental design to investigate the match-up between the 1eve1s of models' physica1 attractiveness and that of sophistication, excitement, and competence used to communicate brand persona1ities. This study found that brand persona1ity and mode1s' physica1 attractiveness have a significant impact on the communication effect. For attitude toward advertisement, no matter whether brand persona1ities convey high or 10w excitement, high or 10w competence, or high or 10w sophistication, high1y at仕active mode1s are 1iked more than normally attractive ones are. For attitude toward brand, there is no significant difference between high1y and normally attractive models when the brand persona1ity conveys high excitement, whi1e normally attractive mode1s are marginally significant1y 1iked more than high1y attractive mode1s when the brand persona1ity conveys 10w excitement. For the remaining brand persona1ities, regard1ess of high or low competence or high or 10w sophistication, highly attractive mode1s are more liked than normally attracüve ones are.

Keywords: Content ana1ysis; Brand persona1ity; Physica1 attractiveness; Match-up hypothesis; Communication effect

1. Introduction

Branding is one of several marketing too1s used by companies and businesses. A unique brand is memorab1e for consumers and makes a company's

(3)

Chiao Da Management Review Vol. 33 No.2, 2013 141

products easily distinguishable and quickly recognizable in the market. The emphasis nowadays is for individuals to possess their personal styles. Similarly,

when a company launches a product or service

,

it must be packaged with a symbol that represents the company's image and style

,

leading to the growing attention towards the concept of branding. ln recent years

,

many researchers have even anthropomorphized brands, endowing them with individual personalities.

Among the various contributing factors

,

advertisements defmitely play an important role in the creation of brand personalities. Aaker (1997) considered advertisements to be a critical avenue through which brand personalities are established. Keller (1998) also highlighted advertisements as one of the most influential factors in the establishment ofbrand personalities.

Taiwan's automobile market became highly competitive upon its admission into the WTO in 2001. Up unti12005, the market remained fairly robust. However,

it gradually declined in 2006, with the volume of domestically produced cars falling by 28.8% year - on - year. In 2007

,

the automobile market slumped to its worst state in two decades. Reports by the lndustrial Development Bureau of the Ministry of Economic Affairs attributed the causes to market saturation as well as increases in intemational prices of oil and raw materials. Govemment intervention and rescue efforts in November 2008 not only failed to revive the ailing industry but also made matters worse. Despite the general malaise

,

automobile manufacturers spared no effort in promoting their brands of cars. A study of Taiwan's media market by Rainmaker XKM Intemational Corp.revealed that car

advertisements typically ranked in the top two in volume in all forms of media,

including television, magazines, and newspapers, indicating the industry's heavy reliance on advertisements. This study has investigated the use of brand personalities in advertisements seeking to increase vehicle sales during the downtum ofthe automobile industry between January 2006 and October 2008.

Wimmer and Dominick (2002) believed that content analysis could be used to study and determine the authenticity of media content. This method uses reality as the criterion to check media descriptions of specific groups, phenomena,

temperaments, or characteristics for consistency and accuracy. In this study,

(4)

142 The Stu砂 onBrand Personality 01 Automobile Advertisements and the Communication 01 Brand Personality

three automobile magazines and to classify different 可pes of brand personalities in order to gain a systematic understanding of the automobile industry's brand personalities in marketing and communication.

It is common practice for advertising agencies to appoint attractive-looking models as product spokespersons. McCracken (1989) emphasized the irnportance of selecting an appropriate spokesperson as he or she can directly shape the brand's personality. Baker and Churchill (1977) believed that highly attractive models (HAMs) have an overall advantage over normally attractive models

(NAMs) in creating positive attitudes towards the advertisements and products that they represent as well as the ability to instill in consumers the intention to purchase those products. This theory was supported by the fmdings of subsequent research

,

including studies by Caballero and Pride (1984) and Petorshins and Crocker (1989). Moreover

,

Kahle and Homer (1985) pointed out th剖 a good effect would result 企om having a good fit between HAMs and the advertised products. Solomon, Ashmore, and Longo (1992) used the match-up hypothesis to prove the importance of matching the spokesperson's level of attractiveness with the products in advertisements. Till and Busler (2000) and Koemig and Page (2002) highlighted that when the spokesperson 's image is congruent with that of the product

,

consumers would have positive assessments of both the product and the advertisement. Lee and Thorson (2008) also proved that celebrity endorsements were less persuasive when there was incongruence between the celebrity and the product or when the degree of fit varied. Overall

,

the communication effect of an advertisement is significantly influenced by the complementarity between the product and the spokesperson.

This study has discussed the brand personalities of car advertisements and what level of model attractiveness could match the different brand personalities of cars and generate good communication effects. The findings can be used as reference when brand personalities are being created in order to generate the desired communication effect. The next section of the paper is the literature review, which is followed by elaborations on the research assumptions,

(5)

Chiao Da Management Review Vol. 33 No.2, 2013 143

by discussing the management implications of the findings and providing recommendations for future studies.

2. Literature Review

The match-up hypothesis was used to investigate the best fit between different brand personalities and the level of attractiveness of spokespersons in order to achieve the communication effect (the main variable being studied). The paper then elaborates upon schema congruence and the match-up hypothesis, both of which are underlying theoretical bases for this study.

2.1 Communication Effect

The effects of an advertisement can be manifested in terms of sales effect and communication effect (Lavidge and Steiner, 1961). Sales effect is measured by an actual increase in sales volume of the advertised product. Communication effect is the degree to which the information conveyed by the advertisement has been noticed

,

understood

,

and accepted by consumers

,

in turn causing a change in their attitude and/or behavior. The metrics for both 可pes of effects have their respective pros and cons. ln general, the ultimate goal of an advertisement is to increase sales volume

,

so advertisement effects are mainly rated by sales effect. However, sales e旺ect is often affected by other factors, inc1uding competition and promotional activities

,

making it harder to measure compared to communication effect (Lin and Lin, 1994). Therefore, the main variable chosen for our study was communication effe仗， which is divided into the advertising and brand attitudes (Peng, 2000). Advertising attitude refers to the tendency of consumers to react positively or negatively to specific stimulation by an advertisement exhibited under specific circumstances (Lutz, 1985). Measurement of the attitude provides an understanding of consumers' evaluation of and response towards an advertisement. On the other hand, brand attitude refers to the sustained psychological tendencies that consumers have developed towards a specific brand (Lutz

,

Mackenzie

,

and Belch

,

1986).

