NSSA - Country of Origin Cues in Brand Personality Assessment

Ryan Fehr
The College of New Jersey

     During the past several decades, research on the quantification and classification of human personality has increased dramatically. Of particular interest is the Big Five Model, which posits a set of five latent dimensions of personality including Neuroticism, Extraversion, Openness to Experience, Agreeableness, and Conscientiousness. A detailed explanation of the theory’s history can be found in Digman’s (1990) comprehensive literature review. The applicability of the model is well-documented. Narcissistic and antisocial behavior, for example, correlate with low levels of Agreeableness, and may therefore be measured with the Big Five (Costa & McCrae, 1990). In both children and adults, Neuroticism is associated with depression (Ehrler, Evans, and McGhee, 1999).
     In an effort to expand the applicability of the theory behind the Big Five, Aaker (1997) utilized factor analysis to develop a theoretical framework for quantifying the assessment of brand personality in psychological and marketing research. Defined formally as “the set of human characteristics associated with a brand,” researchers have examined the phenomenon of brand personality for many years (Birdwell, 1968; Thakor, 1996; Aaker, 1997).
     The associations individuals make with brands are strong, and have been shown to both contribute to and reflect our personal identities (Belk, 1988). In more recent research, it has also been shown that our relationships with brands are dynamic, much akin to human relationships (Aaker, 2004). The importance of brand personality is not lost to marketers either, who have for years utilized the concept as a form of product differentiation and identification (Halliday, 1996; Elliott, 2004; Ives, 2004).
     Academic research on brand personality has suffered, however, due in part to a lack of consensus on the definition of brand personality as a psychological construct, and the methodology of its analysis. In 2001, an attempt was made to directly apply the penta-factorial Big Five Model of personality in the context of brand personality assessment. However, the resulting data proved neither tenable nor replicable according to the scree test and Everett method, respectively (Caprara, Barbarnelli, and Gianluigi, 2001). Of greater promise is the Aaker (1997) study which, while parallel to the Big Five Model in terms of statistical methodology, developed a latent factor model unique to the brand personality construct. A Brand Personality Inventory was also developed and validated in the same study, which utilized a 42 question Likert-scale test to measure participant ratings of brand personality.
     In a follow-up experiment, the Big Five of brand personality was assessed in a cross-cultural framework (Aaker, Benet-Martinez, and Garolera, 2001). The research assessed brand personality from the indigenous perspective, replicating the factor analyses of the original Big Five brand personality study with a combined emic-etic approach. Results found that while some brand personality dimensions remain consistent across cultures, others were unique to the individual countries (Aaker, Benet-Martinez, and Garolera, 2001).
     Despite this detailed analysis of brand personality from the indigenous perspective, no research has yet utilized the Big Five of brand personality to examine culture as an externally imposed entity. In psychology and marketing literature, the concept of culture as an external force is generally referred to as “country of origin” (COO) and, as an influencing factor in consumer decision-making processes, has been studied heavily.
     The first research study on country of origin was conducted in 1965, and concluded that country of origin did indeed influence product perception (Schooler, 1965). An early literature review, now regarded as a classic in the field, generally agreed that country of origin influences product perceptions. Findings from this early research include a positive correlation between product evaluation and the degree of a country’s economic development, a higher rating of foreign products by females than males, higher rating of products from countries of a person’s ancestors, and a view of foreign products as “riskier” than domestic products (Bilkey, and Nes, 1982).
     More recently, a number of studies have contradicted the claims of early work cited by Bilkey and Nes. For instance, while early studies found a negative relationship between the use of country of origin cues and product familiarity, recent findings show a positive relationship (Huddleston, Good, and Stoel, 2000).
     In much the same manner as with brand personality research, COO literature has suffered from a lack of consistency in methodology and definitional constructs. This fact is readily admitted by at least some members of the academic community, with a recent literature review concluding that “most of the recent country-of-origin studies provide us with little generalizable knowledge” (Ozomer, and Cavusgil, 1991, p.274).
     It is the goal of the present research, therefore, to bridge the gap between COO and brand personality research through use of the Big Five of brand personality. This is accomplished through analysis of Aaker’s (1997) Brand Personality Inventory, and the degree to which changes in COO cues influence participant response to the scale. Such a design is of additional import, as it represents the first attempt to utilize the BPI in analysis of a single consumer cue through the development of a hypothetical brand model.
     Breadth is added to the design by examining ethnocentricity as an additional independent variable. The concept of ethnocentricity in consumers has been examined in a number of studies, most often through use of Shimp and Sharma’s (1987) previously validated Consumer Ethnocentricity Scale (CETSCALE). The CETSCALE utilizes a variety of Likert scale questions to assess ethnocentricity, operationally defined as the belief held by consumers about the appropriateness and morality of purchasing foreign-made products in place of locally made products.

