Hispanics and Patronage Preferences for Shopping From the Internet

Authors

  • Pradeep Korgaonkar,

    Corresponding author
    1. The Internet Coast Institute Adams Professor of Marketing at Florida Atlantic University. His research interests are in the areas of e-commerce, advertising, and retailing. His research has been published in numerous journals including the Journal of Advertising, the Journal of Advertising Research, the Journal of Business Research, the Journal of Consumer Marketing, the Journal of Current Issues and Research in Advertising, the Journal of Internet Research, the Journal of Retailing, and the Journal of Marketing Research among others.
    Search for more papers by this author
  • Ronnie Silverblatt,

    Corresponding author
    1. Associate Professor of Management and International Business, College of Business Administration, at Florida International University. She holds a Ph.D. in Business Administration from Georgia State University. Some of her prior works have been published in Industrial Relations, Journal of Labor Studies, Journal of Business Research, and theJournal of Applied Business Research. Her areas of interest include human resource management, Hispanics and e-commerce.
    Search for more papers by this author
  • Enrique Becerra

    Corresponding author
    1. Instructor in the College of Business at Florida Atlantic University.
    Search for more papers by this author

Address: College of Business, Florida Atlantic University, 220 SE 2nd Avenue, Ft. Lauderdale, FL 33301. Phone: 954-762-5218.

Address: College of Business, Florida International University, University Park, Miami, FL, FL 33152.

Address: College of Business, Florida Atlantic University, 220 SE 2nd Avenue, Ft. Lauderdale, FL 33301.

Abstract

The Hispanic market in the U.S. offers promising and lucrative online business opportunities. Spain's Terra Network, one of the biggest online content and access providers in the Latin countries, has teamed up with New Jersey based IDT to provide access to U.S. Hispanics (Folpe, 2000). Approximately, 1.3 million households and 2.3 million Hispanic small businesses are using the Web. Still, little published research exists documenting the shopping preferences for buying from the Web by this growing segment of the U.S. population. Applying the product classification and perceived risk literature from the previous published studies of Girard, Korgaonkar, Silverblatt (2002), the authors explore the Hispanic Web user's preferences for shopping from the Web. The study results and implications for managers are discussed.

Introduction

The recent estimates released by the U.S. Census Bureau show explosive growth in the number of Hispanics in the U.S. The U.S. Hispanic population has increased from 6 percent of the population in 1980 to 13.3 percent of the U.S. population in 2002. Today, the number of Hispanics in the U.S. is about 37 million. Although 9 out of 10 Hispanics currently live in only 10 states (CA, NV, AZ, CO, NM, TX, FL, NY, NJ, CT), the ethnic market is becoming more mainstream as reflected in the fact that in the years from 1990 thru 2002 the Hispanic population more than doubled in states of Arkansas, Nevada, North Carolina, and Georgia. By the U.S. Census estimates, in the year 2010 the Hispanic population is projected to reach 43.7 million. Small wonder that, increasingly, the Latin culture has captured the attention of American business (Assimilation, 1997; Vijayasarathy, 2000). According to one estimate about 1.3 million U.S. Hispanic households are users of the Internet (Greenberg, 1999). A recent study sponsored by the Association of Hispanic Advertising Agencies (2000) states that 38 percent of Hispanics 16 years and older are regular users of the Internet. The study also estimates the U.S. Hispanic buying power at $458 billion. Espanol.com, the first large scale Internet site for Spanish speakers, surveyed 2000 U.S. Hispanics and reported that Hispanic shoppers spend an average of $547 online (Espanol.com, 2000). However, little is known regarding what kinds of products are suitable for selling on the Internet to this lucrative and growing ethnic group. The purpose of the study reported here was three-fold. First, we seek to understand the influence of the type of product on Hispanic consumers' shopping preferences for buying from the Web. Second, we seek to find significant relationships, if any, between the type of e-tailer and consumers' shopping preferences, and finally, we propose to determine if there is an interaction between the type of product and the type of e-tailer on Hispanic consumers' preferences for shopping on the Internet.