(6)

144 The Stu咿 onBrand Personality

0/

Automobile Advertisements and the Communication

0/

Brand Personality

2.2

Schema Congruence and Match-up Hypothesis

2.2.1 Definition of Schema Congruence

Schema congruence occurs when the message in a stimulus is in line with the expectations of consumers' schema (intemal standards). Conversely, there is schema incongruence when that message interferes with the schema. Mandler

(1982) used the schema the。可 to explain how the degree of congruity between a message and consumers' schema affected the latter's emotional response and evaluation of the stimulus.

2.2.2 Effects of Schema Congruence

Mandler (1982) proposed that when consumers receive a stimulus, if it matched the cognitive structure of an existing schema, they would be able to quickly and easily categorize the message and then integrate it with the original schema to form a more complete schema, which would in turn produce a positive emotional assessment of the information: this is called

“

assimilation". However, if there was a divergence between the message and the existing schema, consumers would make adjustments to eliminate the inconsistency and resolve the difference. Consumers would usually have to expand their minds or modify their psychological structure in order to comprehend the incongruent message; this is called

“

accommodation. "

Mandler (1982) proposed modes that would resolve different messages by customers: consumers would process different received messages different旬，

leading to an evaluation of the information in the stimulus and various degrees of emotional intensity. Fiske and Taylor (1991) pointed out that humans form schema through the leaming process or past experiences and would have existing expectations of stimuli, messages, or events that would affect the way in which they deal with and judge a stimulus. Researchers applying the schema theory found that people could distinguish whether information was schema-congruent or not and tended to remember schema-related information while forgetting information that was unrelated. Under normal circumstances, established and good schema would lead to a preference memory for information in line with one's memories and expectations. Schema-congruent messages would be quickly

(7)

assimilated by consumers, leading to a rapid consumer evaluation and positive emotions (Hast間， 1981; Fiske and Taylor, 1991).

2.2.3 Definition of Match-up Hypothesis

The match-up hypothesis is that when a spokesperson's level of physical attractiveness matches that of the advertised product

,

the advertisement will be more effective (Kahle and Homer, 1985). Kamins (1990) explained that when the physical at甘activeness of a spokesperson is related to that of the product

,

it would create a positive attitude in consumers and that the process could be explained by assimilation under the schema theory. Assuming that consumers originally have an established schema of the product

,

assimilation would take place when the schema of the spokesperson matches that of the product, and consumers would create associations between the spokesperson, the advertisement, and the schema of the product

,

resulting in positive attitude.

2.2.4 Literature on Match-up Hypothesis

Kahle and Homer (1985) and Kamins (1990) studied celebrities' level of attractiveness and found that highly attractive celebrities were able to increase the advertising effect of the products they endorsed, and vice versa. The effects of the MUH can also be applied to non-celebrities. Parekh and Kanekar (1994) also verified that when a model whose level of physical attractiveness matches that of the product

,

the advertisement would be more effective. Solomon et al. (1992)

investigated the match-up effect between different types of aesthetics and products and found significant correlations between them. Bower and Landreth (2001) also found that for advertisements of beauty - enhancing products

,

HAMs exuded a greater degree of professionalism compared to NAMs and had better advertising effects.

The aforementioned studies focused on products, while Koeming and Page (2002) instead examined the application of the MUH to services. They found th的 when there is a match between a service and attractiveness, consumers fmd the service provider more physically attractive and become more willing to pay the price for the service. In summary, the MUH is indeed an appropriate explanation

(8)

146 The Stu砂 onBrand Personality 01 Automobile

for why consumers have a more positive assessment of a product or service and its advertisement when there is a match between the image of the spokesperson or service provider and that of the product or service.

2.3 Brand Personality

2.3.1 Definition of Brand Personality

Aaker (1997) defined a brand as a distinguishable name or symbol

,

for example

,

a logo

,

trademark

,

or packaging design

,

in order to distinguish the product or service of a seller (or a group of vendors) 台om that of its competitors

,

ensuring that consumers c1early know the source of the products and can recognize these products by their brand among similar products.

Brands

,

like humans

,

have their own personalities (Dubo缸~ 1986; Ca訂， 1996).

The majority of advertising agencies and brand managers also hold this view (Alt

and Griggs

,

1988). According to Aaker (1997)

,

brand personality refers to a set of human characteristics or 仕aits that can be attributed to a brand. It helps to create differentiation, glvmg consumers various substantial rights and benefits in different forms (Aaker

,

2000). Wee (2004) found th前 brandpersonalities function like personality 仕aits in humans, exhibiting stability over the long term. Yet,

human and brand personalities are not entirely alike in how they develop. Human personality is influenced by cultural and family backgrounds, social c1ass,

attitudes

,

beliefs

,

genetics

,

and body type. Brand personality is not inherent in the product or the brand itself: it is an additional a仿ibute that is shaped by the brand's marketing and promotional activities as well as by the reactions arising

企omconsumers' interactions with the brand. 2.3.2 Dimensions of Brand Personality

Alt and Griggs (1988) were the frrst to propose four dimensions of brand personality: ex仕oversion， social acceptability

,

virtue, and potency. Ba仕a，

Lehmann, and Singh (1993) came up with 14 dimensional pairs, inc1uding reliable-unreliable

,

technical-non-technical

,

interesting - non - interesting

,

etc.

(9)

Macrae (1996) divided brand personality into six types: ritualistic, symbol, heritage, exclusive

,

belonging, and legendary.

The brand personality scale (BPS) developed by Aaker (1997) is a combination of special features that can be used to measure and structure brand personalities (Table 1). Her study covered many product categories, while the five personality dimensions that she proposed - sinceri紗， excitement

,

competence

,

sophistication

,

and ruggedness - could explain up to 93% of the di宜erences

between all brands. Hence, the BPS is used as the criterion when constructing categories for content analysis.

Table 1

Brand Personality Scale (BPS) of Aaker (1997) Dimension szncenty excitement competence sophistication ruggedness Source: Aaker( 1997) Factors

Down-to-earth: down-to-earth, family-oriented, small-town

Honest: honest, sincere, real Wholesome: wholesome, original Cheerful: cheerful, sentimental，企iendly

Daring: daring, trendy, exciting Spirited: spirited, cool, young Imaginative: imaginative

,

unique

Up-to-date: up-to-date, independent, contemporary

Reliable: reliable, hard working, secure Intelligent: inte11igent, technical, co中orate Success扣1: successful, leader, confident

Upper: upper class, glamorous, good looking Charming: charming, feminine, smooth Outdoorsy: outdoorsy, masculine, westem Tough: tough, rugged

(10)

148 TheS，的dyon Brand Personality 01 Automobile Advertisements and the Communication 01 Brand Personality

2.4

Physical Attractiveness

2.4.1 Definition and Degrees of Physical Attractiveness

Patzer (1985) defined physical attractiveness as

“

the extent to which the target is pleasing to the eyes." Smith (1985) highlighted the significant correlation between facial and overall attractiveness

,

implying that the face is the most important element of physical attractiveness. Richins (1991) thought that a beautiful countenance constitutes a high degree of physical at仕activeness. As this study has focused on the attractiveness of spokespersons in print advertisements

,

Richins (1991) been adopted to de曲le physical attractiveness based on the models' countenance.