Design
     The study was designed as a questionnaire, to be taken in two subsequent parts. First, students were given a description of “Lasri Shirt Company,” a fictitious clothing company. The 150 word description contained mostly neutral details of the company, including sales figures, number of employees, the most profitable line of merchandise, and the slogan “Lasri: Better clothes for a better day.” Also included in each description was a COO cue. Four countries were included in the project, including the United States, Syria, Spain, and Japan. The choice of using the United States is self-explanatory, as it serves for an appropriate baseline in analyzing responses to other countries. Japan and Spain were both chosen as a logical extension of Aaker’s (2000) previous work with cross-cultural issues in brand personality. Syria was selected as a recognizable country which, while clearly Middle Eastern, is not found in the news as often as countries such as Iraq and Iran. A logical extension from this decision is that the cues will be based more on culture, and less on specific political or otherwise temporal issues. Therefore, in the choice of Syria it is possible to measure a culture distinct from Japan and Spain, yet representative of a major global culture.
     Participants were randomly assigned to one of the four countries. After reading the brief company description, participants were provided with an explanation of the project and directions on how to proceed. This text was based on Aaker’s directions, and altered minimally to fit the context of the current task. Next, participants responded to the Brand Personality Inventory, a 42 question survey measured with the Likert system. Participants were prompted with a series of adjectives and asked to rate them on a scale of 1-7 from “not at all descriptive” to “very descriptive.” See Table 1 for a listing of the adjectives and their factors. Since the adjectives were presented in the pertinent body of research in groups according to their factors, it was necessary to randomize their order. This ensured that no primacy or recency effects could influence one or two factors unequally. The adjective order was randomized by first assigning them numbers in order, from 1-42, then randomizing them with a random number generator.
     The second part of the questionnaire included the CETSCALE for ethnocentricity and a demographic sheet. The CETSCALE was originally developed by Shrimp and Sharma (1987) to measure participant ethnocentricity. It consists of 17 questions related to the country-of-origin perceptions of American consumers. The questionnaire utilized a seven-point Likert scale, 1= strongly disagree, 7= strongly agree. The demographic sheet was open-response, and asked participants to list their gender, age, ethnic background, and class year.

Materials and Procedures
     Prior to beginning the survey, participants were provided with informed consent forms, verifying that they were above 18 years of age and agreed to the terms of the study. Participants were instructed to sign the form and return it to the administrator. The questionnaire was provided in two separate parts. First, participants were given the brand description, instructions for the Brand Personality Inventory, and the 42-point inventory itself. Next, students were instructed to complete the second questionnaire, which included the CETSCALE and a brief demographic profile. Once all students completed both sections of the survey, they were thanked for their participation and given debriefing forms, which explained the study in further detail and encouraged them to email the research team if interested in learning the results of the study.

Table 1
Brand Personality Inventory

Table 2
Mean Scores and Standard Deviations for BPI Factors

Table 3
Mean Scores and Standard Deviations for Item 26