The World Wide Web presents advertisers with opportunities and challenges, including the need for understanding Web users' attitudes toward and beliefs about this new medium's advertising potential. Unquestionably, the burgeoning Internet has fast become an important new sales, distribution, and advertising channel for commerce. America Online is pursuing the Hispanic market aggressively for AOL Latino 9.0 by showing commercials featuring its Running Man icon learning Spanish. However, little published literature exists documenting shopping preferences of the Hispanic consumers in this growing market. Next, we will review the relevant literature

Literature Review

Product Type

Although prior research studies suggest that consumers' buying behavior varies across product categories (Porter, 1974), little empirical research exists examining the effect of product categories on online shopping preferences (Ainsile & Rossi, 1998; Girard, Silverblatt, & Korgaonkar, 2002; Klein, 1998). A number of relevant research studies use a product classification model based on search, experience, and credence paradigms to examine the Internet's influence on information search (Klein, 1998), explain the buying decision process (Asch, 2001), and purchase intentions and loyalty to certain Web sites (Lynch, Kent, & Srinivasan, 2001). Similarly, Brucks, Zeithaml, and Naylor (2000) use search, experience, and credence product classifications along a continuum to develop a typology of quality dimensions for durable goods. Although not testing empirically, Asch (2001) distinguishes between search, experience, and credence good and service properties that affect consumers' purchasing decisions and the involvement level of their shopping experience., As explained by Girard, Silverblatt, & Korgaonkar (2002) the past studies draw the definitions of search and experience goods from the seminal work of Stigler (1961) and Nelson (1970) which was further refined by Wright and Lynch (1996).

Originally derived from Stigler's (1961) clarification of the “search” phenomena and the theory of economics of information, the qualities of search and experience goods were first defined and used by Nelson (Nelson, 1970) to describe consumers' skepticism about advertisers' exaggerated claims and consumers' attempts to verify the reliability of those claims. Nelson (1974) defines a search good as one whose qualities a consumer can determine by inspection prior to purchase. Therefore, information about the dominant attributes can be identified before purchasing the product. Nelson (1974) defines an experience good as one whose qualities cannot be determined prior to purchase.

Building on Nelson's (1970) definition of experience good, Klein (1998) presents two circumstances in which a good is considered an experience good. First, “full information on dominant attributes cannot be known without direct experience.” Second, the “information search for dominant attributes is more costly or difficult than direct product experience” (p. 199). Wright and Lynch (1996) further refine Nelson's definition of experience goods by including “after using” rather than “after purchasing,” taking into consideration that consumers may experience the product by receiving free samples in a store without having to purchase the product.

The definition of a credence good originated by Darby and Karni (1973) builds on the notion that the consumer cannot know the quality possessed by a product even after its use. Consumers are unable to evaluate credence products with confidence before and after use; therefore, they rely on outside experts (Ford, Smith, & Swasy). Credence qualities are usually found in products such as medicines and anti-wrinkle creams, or in financial services like pension plans and professional services like insurance or medical treatments (Asch, 2001).

Based on the product classification theory developed by Nelson (1970, 1974), Klein (1998), and Darby and Karni (1973), this research study uses four product categories: search, experience-1, experience-2, and credence. Because the products and services are classified by these research studies based on differences in the availability and/or cost of the relevant product attribute information to consumers, the authors expect that consumers' preference to purchase on the Internet will significantly differ across product categories irrespective of the type of e-tailer.