Martin and Kennedy (1993) and Bower and Landreth (2001) classified attractiveness into highly attractive and normally or moderately attractive. A

highlyat仕active person has a not only has beautiful countenance but also shapely physique (S甘iegel- Moore and Rodin, 1986; Bower and Landre血， 2001). On the other hand

,

the looks and physique of a normally attractive person are more general and akin to that of most people (Bower and Landreth, 2001). For the

pu中ose of this study, a highly attractive model is one who has a beautiful countenance

,

while a normally at仕active model has a less beautiful countenance than a high1y attractive model with significant differences between the two.

2.4.2 Stereotypes about Physical Attractiveness

Pendry and Macrae (1994) believed that people often classi勾r others based on their external appearances. Following the classification, one would have psychological expectations, which are then used to evaluate others (Hosoda,

Stone-Romero, and Coats, 2003). Many studies have revealed the tendency for humans to form certain stereo可pes based on physical attractiveness. For example

,

Dion

,

Berscheid

,

and Walster (1972) pointed out that it is generally thought that those who look physically attractive also tend to have successful careers. Cash and Gillen (1977) thought that highly attractive people were generally regarded as being more outgoing. Eagly et al. (1991) found that those who are better looking tend to be thought of as being more kind-hearted

,

highly at仕active people are

(11)

often regarded as possessing better social skills but a lower level of intellect and compassion

,

and at仕active-looking people give others the impression of being vain and conceited. Langlois et al. (2000) found th剖 highly attractive persons are deemed to have a more positive personality. Many studies also highlighted the consistency between physical appearance and the likelihood of being responsible or outgoing (Passini and Norman, 1966; Albright, Kenny and Malloy, 1988; Watson, 1989).

2.4.3 Match-up between Models' Physical Attractiveness and Brand Personality

Cash and Gillen (1977) and Langlois et al. (2000) reached the same conclusion that highly physically attractive people are generally regarded as being more outgoing than most people are. This impression of being outgoing corresponds to the characteristic of the excitement dimension

,

as proposed in BPS of Aaker (1997) (Table 1). Hence

,

a highly attractive model is more compatible with a high excitement brand personality. Conversely

,

a normally at仕active model is often seen to be less outgoing than a highly attractive model. According to the match-up hypothesis

,

to create better advertising effects

,

a high excitement brand personality should be matched with a highly attractive model; and a low

excitement brand personality, with a normally at甘active model. Based on the aforementioned arguments, the following hypotheses are made:

Hl-l: A better advertisement attitude will be generated when a high excitement

brand personality is matched with a highly attractive model or when a low

excitement brand personality is matched with a normally attractive model.

HI-2: A better brand attitude will be generated when a high excitement brand personality is matched with a highly attractive model or when a low excitement

brand personality is matched with a normally attractive model.

It was stated earlier that Dion et al. (1972) had found th剖 physically at仕active people tend to be viewed as having successful careers; hence

,

a highly attractive person possesses more characteristics of a successful person. This impression of being successful corresponds to the characteristic of the competence

(12)

150 The Stu吵 onBrand Personality of Automobile Advertisements and the Communication of Brand Personality attractive model is more compatible with a high competence brand personality.

Corresponding旬， a normally attractive model is often seen as having fewer characteristics of those who are successful compared to a highly at仕active model.

Based on the match-up hypothesis, a brand personality with low competence should be matched with a normally attractive model for a better advertising effect.

These viewpoints have led to the following hypotheses:

H2-1: A better advertisement attitude will be generated when a high competence brand personality is matched with a highly attractive model or when a low competence brand personality is matched with a normally at仕active model.

H2-2: A better brand attitude will be generated when a high competence brand personality is matched with a highly attractive model or when a low competence brand personality is matched with a normally attractive model.

Patzer (1985) viewed physical attractiveness as the extent to which the target is pleasing to the eyes, while Richins (1991) thought that a high degree of physical attractiveness means having a beautiful countenance. Hence, a person who is highly attractive is often better looking and has nice features. The impression of being good looking corresponds to the characteristic of the sophistication dimension, as proposed in BPS of Aaker (1997) (Table 1). Hence, a highly at仕active model is more compatible with a high sophistication brand personality. Moreover, a normally at仕active model is 0丘en viewed as less good looking compared to a highly attractive model. Again

,

using the match-up hypothesis

,

a brand personality with low sophistication should be matched with a normally attractive model for a better advertising effect. The hypotheses listed below were derived 企omthe aforementioned arguments:

H3-1: A better advertisement attitude will be generated when a high sophistication brand personality is matched with a highly attractive model or when a low sophistication brand personality is matched with a normally attractive model.

H3-2: A better brand attitude will be generated when a high sophistication brand personali可 is matched with a highly attractive model or when a low sophistication brand personality is matched with a normally attractive model.

Based on the literature review, a research 企amework for Study 1 - 3 of this research was developed

,

as shown in Figure 1.

(13)

Chiao Da Management Review Vol. 33 No.2, 2013

Figure 1

The Research Framework for Study 1-3

.-

'

-

'

-

'

-

'

-

'

-

'

-

'

-

'

-

'

-'-

'

-

'

-

'

-

'

-'-

--'

-'

_' Match-up Hypothesis Brand Personality Physical Attractiveness ' ,_._._._._._._._._._._.-._._._._._._. Communication Effect

3. Methodology

151

The current study was divided into two stages: content analysis was carried out during Stage 1

,

while experimental design was used for analysis in Stage 2. Details of the two stages are elaborated upon in the fol1owing sections.

3.1

Stage

1:

Content Analysis 3.1.1 What Is Content Analysis?

The use of content analysis originated in Sweden in the 18th century and became popular in the 1930s with the development of propaganda analysis and communication research. In the recent half-century

,

content analysis has become one of the important research methods in academic dissemination and other social sciences; it is often used to explore the significance of messages. A unique feature of content analysis is its ability to describe and analyze events that happened far away or in the past and people's innermost thoughts (Wang, 1999). Wimmer and Dominick (2002) indicated that content analysis is helpful in describing

(14)

152 TheS，ω砂 onBrand Personality 01 Automobile

disseminated contents

,

viewing characteristics of messages

,

comparing media content and the real world, assessing the image of particular social groups, and establishing research on media effects. As this method can be used to analyze the nature of disseminated contents and to examine the forms in which the contents are expressed

,

it was used in this study to examine the brand personalities adopted by car manufacturers to establish their respective brands through car advertisements.

