Modeling links between the decision-making process and luxury brand attachment: An international comparison

Abstract

This research aims to extend the analysis of the factors influencing consumer purchase of luxury goods. The effects of brand and country of origin (CoO) on the purchase decision are specifically considered. A total sample of 1103 respondents in seven countries (China, France, India, Italy, Japan, Russia, USA) is analyzed through a structured questionnaire administered online. The research tests the scales of decision-making and brand attachment through exploratory and confirmatory factor analysis. It also tests the causal model of relationships between decision-making and luxury brand attachment.

The richness of this research is related to the intercultural analysis of the results obtained in seven countries. These results concern the differences and similarities in: (1) defining the concept of country of origin (country of design, country of manufacture/country of assembly); (2) the combined effects of factors, including brand, CoO, price, guarantee, design and advertising, in the purchasing decision for luxury brands; and (3) the links between decision-making and attachment of the consumer to the luxury brand.

The results are interesting in terms of management recommendations for a company that wishes to expand internationally into one of the geographic areas covered by this study, since significant differences are observed.

Abstract

此项研究旨在扩展影响消费者购买奢侈品的因子分析。特别考虑了品牌原产地对于购买决策的影响。已知的消费者行为研究采取了两种理论解释奢侈品消费。他们最初的灵感来自于专门面向全社会动机的社会心理学。最近，这一趋势已经得到补充基于个人愿景的理论解释。

无论何种视角，先前的奢侈品领域的研究展示了品牌作为战略的载体，使用于消费者制定决策过程 (Kapferer, 1997, 1998; Vigneron & Johnson, 1999, 2004, Dubois, Czellar & Laurent, 2005) 。此研究适合这一理论视角，考虑了品牌在消费者决策过程中的特别作用。

品牌虽然在传统认识中认为是消费者决策的核心驱动，我们观察到公司也将其与原产地联系在一起开发他们的国际营销策略。原产地对于消费者评估和购买意愿的影响是国际营销研究中普遍的主题 (Usunier, 2006; Bloemer, Brij & Kasper, 2009) 。然而，据我们所知，在奢侈品行业并没有这一相关主题的研究。原产地的定义在整个学术界广泛的蔓延，从而支持国际贸易的经济发展 (Dichter, 1962; Schooler, 1965; Roth & Romeo, 1992; Nebenzahl & Samie, 1994, Nebenzahl, Jaffe & Lampert, 1997, Jaffe & Nebenzahl, 2001; Samiee, Shimp & Sharma, 2005; Roth & Diamantopoulos, 2009)

此研究的1103份样本来自于7个国家（中国、法国、印度、意大利、日本、俄罗斯、美国），问卷通过在线收集整理。此研究通过探索性因子分析和验证性因子分析测试了制定决策和品牌附属品的范围。同时也对制定决策和奢侈品牌附属品关系的因果模型进行了测试。

此研究最有价值的是对7个国家的数据进行了跨文化分析。研究结果涉及以下不同点和相同点：i) 定义了原产国概念 (设计地, 生产地/装配地); ii) 奢侈品购买决定的影响因子组合：品牌、原产地、价格、保证、设计和广告； iii)奢侈品消费者的决策和附属品之间的关系。

研究结果对于公司管理建议很有趣，对于公司扩展国际市场很有帮助。

Keywords::

• country of origin
• luxury
• international
• cross-cultural analysis
• consumer behavior

Keywords::

• &原产地
• 奢侈品
• 国际性
• 跨文化分析
• 消费者行为

Notes

^1. See for example forecasts from Bain & Company's study “Luxury Goods Worldwide Market”, Spring 2011.

^2. With exploratory factor analysis, we are trying to identify underlying variables that explain the origin of correlations within all of our observed variables.

^3. Confirmatory factor analysis aims to validate a posteriori the structure of a measurement scale. It relies on a reverse process of exploratory factor analysis, in that it is data that will confirm the factors. It also provides additional goodness-of-fit statistics of the model to the data in order to judge the likelihood of the model (Hu & Bentler, 1999) and allows the calculation of indicators for measuring reliability and validity. The estimation method most commonly used in structural analysis is maximum likelihood (ML), which tolerates moderate violations of multi-normality.

^4. The reliability or internal consistency of a measurement scale measures the degree of stability of results when applying the instrument again in identical conditions. In this phase of analysis, reliability is measured by Cronbach's α (1951). It must be greater than .60 for exploratory research and .80 for applied research (Nunnally, 1978; Peterson, 1994).

^5. This first measure of reliability was complemented in the confirmatory factor analysis with Jöreskog's ρ (1971), which is considered more reliable than Cronbach's α as it is less sensitive to the number of items in the scale. In this same phase of analysis, convergent validity was measured through the ρcv, which must be greater than .50.

^6. The confirmatory analysis provides lambdas (λi), which are the standardized correlation coefficients of variables with latent variables. They are all statistically significant at the 5% significant level, since the critical ratios are all above 1.96.

^7. (Ranking, Relative Standardized Regression Weight, Gap with the overall model).