Bhatnagar, Misra, and Rao (2000) assert that as consumers become more knowledgeable, their perception of risk decreases. Since the information for search products is easy to obtain and largely available, consumers' knowledge would be higher for those products than the products in other categories. Therefore, consumers' preference to purchase search products will be greatest. Based on Klein's (1998) definitions of the two types of experience goods (experience-1 and experience-2), information search for experience-2 products is more costly and difficult than for experience-1 products. According to Darby and Karni (1973), the credence products are the hardest to evaluate even after purchase or consumption; thus, consumers' preference to shop online for credence products would be the lowest for the credence products compared to others. Therefore,

H1: Consumers' overall preferences for shopping online will be highest for search products, high for experience-1 products, medium for experience-2 products, and the lowest for credence products.

Online Retail Store Type

Although a limited number of empirical research studies exists in the marketing literature studying the role of consumer perception of risk in selecting different types of traditional shopping channels, no attention has been devoted to the role of perceived risk in online shopping from different types of retailers (i.e., prestigious or discount store based e-tailer, pure play e-tailer). Due to the limited research on consumer perception of risk in online retailer types, the hypotheses of this study are built based on the empirical findings of the extant empirical research on the level of consumers' perceived risk in the traditional in-store shopping channels (i.e., department, specialty and discount retail stores, catalog showrooms) and non-in-store shopping channels (i.e., catalog orders by mail or telephone, Internet) (Bhatnagar, Misra, & Rao, 2000; Cox & Rich, 1964; Festervand, Snyder, & Tsalikis, 1986; Hisrich, Dornoff, & Kernan, 1972; Korgaonkar, 1982; Korgaonkar & Moschis, 1989; Miyazaki & Fernandez, 2001; Prasad, 1975).

Findings from previous studies suggest that in-store shopping is perceived to be less risky than non-in-store shopping via telephone or mail catalog order (Cox & Rich, 1964; Festervand, Snyder, & Tsalikis, 1986). Additionally, a small number of research studies investigated the role of perceived risk of product in the selection of a shopping channel (Bhatnagar, Misra, & Rao, 2000; Hisrich, Dornoff, & Kernan, 1972; Korgaonkar, 1982; Korgaonkar & Moschis, 1989; Prasad, 1975). The results of these studies indicated that consumers perceived a lesser amount of risk for the department stores, specialty stores and catalog showrooms for high social risk products than that of the perceived risk for the discount stores. These results indicate that “perceived risk of a product is transferable to the store that sells the product” (Korgaonkar, 1982, p. 78]. This assumption can be justified based on the fact that the department stores generally carry a better quality and selection of merchandise than the discount stores. Given this concern, we predict that online shoppers will perceive the lowest risk for shopping from a prestigious store Web site, a medium amount of risk for shopping from a discount store Web site, and the highest risk for shopping from a pure play e-tailer (see Table 3). The following hypotheses overall suggest that consumers' online shopping preference will vary by the type of Internet store irrespective the type of product. More specifically,

Table 3.  Pairwise comparison of product category adjusting for number of comparisons (Bonferroni).
Pairwise ComparisonMean DifferenceSignificance*
  1. *Adjustment for Multiple Comparisons: Bonferroni

Search Products – Experience11.8290.00
Search Products - Experience 22.4330.00
Search Products – Credence1.5400.00
Experience 1 – Experience20.6030.01
Experience1 – Credence−0.2890.77
Experience2 – Credence−0.8920.00

H2: Consumers' online shopping preferences will be higher for a prestigious department store Web site than a discount store-based Web site for all product categories.

H3: Consumers' online shopping preferences will be higher for a discount store-based Web site than for a non-store-based Internet retailer for all product categories.

H4: Consumers' online shopping preferences will be higher for a prestigious department store Web site than a non-store-based Internet retailer for all product categories.