3.1.2 Scope and Duration of Research a. Scope of Research

Magazine advertisements have always been popular subjects for research analysis. One reason is that static advertisements are easier to observe than dynamic advertisements: there is also greater design 仕eedom for magazine advertisements compared to newspaper advertisements (Liu, 1992). Another reason is that more information is contained in and presented by magazine advertisements compared to other forms of print advertisements. Therefore

,

magazine advertisements are used in this study for analysis.

To select the magazines, this study referred to the Government Information Office 2007 】告arbook，which stated that the top and second bestselling automobile magazines in 2006 were Taiwan Motor and Car Guide respectively. A survey by Rainmaker XKM Intemational Corp. on the readership of automobile magazines in Taiwan for the 2nd quarter, 2006, placed Car News at the top (0.2%). Joint runner-ups were 日iwan Motor and Car Guide

,

with a readership of 0.1 % each. Hence, the car advertisements in these three popular automobi1e magazines were studied. In addition

,

a survey conducted by the Industrial Technology Research

Institute in 2007 found that Taiwan's automobile market was still dominated by smaller passenger cars. The output value of these cars amounted to NTD93 billion

,

accounting for 63.99% ofthe total output value for automobiles. Hence

,

the focus of this study was .on general passenger cars

,

excluding larger vehicles

,

such as sport utility vehicles (SUV s) and trucks, to remove the impact of different

(15)

Chiao Da Management Review Vol. 33 No.2, 2013 153

b. Duration of Research

The annual sale of automobiles in Taiwan reached a peak of 560,000 vehic1es in 1994, but gradually dec1ined year-on-year afterwards. With its admission into the WTO in 2001, the Taiwanese automobile market became highly competitive. This robust scenario lasted until the end of 2005. In 2006, the output volume of domestically produced cars fell by 28.8% year-on-year. According to a 2007 market report on the state of the automobile industry published by the Indus仕ial Development Bureau of the Ministry of Economic Affairs

,

this dec1ine to a 20-year-low was caused by market saturation as well as increases in the intemational prices of oil and raw materials. This led to the govemment's decision to intervene and rescue the automobile market through various measures. Unfortunately, the news was leaked prematurely, causing consumers who had intended to purchase cars to instead adopt a wait-and-see attitude. Those who were anticipating preferential schemes even decided to cancel their earlier orders. Consequently, vehic1e sales in November 2008 were adversely affected, which was an unexpected development. Therefore

,

the duration of the study was 企om

January 2006 to October 2008, during which lO2 issues regarding the malaise in Taiwan's automobile industry and the brand personality strategies used in advertisements were discussed.

3.1.3 Construction of Categories and Reliability Analysis

The core task of content analysis is the construction of categories. Categories are the basic 凹lÏtsof content analysis and refer to the standards by which contents are c1assified. The categories adopted for this study are sinceri秒， excitement, competence, sophistication, and ruggedness, based on BPS of Aaker (1997). The characteristics associated with the various categories are shown in Table 1.

Reliability analysis when using content analysis refers to the ability of the adopted categories and units of analysis to withstand the test of different coders,

that is

,

whether different coders c1assify the contents being analyzed into similar categories to achieve a consistent result. The higher the consistency in content

c1assification by different coders, the more reliable the content analysis. Wang

(16)

154 The Stu砂 onBrand Personality 01 Automobile Advertisements and the Communication 01 Brand Personality during CA and advocated a mllllmum benchmark of 0.8 or greater as the reasonable standard. Wang's formula, as adopted for this stu旬" is elaborated upon as follows:

Mutual Agreement = 2M / N 1

+

N2

Reliability = (N x Average Mu仙al Agreement) / [1

+

(N -1) x Average Mutual Agreement], where

M: Number of categories that are in full agreement

Nl: Number of categories in full agreement that Coder #1 should have N2: Number of categories in full agreement that Coder #2 should have N: Total number of coders

3.2

Stage

2:

Experimental Design 3.2.1 Experimental method

In Stage 2, the experimental design was used to investigate the advertising effect created by matching three brand personali可 dimensions . excitement,

competence, and sophistication ﹒ Wl血 the level of models' physical at仕activeness.

Specifically

,

factorial design was used and three experiments were carried out

,

Study 1-3. In Study 1, excitement was manipulated as high and low and then matched with a highly attractive model and a normally at仕active model,

respectively. In Study 2 and Study 3, similar manipulations and matching were done with competence and sophistication, respectively.

3.2.2 Preparation of experimental materials a. Selection of Models

i. Potential Models

Photographs of models were selected from men's magazines that currently sell well in the Taiwan market, including Men

s

Health, Men

s

Uno, and GQ (Taiwan). Photographs with unclear or blurred images were excluded. It was further verified that the models in the photos had not previously endorsed any products professionally, nor had they been exclusive spokespersons for any

(17)

particular brands. Photographs of five male models

,

labeled 1-5 sequentially

,

were chosen for pretest.

ii. Scales

Two questions,“The model was good looking." and

“The model was very

attractive." were devised for the measurement ofthe physical attractiveness ofthe models, in accordance with Bower and Landreth (2001) and Tsai (2006). The 5-point Likert scale was used for rating.

iii. Selection Process

Undergraduates 企om I-Shou University participated in the selection process. To prevent sequence effì侃侃， participants were randomly assigned to one of five groups

,

with each group being represented by a male model as labeled above. Participants had 5-10 rninutes to complete a questionnaire after looking at the photograph, and all questionnaires were collected immediately upon completion. iv. Selection Results

For each model, the points rated by the participants to the two questions were added up. The average score was the level of attractiveness for that model. Model No. 1, with the highest score of3.500, was considered the highly attractive model. For the selection of normally attractive model

,

t-test was conducted to rate the scores of the other models against that of Model No. 1. The p-value for Models No. 4 and No. 5 was 0.177 and 0.075 respectively

,

indicating a lack of significant difference. The p-value for both Models No. 2 and No. 3 was less 血an 0.05 (0.000 and 0.020 respectively), indicating a significant difference in physical attractiveness between Model No. 1 and these two models. The average score for Model No. 2 was 2.667, which was lower than the median value ofModel No. 3. Considering that advertisement models are generally not unat仕active， Model No. 3 was selected as the normally attractive mode. The photographs of the highly attractive mode and normally attractive mode are shown in the appendix.

b. Selection of Automobile Brands

The frequency of the various brand personalities of cars

,

as deterrnined by content analysis, is listed in Table 2. Sophistication had the highest count (247)

,

accounting for 40% of brand personalities, followed by excitement (230, 38%) and competence (229, 37%). The remaining two dimensions had substantially

(18)

156 The Stu砂 onBrand Personality

0/

Automobile

Advertisements and the Communication

0/

Brand Personality

lower counts: sinceri砂 had 74 counts (12%), while ruggedness had only 39 counts (6%).

