Consumer Innovativeness for Fashion as a Second Order Construct: A Cross-Cultural Study
Ronald E. Goldsmith,Daekwan Kim,Leisa Flynn,Wan-Min Kim
DOI: https://doi.org/10.1080/20932685.2010.10593057
2010-02-01
Journal of Global Fashion Marketing
Abstract:Setting and changing prices is difficult for marketing managers, especially for new products. Consumer price sensitivity (how consumers react to prices or to price changes) makes a difference in the likelihood consumers will buy (Ainslie and Rossi 1998). For many new products, the reactions of the earliest buyers or innovative consumers are critical to success (e.g., Munnukka 2005). But how important is the price of new products to innovators? The answer to this question is of theoretical interest and of practical importance to marketers, especially when marketing new fashions, where innovators play a crucial role in consumer acceptance. Existing research shows that innovators are relatively less sensitive to prices of new products than are later adopters; that is, they are more willing to pay higher prices (Goldsmith 1996, 1999). However, the evidence in favor of the negative relationship between innovativeness and price sensitivity is based largely on studies that conceptualize and measure innovativeness as a unidimensional construct (e.g., Goldsmith and Newell 1997). Although this approach is valid, it can be criticized because the concept of consumer innovativeness is multidimensional (see Muzinich, Pecotich, and Putrevu 2003). We propose that consumer innovativeness can be conceptualized as a higher-order construct consisting of four dimensions: involvement, information search, subjective knowledge, and opinion leadership. The purpose of the present study is to reevaluate the negative innovativeness/price sensitivity relationship using a causal model analysis in which innovativeness is represented by a multi-dimensional, second-order factor. To enhance this relationship’s generalizability, we perform the analysis on datasets from samples of consumers from Korea and the U.S., offering a cross-national comparison. Price sensitivity refers to individual differences in how consumers react to prices and price changes and is influenced by economic circumstances, demographics, situational factors, personality, amount of product use, and product involvement (Munnukka 2005). Price sensitivity varies across product categories as well as reflecting an overall marketplace disposition. Studies show that when consumer innovativeness is measured directly, a negative relationship exists between price sensitivity and innovativeness; consumers who want and buy the newest products are willing to pay more for them than less interested consumers (Goldsmith 1996, 1999; Goldsmith and Newell 1997). U.S. and Korean adults and college students completed the survey. The U.S. data came from a convenience sample of adult consumers. The sample consisted of 164 men and 151 women, with eight missing values for sex. The respondents ages ranged from 19 to 80, with a mean of 34.4 years (SD=12.5) and a median age of 29.5 years. In Korea the 857 respondents were aged 20 to 73 years, with a median of 27 years and included 398 men and 462 women. Students completed surveys in the classroom. Adults were surveyed via intercept interviews conducted in selected locations around the city of Pusan, S. Korea. The questionnaires contained multi-item scales used to measure fashion involvement, fashion information search, subjective fashion knowledge, fashion opinion leadership, and fashion price sensitivity. To test the hypothesis, we estimated the proposed second-order model with all measurement items from the second-order partial variance invariance CFA using EQS for Windows 6.1. The results revealed an excellent fit of the model with the empirical covariances with χ2= 1371.79 on 426 df, NNFI=.940, CFI=.944, CAIC= −2063.07, and RMSEA=.044 (Hu and Bentler 1999). Relying on the excellent fit of the structural model, we tested the hypothesis of a negative effect of innovativeness on price sensitivity. The empirical results support the hypothesized negative effect (p<.01) with a standardized path coefficient of −.546 in Korea and −.569 in the U.S. Although the two coefficients are significantly different at .05, comparing these two coefficients is inadequate as the second order measurement model only supported partial variance invariance (Steenkamp and Baumgartner 1998). These results provide evidence for our novel conceptualization of consumer innovativeness and show that the negative innovativeness/price sensitivity relationship observed in U.S. consumers can be generalized to Korean consumers as well. Managerially, we see that managers can price new fashions at higher levels when they seek to sell them to fashion innovators than when they sell to later adopters. Basically, having the newest thing is worth the extra cost. Moreover, the findings extend the scope of the innovativeness/price sensitivity relationship to a new culture in which it has not yet been studied. This is evidence for the robustness of the theory of consumer innovativeness. The study supports the argument that consumer innovativeness can be conceptualized as a multi-dimensional construct. This may enable researchers to incorporate many variables into their models of consumer behavior that they could not do so previously because they lacked valid, reliable measures of these constructs. Methodologically oriented researchers should also examine possible negative aspects of this approach as well so that it is not accepted unquestionably by the research community. In conclusion, we feel that the study achieves its goal of introducing a new view of fashion innovativeness and a new way to operationalize it cross-culturally. By conceptualizing fashion innovativeness as a second order concept, this study offers a more comprehensive approach to its measure. Furthermore, the dual study context, Korean and U.S. consumers, suggests a possibility that fashion innovativeness results in price insensitivity cross culturally.