Interaction Effect between Product Type and Online Retail Store Type on Preference for Shopping Online

The studies by Bhatnagar, Misra, & Rao (2000), and Miyazaki and Fernandez (2001) suggest that the perceived product risk affects consumers' preferences for not only selection of retail stores (in-store or non-store) but also the product types purchased. Based on the findings of these studies,we predict that online retail store type and product type will have an interaction effect on consumers' preferences for shopping from different types of Internet retailers. That is, for the different levels of perceived risk in different product categories, consumers will prefer to shop online from different type of Internet retail stores because of the preceived store risk. For search products, however, the authors hypothesize that consumer preferences will be greates for shopping online because information about the search product category is available regardless of the type of the Internet retail store by which the product is sold. That is, consumer preferences to shop online for search products such as books and CDs will be the same for the three types of Internet retail store (prestigious store-based, discount store-based, and non-store-based Web sites). Therefore,

H5: For search products, shopping preference will be similar for all types of e-tailers.

Based on the findings of the study by Bhatnagar, Misra, & Rao (2000), consumers would perceive low risk for experience-1 products such as clothing and perfume. That is, consumers can be expected to prefer to shop for experience-1 products from a non-store- based e-tailer since the low perceived product risk will help reduce the perceived store risk for the non-store-based e-tailer and for a discount store-based e-tailer. Since the perceived store risk was predicted to be the highest for a non-store-based e-tailer, medium for a discount store based e-tailer, and the lowest for a prestigious store-based e-tailer, consumer preference to shop online will be the highest for experience-1 products from a non-store-based e-tailer, medium from a discount store based e-tailer, and the lowest for a prestigious department store based e-tailer. Therefore,

H6: For experience-1 products, online shopping preference will be low for a prestigious department store e-tailer, medium for a discount store-based e-tailer, and high for a non-store-based e-tailer.

The findings by Bhatnagar, Misra, & Rao (2000) also suggest that perceived product risk was higher for technologically complex and expensive products (i.e., electronics, hardware) purchased online. For experience-2 products such as mattresses and televisions, getting the relevant attribute information would be more difficult/costly than actual product/service experience prior to its purchase. Therefore, online shoppers would not be confident of making the purchase decision without using/sampling the product/service prior to its purchase. With this concern, consumers are most likely to purchase experience-2 products from a prestigious department store based e-tailer, next most likely from a discount store-based e-tailer, and least likely from a non-store-based e-tailer. Therefore,

H7: For experience-2 products, online shopping preference will be high for a prestigious department store e-tailer, medium for a discount store-based e-tailer, and low for a non-store-based e-tailer. Credence products such as vitamins and anti-wrinkle cream cannot be evaluated until after they are purchased and used. Because the perceived product risk is highest for the credence category, consumers will select the Internet store for which they perceived the lowest risk. The prestigious department store-based e-tailer would be preferred as a first choice, discount store-based e-tailer as a second choice, and a non-store-based e-tailer as the last choice. Therefore,

H8: For credence products, online shopping preference will be high for a prestigious department store e-tailer, medium for a discount store-based e-tailer, and low for a non-store-based e-tailer.

Prior Internet Purchase Experience: A Co-Variate

Several past studies in the literature suggests that consumers' preferences for shopping from Internet retailers will be significantly related to their past purchase experiences [e. g. Ghani & Deshpande, 1994; Pitt, Berthon, & Watson, 1996; Sultan, 2002; Van den Poel & Lenuis, 1999) Hence, the aforementioned hypotheses were tested by controlling for the past Internet purchase experiences of the shoppers in the study.

Method

The First Stage Sample Data

The data were collected from two independent samples in two different stages. The first stage sample data were collected from 32 Hispanic students from a state university in the United States. Given the definitions of each product type (search, experience-1, experience-2, and credence) derived from the literature (Ford, Smith, & Swasy, 1990; Klein, Ai-Mei, & Whinston, 1998; Nelson, 1974), the respondents were asked to list a maximum of four products that they felt represented each category. The sample consisted of an equal number of males and females, from various professions. The four descriptions are adopted from (Darby & Karni, 1973) and are provided below:

The first product class description was for a search product/service whose relevant attribute information could be easily obtained prior to the use or purchase. They would be confident of making the purchase decision without using/sampling the product/service prior to its use or purchase.