Table 2

The Frequency of Brand Personalities of Automobiles

Sophistication Excitement Competence

41 1.Protonlotus 32 1. Protonlotus 68

26 2.M紹dd 27 2.區亞 26

3.Porsche 21 3.Honda 26 3.Suzuki 21

4.Lexus 21 4.Volkswagen 15 4.Honda 19

5.Honda 20 5.Benz 13 5.Be也 19

6.Audi 19 6.BMW 15 6.Renault 17

7.Volvo 17 7.Lexus 12 7.Volkswagen 16

8.Be血 16 8.Audi 11 8.BMW 14

9.Renault 11 9.Ford 11 9.Daihatsu 10

10 .Infini ti 11 10Infiniti 10 1O.Ford 6

11.BMW 8 11.Skoda 8 11.Alfa Romeo 6

12.Mazda 7 12.Toyota 7 12.Lexus 7

13.Peugeot 6 13.Peugeot 7 13.Skoda 3

14.Jaguar 6 14.Renault 7 14.Peugeot 3

15.Ford 5 15.Smart 7 15.Infmiti 2

16.Suzuki 5 16.Daihatsu 6 16.Toyota 2

17.Skoda 4 17.此1ini 6 17.Nissan 2

18.Buick 3 18.Suzuki 4 18.Porsche

19.扎1asersti 3 19.Fiat 3 19.Youlon GM

20.Toyota 3 20.Alfa Romeo 3 20.Flat

21.Yulon GM 3 21.Volvo 2 21.~

22.Daihatsu 3 22.Jaê!ar 22.Buick

23.Nissan 3 23RlSSE 23.扎1itsubishi

24.Lancia 2 24.扎1itsubishi

25.Mitsubishi 25.Citroen

26.C趾ysler 26.Porsche

27.Citroen 27.扎1arserati

(19)

It can be seen from Table 2 that Proton Lotus was ranked high while Mitsubishi was ranked low in terms of excitement

,

sophistication

,

and competence. Both automobile brands were excluded from the study because they each embodied more than one brand personality. Considering consumers' awareness of the various automobile brands

,

Mazda and Nissan were selected as the representatives for high and low excitement respectively. High and low competence were represented by Volvo and Audi respectively

,

while high and low sophistication were represented by Volkswagen and Hyundai respectively.

3.2.3 Experiment

Following the selection of automobile brands

,

four full-colored advertisements were designed for each of Study 1 - 3. Besides the variable being manipulated for each study, all other variables remained the same.

a. Scale Design

The scales included ratings of the model's attractiveness and the communication effect of the advertisement

,

encompassing the two elements of advertisement and brand attitude. The two questions for rating the model's attractiveness were similar to those used in the pretest according to Bower and Landreth (2001) and Tsai(2006). For advertisement attitude

,

six questions were designed based on Holbrook and Batra (1987) and Liang (1992). The five questions related to brand attitude were made with reference to Leclerc

,

Schmitt and Dube (1994). All questions were rated using a 5-point Likert scale.

b.Participants and Process

Considering that office workers are likely to have the economic capability to purchase a car, and for the sake of convenience, employees of the Kaohsiung Oil Refmery were recruited as participants. Study 1-3 each had four experimental groups, with 30 pa叫cipants in each group, making a total of 360 subjects. Respondents were assigned randomly to one of the units and conducted a simple briefing before the experiment to explain the rules. The subjects were asked to complete a questionnaire immediately after viewing the advertisement copy assigned to their group. The questionnaires were coUected after being duly completed. The entire experiment process lasted approximately 5-10 minutes.

(20)

158 The Stu砂 onBrand Personali秒 0/Automobile Advertisements and the Communication 0/ Brand Personality

4. Results

4.1

Results of Content Analysis

Through content analysis, it was found that between January 2006 and October 2008

,

604 car print advertisements appeared in Car News

, Taiwan Motor

,

and Car Guide. Among these automobile magazines

, Car

Guide had the most car advertisements (261)

,

followed by Car News (189)

,

and lastly

,

Taiwan Motor (154).

4.1.1 Reliability Test of Content Analysis

A total of 48 car advertisements 企omthe previous one and a half years were used in a pre-test to establish the consistency level between the three coders, A

-C

,

the more consistent the coding

,

the higher the degree of reliability. After the coding results were entered into the formula proposed by Wang (1999), a reliability rating ofO.9196 was derived. The rating was higher than the benchmark of 0.8 set by Wang (1999), and thus, was deemed reliable. The three coders then officially began to separately analyze the brand personalities in the 604 car advertisements. The brand personality scores analyzed by Coders A, B, and C were 1044

,

1072

,

and 1012 respectively. Table 3 shows the number of brand personalities that are in full agreement between any two coders as well as the degree of agreement between the two. The results were inputted into Wang's formula, deriving an overall reliability rating ofO.93859.

Table 3

Frequency in Full Agreement between Coders and Mutual Agreement

Coder A Coder B Coder C

Coder A

Coder B 863(0.815689)

Coder C 894(0.869649) 857(0.869649) Note: In Parentheses are Mutual Agreements between Coders

(21)

4.1.2 Frequency of Brand Personality Dimensions

To completely present all the brand personalities used in the car advertisements

,

whenever two personalities were identified in the same advertisement

,

both would be included in the analysis. Furthermore

,

as the analyses of the three coders were not identical, all discrepancies were elirninated and only those findings for which all three coders were in agreement were eventually compiled. The majority ofthe car advertisements (247) showcased the

sophistication dimension of brand personality. It was followed by excitement

(230), competence (229), sincerity (74), and ruggedness (39).

4.2

Results of the Various Studies 4.2.1 Reliability and Validity

From the three studies, the Cronbach's alpha of physical attractiveness,

advertisement attitude, and brand attitude are listed in Table 4. Since all the reliability values are greater than 0.7, it can be established that the scales used in this study were reliable (DeVellies

,

2003).