The second product class description was for an experience-1 product/service whose relevant attribute information could not be known until the use of the product. They would not be confident of making the purchase decision without using/sampling the product/service prior to its purchase. The third product class description was for an experience-2 product/service for which it was more difficult/costly to get the relevant attribute information than actual product/service experience prior to its purchase. They would not be confident of making the purchase decision without using/sampling the product/service prior to its purchase.

Lastly, the fourth product class description was for a credence product/service for which relevant attribute information was not available prior to as well as after the use of the product/service. They would not be confident of their purchase decision even after using/sampling the product/service.

A total of 121 products were listed by the respondents under the search product category, 112 products were listed under the experience-1 category, 88 products were listed under the experience-2 category, and 99 products were listed under the credence product category. Based on the examination of the product list provided by the respondents, the authors selected the two most representative products for each product category to be used in the second phase of the study. The authors also focused on those products that are more readily available on the Internet so as to make the second phase of the study relevant to the respondents. The products are: books and music CD search category, clothing and perfume experience-1 category, mattress and television experience-2 category, vitamins and anti-wrinkle cream credence category.

The Second Stage Sample Data

The second stage data were collected in two counties with a total of approximately 4 million residents in a southern state of the United States. The respondents were over 18 years old and regular users of the Internet. The study sample consisted of 350 consumers from a large southeastern metropolitan area. They were contacted on different days of the week and times of the day for their study participation. Given the nature of the study topic, only those who indicated they had used the Web were selected to participate in the study. Probability sampling for the project was difficult given that Hispanics tend to be more wary of participating in survey research than other ethnic groups for a variety of reasons. Concerns about participating in a study are exacerbated because of the fears that personal information could be used against them by immigration/government authorities as well as fears that unethical businesses will use the information to exploit them (e.g., Marin & Vanoss Marin, 1991). Hence, following the recommendations that the survey of Hispanic population be conducted by persons familiar with the community and/or of similar background characteristics, including ethnicity, in a personal face-to-face situation (Marin & Vanoss Marin, 1991) data for the study was collected via personal interviews. The respondents were given the choice of responding to English or Spanish questionnaires. Although attempts were made to sample respondents to reflect the Hispanic profile of the community from which they were selected, several factors were inhibiting. One was the fact that we only were interested in those who had used the Web before. In additions to the fears alluded to earlier, the lack of accurate listing of population, the presence of illegals, and the practice of relatives living in a common household, all limited our ability to select and administer a probabilistic sample (Pl. see especially Marin & Vanoss Marin, 1991, for a detailed discussion).

The demographic profile of the respondents is depicted in Table 1. The study results are based on the complete responses of 258 consumers. In comparing the sample to the Census 2000 of the local area the authors found that the sample was skewed towards more highly educated, higher income, and skilled individuals as well as professional types of occupations (U.S. Census Bureau, 2002). This was expected because the authors surveyed only those who were regular users of the Internet. However, the sample profile is similar to the national profile of Internet users as reflected in the 1998 GVU survey.

Table 1.  Sample characteristics.a
Characteristic**Percent
Gender 
Male50.0
Female50.0
Age 
Under 20 years0.9
20–30 years62.3
31–40 years18.1
41–50 years13.7
51–60 years3.7
Over 60 years0.9
Education Level 
High school22.3
Trade school12.5
Some college38.8
College graduate20.8
Post graduate5.5
Annual Household Income 
Under $20,00021.2
$20,000 –$40,00035.2
$40,001 –$60,00018.0
$60,001 –$80,00010.8
$80,001 –$100,0008.4
Over $100,0006.4
Country of Origin*** 
Cuba15.7
Colombia18.8
Puerto Rico2.4
U.S.A21.2
European2.4
Other39.6

The Second Stage Survey Instruments

The survey was administered to different groups of respondents for three online retail store types and four product categories. The survey instrument contained several sections. In section one, the authors asked the respondents to indicate their preference for purchasing from the Internet each of the eight products selected from the stage one of the study. The preference for purchasing each product from an Internet retailer was measured on a five point scale ranging from (1) “May Never Buy” to (5) “May Prefer Buying” for each of the selected eight products.

In the second section, the subjects were given a list of 50 statements, adopted from Vijaysarathy and Jones (2000) and Eastlick and Feinberg (1999) after an exhaustive search of the literature in the area of Internet retailing as well as direct marketing. The 50 statements were designed to capture 11 dimensions of Internet retailers. The eleven dimensions that the fifty statements measured are Perceived Value (6 items); Convenience (6 items); Economic Utility (6 items); Home Shopping (3 items); Merchandise Assortment (4 items); Order Services (4 items); Company Clientele (4 items); Information Services (7 items); Customer Service (4 items); Security/Privacy (3 items), and Internet Retailer Reputation (3 items). The respondents were asked to rate the importance of the 50 Internet retailer attributes for purchasing the two products from one of the four product classifications on a scale of (1) “Not Important at All” to (5) “Extremely Important.” For example, one group of respondents was asked to assume that they wanted to purchase a product such as a book or music CD (search products) from an Internet retailer and they were asked to rate the importance of various Internet attributes in choosing a specific Internet retailer to purchase from.

A separate set of respondents was asked to assume they were purchasing products such as clothing and perfume (experience-1 products). A third group of respondents was asked to assume they were purchasing products such as mattresses and televisions (experience-2 products). The fourth group of respondents was asked to assume they were purchasing products such as vitamins and anti-wrinkle cream (credence products). The administration of the survey instruments was randomized to prevent a response bias. There were no statistically significant differences in the demographic profiles of the four groups of respondents. A total of 250 valid surveys were obtained: 76 observations came from the discount store-based e-tailer survey; 88 observations were collected from the prestigious store-based e-tailer survey, and 86 observations came from the non-store-based e-tailers. The breakdown of the sample size in each product category for the three types of Internet stores is as follows: 250 for books and music CD, 246 for clothing and perfume, 248 for mattresses and televisions, and 249 for vitamins and anti-wrinkle cream (see Table 2).

Table 2.  Sample size for each type of Internet stores, and each type of product, means and standard deviations for shopping preferences.
 Search Products Mean/St. DevExperience-1 Products Mean/St. DevExperience-2 Products Mean/St. DevCredence Products Mean/St. DevTotal Mean/St. Dev
Discount store based e-tailer7.59/1.73 n=765.68/2.04 n=734.95/2.29 n=756.21/2.49 n=756.12/2.35
Prestigious store based e-tailer6.99/1.70 n=885.29/2.01 n=874.94/2.20 n=885.47/2.43 n=885.67/2.24
Non-store-based e-tailer7.45/1.70 n=865.57/1.98 n=864.85/2.10 n=855.73/2.21 n=865.90/2.26
Total7.73/1.745.50/2.014.91/2.215.78/2.415.89/2.21

The past Internet purchase experience of the respondents was measured on a seven-pont scale. Finally, we collected demographic information including gender, household income level, occupation, education, and ethnicity. The demographic information collected in the study included gender measured as (1) male, (2) female; age measured on a scale of (1) under 20 years through (6) over 60 years; education measured on a five point scale of (1) high school through (5) post graduate/professional training; annual household income measured on a six point scale of (1) under $20,000 through (6) over $100,000. Finally, the respondents were asked to indicate if they were born in the U.S. or not. Those who were born in another country were asked to indicate the country of birth.

The data for the study were collected by personal interviews using English and Spanish languages questionnaires. The English version was pretested and the validity as well as the reliabilities of the constructs was well established prior to its translation into Spanish. The surveys were first translated from English to Spanish by a bilingual translator. A second bilingual translator translated the Spanish version back to English to ensure the exactness of meaning between the two versions of the instrument. A few discrepancies between the two were resolved by a common agreement.