To test for validity

,

factor analysis was separately performed on the three factors of physical attractiveness, advertisement attitude, and brand attitude (Table 4). For each construct, only one factor was ex仕actedbased on the criterion of its eigenvalue being larger than 1. This was to confirm the singularity of the construct. The factor loadings of the various construct measures and the percentages of variation extraction were relatively high, indicating that the scales used in this study had convergent validity. The discriminant validity between physical attractiveness, advertisement attitude, and brand attitude are shown in Table 5. According to Fomell and Larcker (1981), there is discriminant validity

between two constructs when the percentage of variation extracted 企om them separately is greater than the square of the correlation coefficient between them. Hence, based on Table 5, it can be concluded that the scales used in this study had discriminant validity.

(22)

160 The Study on Brand Personality

0/

Automobile Advertisements and the Communication

0/

Brand Personality Table 4

Reliability and Validity of Study 1-3

Factor Loadings Cronbach's Alpha Percentage of Extracted

Items Variation

(%)

SI S2 S3 SI S2 S3 SI S2 S3

The model was good

0.965 0.971 0.958

Physical looking

0.925 0.938 0.909 93.092 94.331 91.755

Attractiveness 百lemodel was very ₀_.₉₆₅ _0.971 ₀_.₉₅₈ attractJ.ve

1 liked the ad 。.896 0.912 0.905 The ad was great 0.913 0.944 0.907 1 felt positively about

0.889 0.914 0.882

the ad

Ad 1 prefeηed the ad's

0.862 0.934 0.885 0.948 0.964 0.951 79.443 84.915 80.459

Attitude presentatlOn

The ad was very

0.885 0.921 0.901

mterestJ.ng The ad attracted

0.904 0.904 0.903

people's attention 1 liked this brand of

0.755 0.847 0.905

car very much

The quality ofthis ₀_.₉₀₂ _0.883 ₀_.₉₀₇ brand is good

Brand 1 feel positively

0.867 0.852 0.848 0.874 0.888 0.887 67.765 69.212 69.277

Attitude towards this brand This brand gives me

0.848 0.796 0.845

pleasure

1 am very attracted to

0.732 0.778 0.796

也isbrand

(23)

Chiao Da Management Review Vol. 33 No.2, 2013

Table 5

Discriminant Validity of Study 1 / Study 2/Study 3 Physical Attractiveness Ad Attitude Physical Attractiveness 0.930/0.943/0.918 AdAttitude Brand Attitude 0.726/0.845/0.790 0.113/0.143/0.141 0.794/0.849/0.806 0.153/0.141/0.177 161 Brand Attitude 0.678/0.692/0.693 Note: The Diagonal is the Percentage ofEx仕actedVariation, and the Remaining is Square of the Correlation Coefficient

Table 6

The Result of ANOV A-Study 1

Source Dependent SS df MS F p

Variable

Physical Attractiveness AdAttitude 90.133 90.133 375.169 0.000..

Brand Attitude 0.225 0.225 0.458 0.500

Brand Personality AdAttitude 0.181 0.181 0.7558 0.387

Brand Attitude 3.745 3.745 7.608 0.007".

Physical Attractiveness x AdAttitude 0.093 0.093 0.385 0.536

Brand Personality _{Brand Attitude} _0.261 ₀_.₂₆₁ _0.385 ₀_.4₆₈

Error AdAttitude 27.867 116 0.492 Brand Attitude 57.107 116 0.492 Total AdAttitude 118.274 119 Brand Attitude 61.339 119 串串串 Note: p<.05; "p<.OI

(24)

162

4.2.2 Manipulation Test

TheS的砂 onBrand Personality 01 Automobile

F or Study 1, the average score of the highly attractive model group was 3.883

,

while that for the normally attractive model group was 2.475. The t-test resulted in a t-value of 11.983 and a p-value of 0.000 < 0.05. For Study 2

,

the average score of the highly attractive model group was 4.067

,

while that for the normally attractive model group was 1.850. The t-test resulted in a t-value of 18.673 and ap-value ofO.OOO < 0.05. For Study 3, the average score ofthe highly attractive model group was 3.967, while that for the normally attractive model group was 2.225. The t-test resulted in a t-value of 14.857 and a p-value of 0.000 <0.05.

For all of the three studies, the average scores of the highly attractive model group were consistently and significantly higher than that of the normally attractive model group. Hence, the manipulation of physical attractiveness in the three studies can be deemed successful.

4.2.3 Results of Study 1

a. Multivariate Analysis ofVariance

MANOVA was carried out to investigate the impact of physical attractiveness and brand personality on communication effect. The results show that the Wilks' lambda (A) ofphysical attractiveness was 0.232 (p = 0.000 < 0.01)

,

indicating that physical attractiveness had a significant impact on communication effect. The A ofbrand personality was 0.938 (p = 0.024 < 0.05), again indicating a

significant impact on communication effect. There was no significant interaction e宜ect between physical attractiveness and brand personality (A = 0.993

,

p 0.683).

b. Analysis ofVariance

ANOVA was used to investigate the effect of physical attractiveness and brand personality on advertising attitude (Table 6). The results show that physical attractiveness had a significant impact on advertising attitude (F (1, 116) 375.169, P = 0.000 < 0.01). Furthermore, it can be seen from Table 7 that the

(25)

impact of a HAM on advertising attitude was significantly greater than that of a NAM (3.805 versus 2.939). However

,

brand personality had no significant impact on advertising attitude. Neither was there any significant interaction effect between physical attractiveness and brand personality with regard to advertising attitude. Physical attractiveness had no significant impact on advertising attitude,

but brand personality did (F (1

,

116) = 7.608

,

p = 0.007 < 0.01). The impact of a

high excitement brand personality on brand attitude was 3.790, which was significantly greater than that of a low excitement brand (3.437). There was no significant interaction effect between physical attractiveness and brand personality with regard to brand attitude.

c. Simple Main Effect Test- Advertisement Attitude

The average scores of the advertisement attitude of the various groups are listed in Table 7. Through simple main effects test, it was confirmed that for both high and low excitement brand personalities

,

a highly attractive model was significantly more suitable as a spokesperson compared to a normally attractive model; the scores were (F (1, 116) = 152.63, p = 0.000 < 0.01) and (F (1, 116) =

137.26,p = 0.000 < 0.01) respectively. Hence, Hl-l is only partially supported. d. Simple Main Effect Test - Brand Attitude

The average scores of the brand attitude of the various groups are listed in Tab1e 8. Through simp1e main effects test, it was confirmed that for a high

excitement brand personality

,

there was no significant difference regardless of whether a highly attractive model or normally attractive model was used (F (1

,

116) = 1.12, P = 0.292). However, for a low excitement brand personality, a normally attractive model was marginally more suitable than a highly attractive model (F (1, 116) = 3.03, p = 0.085 < 0.10). Hence, Hl.♂ i臼s only partially supported.