The sample consisted of an almost equal number of males (50%) and females (50%), with some college or higher level of education (65.1%), mostly under 40 years of age (80.4%), with income between $20,000 and $40,000 (35.2%). Of those who responded to the national origin question (n=201), the major responses were Cuba (15.7%), South America (39.6%), Colombia (18.8%), and the U.S. (21.2%). Compared to the demographic composition of the area, the sample was over-represented in terms of younger and college-educated respondents. This over-representation was not surprising since we surveyed only those consumers who had previous experience with the Web. Table 1 exhibits our sample's characteristics.

Hypotheses Testing and Results

Hypothesis 1

The preference scores for each pair of products (book and music CD, clothing and perfume, televisions and mattresses, anti-wrinkle cream and vitamins) were combined into four product categories (search, experience-1, experience-2, credence) to compute the overall preference for shopping online for each product category. ANOVA analysis was used to test the hypothesis evaluating whether consumer preferences for each product category are significantly different from those for the other product categories. A pairwise comparison for each product category was performed using the Bonferroni adjustment for the number of comparisons. Hypothesis 1 is supported with the significant p-value at 0.01 indicating that the preference to shop online for the search products was the highest compared to the experience and credence products. There were significant differences between the preference mean scores of experience-1 and experience-2 products (0.01) as well as experience-2 and credence products (0.00) at the 0.01 significance level. However, there were no statistically significant differences between the mean scores of preference for shopping online for experience-1 and credence products(Table 3).

Hypotheses 2 Through 4

To test the hypotheses 2, 3, and 4, the preference scores for the three online retail store types (prestigious store-based e-tailer, discount store-based e-tailer, and non-store-based e-tailer) were combined and Analysis of Variance was conducted(Table 4). The F-statistic result was significant (F=3.420, p =.065). Therefore, hypotheses 2, 3 and 4 were supported. Three paired t-tests among the three types of stores again did indicate significant differences for the online patronage preferences at 0.05 level or better. The results indicate that online retail store type did have a significant effect on preference for shopping online. Specifically, irrespective of the type of product, e-tailer with discount operation was most preferred followed by pure play e-tailer and e-tailer of a prestigious operation, respectively. A post-hoc analysis shows(Table 4) only that the e-tailer with discount operations is statistically significantly different (at .05 level) from the e-tailer with prestigious operations.

Table 4.  Analysis of variance (ANOVA) results : Shopping preference by type of product (4 levels) and type of e-tailer (3 levels) controlling for prior purchase experience, and e-tailer pairwise comparison adjusted for multiple comparisons (Bonferroni)
ANOVA Test Between SubjectsSum of SquaresdfFSignifican
  1. *Adjustment for Multiple Comparisons: Bonferroni

Intercept4250.6651755.8850.00
Purchase15.1713.420.06
Product Category798.9543100.2220.00
E-tailer Type30.39625.7170.04
Interaction Effect (product*e-tailer type)15.91660.5980.73
E-tailer Type Pairwise ComparisonMean DifferenceSignifican
Non-store based e-tailer – Discount store based e-tailer−0.1840.810
Non-store based e-tailer – Prestigious store based e-tailer0.2460.376
Discount store based e-tailer – Prestigious store based e-tailer0.4300.029

Hypotheses 5 Through 8

The hypotheses 5 through 8 were tested by using ANOVA two-factor design with 3x4=12 treatment combinations. The three online retail store types correspond to the three treatments of the first independent variable, and the four product categories correspond to the four treatments of the second independent variable. The consumers' preference scores of the eight products were used as the dependent variable.

The factorial ANOVA results with the three online retail store types indicated no statistically significant interaction effect of online retail store type and product type on consumers' online shopping preference. Similarly, no significant interaction effects of online retail store type and product type were found when two instead of three online retail store types together were tested: prestigious store-based and non-store-based (F=0.558, p <0.65); discount store-based and non-store-based (F=0.297, p <0.90) and prestigious store based vs. discount store based (F=0.933, p <0.50).

Conclusion

The evolution of the Web has important implications for the way businesses perform marketing tasks. Interactive technology has the capability to change the way we do business in a variety of ways. The Web's future includes high fragmentation similar to what we have seen in television, radio, and print media. Much like CNN delivers information, ESPN delivers sports, and The Home Shopping Network delivers shopping goods, the Web is beginning to fragment. America On Line has launched LAOL, an on-line resource for Hispanics focusing on specific needs of the growing Hispanic market. The successful launch of StarMedia, a leading Internet company catering to the Latin market (Schrage, 1999, is another example. The future of the Web indicates further fragmentation. The Internet may be used as tool for entrepreneurs to provide new services and opportunities to ethnic user groups. The findings of this study add significantly to our understanding of why and how the Hispanic consumers use the Web. Thus, for practitioners and researchers alike, understanding why and how these consumers use the Web may be the key to unlocking the Web's capacity. The Web has drastically changed the buyer-seller relationship, tipping the balance of power in favor of consumers, as the interactive feature of the technology puts the consumer in control. How the “new” consumers view the new medium and use it may have a significant impact on the future of marketing to them.

The Hispanic News (http://www.hispanic.bz) suggests that there were 12.4 million Hispanic online users in the U.S. This is stated to be larger than the total online population of Spain and four percent larger than the total online population of Mexico, Argentina, and Columbia combined. The attractiveness of the U.S. Hispanic market has lured even foreign companies such as Spain's Terra Network to establish a significant presence here (Folpe, 2000). This study documents that the Hispanic Web users' preferences for shopping on the Web are influenced by the type of product and the type of Internet retailer. One of the findings of this study is that consumers' overall preferences for shopping online differ based on the product type irrespective of the online retailer types. More specifically, preferences to shop online for the search products were the highest among the four product categories. Preferences to shop online for credence products were significantly higher than experience-1 and experience-2 products. These findings suggest that consumers prefer to shop online for the types of products that they can easily obtain information about. Therefore, online retailers can significantly increase consumers' preferences to shop online for experience products by providing detailed information about the attributes of those products as well as third party reviews of consumers' experience with the products and satisfaction. The studies of Wright and Lynch (1995) and Girard, Silverblatt, and Korgaonkar (2002) suggest that advertising is more suited to search and credence products whereas actual experience or a suitable surrogate of it is more suited for experience products. For example, Autobytel.com, the etailer of automobiles, encourages consumers to visit a dealership and test drive a car before purchasing it on-line. The findings of the present study are in also line with Bhatnagar, Misra, and Rao (2000) in that ease of obtaining product information and increased level of product knowledge decrease the perception of risk in online shopping.

The present study also finds that consumers' preferences do significantly differ among the online retailer types as well as for the product categories. Therefore, different types of retailers practicing business on the Internet, whether they are prestigious bricks and clicks, discount bricks and clicks or non-store-based e-tailers (pure plays), do have a significant advantage over one another. Specifically, the Hispanic shoppers preferred shopping from the Web site of a discount retailer, followed by the pure play e-tailer and lastly from the Web site of a prestigious retailer. However, the interaction between the type of product and type of e-tailer was not supported. In sum, it is clear that practitioners and researchers need to pay more careful attention to the needs of the Hispanic Web users. Examining these needs may provide the means to understanding the under-utilized potential of the Web. The study results are limited to the Hispanic groups investigated in the study. Future studies could benefit by expanding the sample to include more Mexican-Americans as well as by investigating differences if any between the Hispanic subgroups based on their ethnic/country of origin. Investigation of this large and profitable ethnic markets' shopping preferences will benefit researchers and practitioners eager to unravel the Internet's potential for commerce.

Ancillary