(26)

164 Brand PersonalityPhysical Attractiveness Highly Physical Attractiveness Normally Physical Attractiveness Main EfIect-Brand Personality

Note: In Parenthesis is Standard Deviation.

Brand PersonalityPhysical Attractiveness Highly Physical At仕activeness Normally Physical Attractiveness Main EfIect -Brand Persοnality

Not活: In Parenthesis is Standard Deviation.

TheSω砂 onBrand Persona/ity of Automobile

Advertisements and the Communication of Brand Persona/ity

Table 7

Ad Attitude for each Group of Study 1

High Main

Effect-Low Excitement Physical

Excitement Attractiveness 3.872 (0.506) 3.739 (0.378) 3.805 2.083 (0.555) 2.061 (0.504) 2.939 2.978 2.900 Table 8

Brand Attitude for each Group of Study 1

High Main

Effect-Low Excitement Physical

Excitement

Attractiveness 3.880 (0.555) 3.433 (0.784) 3.657 3.700 (0.696) 3.440 (0.749) 3.613

(27)

•

Chiao Da Management Review Vo/. 33 No.2, 2013

4.2.4 Results of Study 2 a.MANOVA

165

The A of physical attractiveness was 0.074 (p = 0.000 < 0.01), indicating

that physical attractiveness had a significant impact on communication effect. The A of brand personality was 0.895 (p = 0.002 < 0.01), indicating a significant impact on communication effect. There was no significant interaction effect between physical attractiveness and brand personality (A

=

0.981

,

p

=

0.332). b.ANOVA

The ANOVA results (Table 9) show that physical attractiveness had a significant impact on advertisement attitude (F (1, 116) = 14.061, p = 0.000 < 0.01). Table 10 shows that the impact of a highly attractive model on advertisement attitude was significantly greater than that of a normally attractive model (4.044 versus 2.881). Brand personality similar1y had a significant impact on advertisement attitude (F (1, 116) = 11.237, p = 0.001 < 0.01). The brand

attitude of a high competence brand personality was 2.778, which was significantly lower than that of a low competence brand personality (2.983). There was no significant interaction effect between physical attractiveness and brand personality with regard to advertisement attitude.

Brand personality had a significant impact on brand attitude (F (1, 116) = 3.087,p

= 0.082 < 0.10). According to Table 11, the brand attitude of a high competence

brand personality was 3.337, which was significantly lower than that of a low

competence brand (3.570). Physical attractiveness had no significant impact on brand attitude. Neither was there any significant interaction effect between physical at仕activenessand brand personality with regard to brand attitude.

c. Simple Main Effect Test - Advertisement Attitude

Table 10 shows the average scores of the advertisement attitude of the various groups. Through simple main effects test, it was found that for both high and low competence brand personalities, a highly attractive model was significantly more suitable as a spokesperson compared to a normally at仕active mode1. The scores are (F (1, 116) = 225.42, p = 0.000 < 0.01) and (F (1, 116) = 185 .41,p = 0.000<0.01) respectively. Hence, H2-1 is only partially supported.

(28)

166 The Stu砂 onBrand Personality 01 Automobile Advertisements and the Communication 01 Brand Personality

Table 9

Source Dependent SS df B宜S F p Variable

•

Physical Ad 162.556 162.556 1441.061 0.000... At仕aCÌ1veness Attiωde Brand Attitude 0.901 0.901 1.703 0.194 Brand Personality Ad 1.268 1.268 11.237 0.001... Attitude Brand Attitude 1.633 1.633 3.087 0.082. Physical Ad 0.156 0.156 1.387 0.241 At仕activeness x A位itude

Brand Personality Brand

Attitude _0.385 _0.385 ₀_.728 0.395 Error Ad 13.085 116 0.113 Attitude Brand 61.379 116 0.529 Attitude Total Ad 64.299 119 A甜tude Brand 64.299 119 Attitude Note:. p< .10; •• p< .05; ••• p< .01 Table 10

Brand

High Low Main

Effect-PersonalityPhysical Physical

Attractiveness Excitement Excitement Attractiveness

Highly Physical 3.906 (0.383) 4.1833 (0.311) 4.044 Attractiveness Normally Physical 1.650 (0.275) 1.7833 (0.364) 2.881 Attractiveness Main Effect- 2.778 2.983 Brand Personality

(29)

Table 11

Brand Attitude for each Group of Study 2 Brand

High Low Main

Effect-PersonalityPhysical Physical

Attractiveness competence competence Attractiveness

Highly Physical 3.480 (0.647) 3.600 (0.515) 3.540 Attractiveness Normally Physical 3.193 (0.713) 3.540 (0.962) 3.453 Attractiveness Main Effect- 3.337 3.570 Brand Personality

Note: In parenthesis is standard deviation.

d. Simple Main Effect Test -Brand Attitude

Table 8 lists the average scores of the brand attitude of the various groups. The simple main effects test revealed that for both high and low competence

brand personalities, there was no significant difference regardless of whether a

highlyat仕active model or normally a前ractive model was used. The scores are (F

(1, 116) = 2.33,p = 0.130) and (F (1, 116) = O.lO,p = 0.750) respectively. Hence,

H2-2 is not supported. 4.2.5 Results of Study 3 a.MANOVA

The A ofphysical attractiveness was 0.284 (p = 0.000 < 0.01), indicating that

physical attractiveness had a significant impact on communication effect. The A of brand personality was 0.945 (p = 0.040 < 0.05), again indicating a significant effect on communication effect. There was no significant interaction effect between physical attractiveness and brand personality with regard to communication effect (A = 0.996, p = 0.801).

b.ANOVA

Table 12 shows the ANOVA results, which indicate that physical attractiveness had a significant impact on advertising attitude (F (1, 116) =

289.683, p = 0.000 < 0.01). Furthermore, the impact of a highly attractive model on advertisement attitude was significantly higher than that of a normally

(30)

168 The Stu砂 onBrand Personality 01 Automobi/e

attractive model (3.881 versus 2.990). Brand personality had no significant effect on advertising attitude. Neither was there any significant interaction effect between physical attractiveness and brand personality with regard to advertising attitude. However

,

physical attractiveness had a significant impact on brand attitude (F (1

,

116) = 3.916

,

p = 0.05 < 0.1). Table 14 shows that the impact of a highly at仕active model on brand attitude (3.837) was significantly higher than that of a normally at仕active model (3.718). Brand persona1ity had a significant effect on brand attitude (F (1, 116) = 4.577, P = 0.035 < 0.05). As seen in Table 14,

the brand attitude of a high sophistication brand persona1ity was 3.847, which was significantly higher than that of a low sophistication brand (3.580). There was no significant interaction effect between physical at仕activeness and brand personality with regard to brand attitude (Table 12).

c. Simple Main E能ctTest -Advertisement Attitude

The average scores of the advertising attitude of the various groups are shown in Table 13. Simple main effects test showed that for both high and low

sophistication brand personalities, a highly attractive model was significantly more suitable as a spokesperson compared to a normally at甘active model. The scores were (F (1, 116) = 199.82, P = 0.000 < 0.01) and (F (1, 116) = 175.77, p = 0.000 < 0.01) respectively. Hence, H3-1 is only partially supported.

d. Simple Main Effect Test - Brand Attitude

The average scores of the brand attitude of the various groups are shown in Table 14. Through simple main effects test

,

it was found that for both high and

low sophistication brand persona1ities

,

there was no significant difference

regardless of whether a highly attractive model or normally attractive model was used. The scores were (F (1, 116) = 0.98, p = 0.324) and (F (1, 116) = 0.11, P =

(31)

Chiao Da Management Review Vol. 33 No.2, 2013 169 Table 12

Source Dependent SS df MS F p Variable Physical Attractiveness Ad 95.111 95.111 289.683 0.000... Attitude Brand Attitude 1.825 1.825 3.916 0.050. Brand Personality Ad 0.752 0.752 2.291 0.133 Attitude Brand Attitude 2.133 2.133 4.577 0.035.. Physical Attractiveness x Ad 0.067 0.067 0.204 0.653

Brand Personality Attitude

Brand Attitude ₀.108 0.108 0.232 0.631 Error Ad 38.086 116 0.328 Attitude Brand Attitude 54.072 116 0.466 Total Ad 134.016 119 Attitude Brand 58.139 119 Attitude Note:. p<.lO;.. p<.05; ••• p<.OI

(32)

170 The Study on Brand Personality 0/ Automobile

Advertisements and the Communication 0/ Brand Personality

Table 13

Highly Physical Attractiveness Nonnally Physical Attractiveness Main Effect. Brand Personality High sophistication 3.983 (0.607) 2.156 (0.678) Low sophistication 3.778 (0.456) es- -e3- te-

el--位叫附一

E. 明廿 -m Mb.m-3

ap--MM

一

Brand PersonalityPhysical Attractiveness 2.044 (0.527) 2.990 3.069 2.911

Note: In parenthesis is standard deviation.

Table 14

Brand Attitude for each Group of Study 3 Brand PersonalityPhysical Attractiveness Highly Physical Attractiveness Nonnally Pbysical Attracbveness Main Effect-Brand Personality

High Low Main

Effect-Physical sophistication sophistication Attractiveness 3.940 (0.520) 3.733 (0.474) 3.837 3.753 (0.714) 3.427 (0.927) 3.718 3.847 3.580

(33)

Chiao Da Management Review Vo/. 33 No.2, 2013 171

5. Conclusion and Discussion

5.1

Conclusion and Contributions

5.1.1 Content Analysis

Content analysis was performed on 604 car print advertisements found in Taiwan's three bestselling automobile magazines, over a period of 102 weeks (January 2006 - October 2008). The advertisements were categorized according to the five dimensions of the BPS proposed by Aaker (1997): sinceri秒， excitement,

competence, sophistication, and ruggedness. The dimension that appeared the most times was sophistication (247), followed by excitement (230), and competence (229). Sincerity and ruggedness were used significantly less

,

appearing in only 74 and 39 advertisements respectively. 5.1.2 Experimental Method

Three studies were conducted

,

and it was found that both physical attractiveness and brand personality had significant impacts on communication effect. However

,

there was no significant interaction effect between physical attractiveness and brand personality with regard to communication effect.

The results of Study 1 show th剖 a higWy attractive model had a significantly higher impact on advertisement attitude compared to a normally attractive model,

while brand attitude was significantly higher for high excitement, rather than low

excitement, brand personality. In terms of impact on advertisement attitude, a highly attractive model was significantly more suitable as a spokesperson for both high and low excitement brand personalities. Hence, Hl-l is only partially supported. For brand attitude, there was no significant difference between using a highly attractive model or normally attractive model for a high excitement brand personality. However, a normally attractive model was marginally more suitable for a low excitement brand personality. Hence, Hl-2 is only partially supported.

The results of Study 2 similarly show that a highly attractive model had a significantly higher impact on advertisement attitude compared to a normally attractive model. For a high competence brand personality, the advertisement and

(34)

172 The Study on Brand Personality 01 Automobile Advertisements and the Communication 01 Brand Personality brand attitude were both significantly lower than for a low competence brand personali可. In terrns of impact on advertisement attitude, a high1y at仕active model was significantly more suitable as a spokesperson for both high and low competence brand personalities. Hence, assumption H2-1 is only partial1y supported. With regard to brand attitude, there was no significant difference between a highly attractive model and norrnal1y attractive model for both high and low competence brand personalities. Hence, H2-2 is not supported.

The results of Study 3 show that for both advertisement and brand attitude, a highly a位ractive model scored significantly higher than a norrnal1y attractive model. The impact of a high sophistication brand personality on brand attitude was significantly higher than that of a low competence brand personality. With regard to advertisement attitude

,

a highly attractive model had a significant impact for both high and low sophistication brand personalities. Hence

,

H3-1 is only partially supported. For both high and low sophistication brand personalities

,

there was no significant difference between a high1y attractive model and norrnally at仕active model with regard to brand attitude. Hence, H3-2 is not supported.

The main contribution of this study is its investigation of the 可pes ofbrand personalities adopted by car manufacturers for communication. The findings allow practitioners and academia to have a more complete understanding of various brand personality strategies adopted by the market. These fmdings can also serve as a reference for car manufacturers when deciding upon a brand strategy in the future.

In the past

,

studies on the physical attractiveness of models have been confmed to how to enhance advertising impact, but have rarely been applied in the communication of brand personality. On the other hand, research on the match-up hypothesis has tended to focus more on the match-up between advertising models and product types, but has rarely considered the match-up of advertising models and brand personality in brand personality communication. In terrns of theo旬， this study expands the scope of interpretation in relation to the match-up hypothesis and the physical attractiveness of advertising models. From a practical viewpoint, the scope of application for the match-up hypothesis and